article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
in cold regions of the interstellar medium dust grains are coated with ices
h@xmath0o , co , co@xmath0 and additional minor ice constituents through a combination of condensation and an active surface chemistry .
these ices account for up to 60 and 80% of the volatile oxygen and carbon budget in star forming regions respectively @xcite . of the three major ice constituents ,
this study focuses on co@xmath0 .
co@xmath0 is an important ice component during all stages of star formation .
co@xmath0 ice has been observed in dense clouds @xcite , protostellar envelopes @xcite and in comets @xcite , the remnants of the solar nebula . in this range of environments , co@xmath0 abundances with respect to water
is surprisingly constant at 2030% .
interstellar and cometary co@xmath0 identifications as well as column density determinations rely on infrared spectroscopy .
co@xmath0 ice has two ir active fundamental modes : the 4.27 @xmath1 stretch ( @xmath2 ) and the 15.2 @xmath1 bending ( @xmath3 ) modes , both of which have been identified in ices observationally .
furthermore , co@xmath0 ice spectra have been shown to depend sensitively on the local ice environment .
the 15.2 @xmath4 m band displays a characteristic davydov splitting in the pure crystalline phase .
amorphous co@xmath0 ice and co@xmath0 in ice mixtures do not show this splitting , and the band is typically shifted with respect to the pure ice band position .
the frequency and shape of the co@xmath0 @xmath2 stretch also depend on the bulk ice structure .
several studies have pointed out that the band displays a low frequency shoulder at 2328 @xmath5 ( 4.30 @xmath4 m ) in pure amorphous co@xmath0 and in some co@xmath0 mixtures with hydrogen - bonding molecules @xcite .
this environmental sensitivity has been used to identify several different co@xmath0 ice phases in observational spectra : in star forming regions , most co@xmath0 is typically mixed with water , but co : co@xmath0 and pure co@xmath0 ice phases are also common @xcite .
co@xmath0 ice spectroscopy has also been used in the laboratory to trace co@xmath0 trapping , segregation and diffusion processes @xcite .
in addition to the normal vibrational modes , co@xmath0 optical phonons , arising from long range collective vibrations in the solid , can be excited by infrared radiation .
a phonon is a quantized vibrational motion in which the lattice atoms or molecules vibrate at a single frequency .
optical phonons occur when the molecules are moving out of phase within the lattice .
these phonons can propagate through thin films of astronomical dimension but can not travel through films of thickness much greater than the wavelength of incoming radiation .
thick films will instead exhibit restrahalen bands in which the change in the refractive index results in a strong reflection .
solid co@xmath0 exhibits both tranverse optical ( to ) and longitudinal optical ( lo ) phonons in which the normal vibrations propagate through the ice lattice perpendicular and parallel to the direction of the ir field respectively .
when a thin film is positioned at an oblique angle to the incoming radiation , the electric field has components both parallel and perpendicular to the film normal and longitudinal optical phonons can be excited by the parallel component of the field vector .
the angular dependent enhancement of the lo mode is known as the berreman effect after experiments conducted by berreman on the angular dependence of the lo mode in ionic crystals @xcite .
the theory of optical phonon mode splitting has been described previously for polycrystalline films @xcite .
the magnitude of the splitting is estimated by : @xmath6 where @xmath7 and @xmath8 are the frequencies of the lo and to phonon modes , @xmath9 is the dielectric constant of the ice , _ m _ is the reduced mass associated with the coordinate _ q _ , _ v _ is the unit cell volume and @xmath10/@xmath11 is the transition dipole moment matrix element . in many cases
the splitting can be accurately described by the lyddane - sachs - teller ( lst ) relationship : @xmath12 where _
_ n__@xmath13/@xmath14 and _ _ n__@xmath15 /@xmath16 are the limiting low- and high - frequency refractive indices/ dielectric constants of the band .
the high - frequency index is often approximated using the electronic dielectric constant , for example @xcite use reported values of the visible refractive index of the sample at the sodium d line ( @xmath17 = 589 nm ) .
the values used by hudgins _ et .
al . _ for co@xmath0 , o@xmath0 , co and h@xmath0o
n__@xmath15 = 1.22 , 1.25 , 1.30 and 1.32 .
the lst approximation was originally developed for cubic crystals but has been shown to apply to a range of disordered materials @xcite .
lo phonon modes are often invisible in laboratory ice spectra due to the common experimental setup in which the infrared beam is normal to the film surface .
co@xmath0 lo - phonon modes have , however , been reported in reflection - absorption infrared spectroscopy ( rairs ) studies of ices , when the substrate is naturally positioned at a grazing angle to the ir beam .
@xcite report rair spectra of co@xmath0 ices at 12 k and show that new modes appear using p - polarized light at 2377 and 676 @xmath5 . @xcite have also observed co@xmath0 lo phonons in both transmission ir and rairs experiments .
they show that the co@xmath0 @xmath18 mode in p - polarized rair specta shifts towards the lo frequency of pure crystalline co@xmath0 ice ( 2381 @xmath5 ) with increasing ice thickness .
the co@xmath0 @xmath18 lo phonon mode has also been observed in asw films exposed to co@xmath0 at 90 k @xcite and for h@xmath0o / co@xmath0 nanoparticles at 80 k @xcite .
co@xmath0 deposited on amorphous solid water ( asw ) ice at 90 k exhibits an lo mode at 2379 @xmath5 , close to that of pure co@xmath0 ice . composite h@xmath0o
/ co@xmath0 nanoparticles show shape effects which manifest in the ir at frequencies between @xmath7 and @xmath8 .
in this paper we present infrared transmission spectra of the co@xmath0 optical phonon modes in low temperature ices .
we demonstrate that the lo phonon mode shape and peak frequency depends sensitively on ice mixing components , such as h@xmath0o , co and o@xmath0 .
we analyze these dependencies and find a systematic frequency shift with ice mixing ratio for all mixing - partners , regardless of whether the ice mixture is deposited directly , or produced photolytically _ in situ_. in section [ sec : discussion ] , we discuss how these spectroscopic characteristic could be used to characterize co@xmath0 in astrophysical environments , as well as in laboratory ice experiments aimed at constraining ice diffusion .
laboratory ice studies were conducted in a high vacuum system ( base pressure @xmath192 x 10@xmath20 torr , dominated by h@xmath0 ) modified from @xcite .
a closed cycle helium cryostat ( air products displex de 202 - 0sp expander and apd cryogenics hc-4 compressor ) was added to cool the ir transparent substrate , a 0.3 @xmath21 kbr pellet pressed into a tungsten grid that is suspended into the chamber by an ofhc copper coldfinger .
the substrate is cooled to 28 k and the temperature is monitored using a k - type thermocouple welded to tantalum foil and attached to the w - grid close to the kbr pellet .
infrared spectra were collected using a bruker tensor 27 ftir spectrometer and a liquid - n@xmath0 cooled mct detector . in the fiducial experiment
the substrate was kept at @xmath1940@xmath22 to the ir beam , but other angles were explored as well .
h@xmath0o was prepared in a he - purged glove - box and freeze - pump - thawed for several cycles . the purity of the co@xmath0 , co , o@xmath0 and h@xmath0o used to prepare the ice mixtures was checked prior to the experiments using an rga 200 quadrupole mass spectrometer ( qms ) .
ice mixing ratios are calculated using infrared absorption spectra and literature band strengths @xcite when possible , i.e. for h@xmath0o , co@xmath0 and co. uncertainties on ice mixing fractions are calculated by analysis of repeat absorbance spectra of the ice mixtures and the pure ice , combined with a 15% uncertainty in the bands strengths used to calculate the ice mixing fraction .
molecular oxygen is not observed in the infrared due to absence of a permanent dipole .
oxygen is mixed with co@xmath0 in the gasline and the measured o@xmath0 gas fraction was used to approximate the ice mixture .
the gas fraction was measured by monitoring the o@xmath0 and co@xmath0 partial pressures with the qms at m / z = 32 and 44 respectively . for photochemistry experiments ,
a microwave discharge hydrogen flow lamp ( opthos instruments ) was used .
it has been described in detail previously @xcite .
the microwave power was kept at 60 w for the duration of the experiments .
a mixture of 90@xmath23 ar , 10@xmath23 h@xmath0 was used to produce uv light with a sharp lyman - alpha feature @xcite .
figure [ fig : pureco2 ] shows the infrared spectrum of pure co@xmath0 ice at 28 k , taken with the substrate at 40@xmath22 to the incident ir beam .
the lo phonon mode appears at 2381 @xmath5 ( 4.2 @xmath4 m ) and is shifted from the @xmath2 ( to ) mode by 39 @xmath5 .
the co@xmath0 @xmath3 mode also exhibits an lo phonon mode at 677 @xmath5 ( 14.8 @xmath4 m ) , split 17 @xmath5 and 22 @xmath5 from the co@xmath0 doublet peaks at 660 and 655 @xmath5 respectively .
the lo phonon mode disappears as the ice film is rotated such that its plane of incidence is parallel to the ir beam ( figure [ fig : anglesco2 ] ) .
the co@xmath0 ices deposited at 28 k have crystalline spectral characteristics including splitting of the @xmath3 mode and both the frequency and bandwidth of @xmath2 mode , as well as an amorphous feature around 2328 @xmath5 that dominates at low coverage .
the co@xmath0 ices in each of the experiments are 70 - 90 nm thick , calculated by integrating the @xmath24 + @xmath2 mode using the band strengths of @xcite and a density of 1.78 g/@xmath25 as reported for crystalline co@xmath0 .
figure [ fig : deriv ] shows the method for determining the @xmath26 by analysis of spectra of the ice mixtures and pure co@xmath0 ices .
the @xmath7 frequency is found by identifying the minimum in the second derivative spectrum .
second derivative spectra were taken in both the opus software and thermo scientific grams / ai spectroscopy software .
a smoothing function was applied in opus software before taking the second derivative .
the frequencies of the @xmath7 modes from the smoothed spectra in opus agreed within 1% with those determined using grams without smoothing .
the uncertainty on @xmath26 is conservatively taken as the spectral resolution ( @xmath27 1 @xmath5 ) .
ccccc co@xmath0:co & [email protected] & 2377.0 & 3.9 + & [email protected] & 2373.3 & 7.6 + & [email protected] & 2367.0 & 13.8 + & [email protected] & 2363.7 & 17.1 + & [email protected] & 2361.7 & 19.2 + co@xmath0:o@xmath0 & [email protected] & 2375.5 & 5.3 + & [email protected] & 2373.4 & 7.5 + & [email protected] & 2371.0 & 9.9 + & [email protected] & 2369.4 & 11.5 + & [email protected] & 2367.8 & 13.1 + co@xmath0:h@xmath0o & [email protected] & 2373.8 & 7.0 + & [email protected] & 2372.1 & 8.8 + & [email protected] & 2371.0 & 9.8 + & [email protected] & 2366.3 & 14.6 + & [email protected] & 2364.6 & 16.3 + the co@xmath0 experimental spectra and @xmath18 lo mode redshifts versus defect ice fraction for co , o@xmath0 and h@xmath0o ice mixtures are displayed in figure [ fig : spectra ] and table [ tab : mixing ] .
figure [ fig : spectra ] shows the effect of introducing co , o@xmath0 or h@xmath0o into the co@xmath0 ice lattice by mixing with co@xmath0 prior to deposition . increasing the fraction of the defect species in the co@xmath0 ice results in an increasing redshift of the lo phonon mode from the pure ice position .
the right panels of figure [ fig : spectra ] present the @xmath7 redshift versus the fraction of each species in the co@xmath0 ice mixture .
these plots display a positive linear correlation between the defect concentration and the lo redshift for each of the ice mixtures considered .
the slope of the @xmath7 redshift versus the defect fraction are 36 @xmath27 7 @xmath5 n@xmath28/n@xmath29 , 19 @xmath27 4 @xmath5 n@xmath30/n@xmath29 and 18 @xmath27 4 @xmath5 @xmath5 n@xmath31/n@xmath29 for co , o@xmath0 and h@xmath0o ice mixtures respectively . the effect of the ice defect on the position of the lo mode may be quantified by comparing these slopes . for example , we would expect the lo mode to shift by 1 @xmath5 if the co@xmath0 ice contains @xmath193@xmath23 co , @xmath195@xmath23 o@xmath0 or @xmath196@xmath23 h@xmath0o .
in addition to shifting the frequency of the lo mode , the three mixing partners also affect the shape of the lo mode spectral band , and these spectral changes are species specific . in co mixtures with co@xmath0 , a second small peak appears in the spectra of mixtures with more than 30@xmath23 co. this peak is also redshifted from the pure lo phonon mode position but does not appear to shift with co concentration beyond 30% , possibly representing a stable mixing environment between co and co@xmath0 such as the formation of a complex .
a similar second peak is present in all of the mixed o@xmath0:co@xmath0 ice spectra .
the spectra of h@xmath0o mixtures with co@xmath0 differ from those of the co and o@xmath0 mixtures in several ways . in particular , the spectra of co@xmath0 in polar h@xmath0o environments show broadening of the @xmath2 mode as well as non - linear dependence of the @xmath2 frequency on the fraction of h@xmath0o in the ice .
the co@xmath0 bending mode for the ice mixtures are shown in figure [ fig : v2 ] .
the lo phonon mode is split from the @xmath32 mode at 660 @xmath5 by 17 @xmath5 , while the @xmath33 peak at 655 @xmath5 does not display appreciable lo / to splitting .
the @xmath3 lo mode behaves similarly to the @xmath2 mode , with linear shifts that are dependent on the fractional abundance of the defect in the ice .
the redshift is likewise the largest for co ice mixtures with respect to the ice fraction . both co and o@xmath0 ice mixtures display decreased davydov splitting of the @xmath3 band with respect to the pure co@xmath0 ices .
the @xmath3 bands in h@xmath0o mixtures are broad and do not display davydov splitting , due to the effect of the hydrogen bonding network of the h@xmath0o lattice on the co@xmath0 crystal structure .
the decrease of davydov splitting in the ice mixtures supports the fact that the co@xmath0 crystal structure is disrupted upon addition of the defect molecules as pairwise binding interactions between co@xmath0 and the introduced molecule can result in only one equivalent co@xmath0 per unit cell .
this is consistent with the fact that the davydov splitting continues to decrease in the ice mixtures with increasing concentration of co or o@xmath0 as more interactions become available in each unit cell .
we also conducted experiments to measure the introduction of defects _ in situ _ by uv irradiation of pure co@xmath0 ices , which is known to produce co. figure [ fig : photo ] shows spectra of pure co@xmath0 ice irradiated with uv photons from a hydrogen discharge lamp for a period of 140 minutes .
the co@xmath0 lo phonon mode clearly redshifts with exposure time .
the redshift is linear with co photoproduct growth , similar to when a defect is introduced through pre - mixing .
there are two differences in the lo phonon mode shifts in these experiments compared to the pre - mixed experiments , however .
first , the lo phonon mode depends more steeply on co concentration in the ice with a linear fit of 54 @xmath27 6 @xmath5 n@xmath28/n@xmath29 compared to the pre - mixed case which gives a fit of 36 @xmath27 7 @xmath5 n@xmath28/n@xmath29 .
second , the photolyzed phonon mode is split into the redshifted ( co : co@xmath0 mixture ) component and a second component that remains at the pure co@xmath0 lo phonon frequency .
this suggests that some pure co@xmath0 is present throughout the photolysis experiment . to test whether this may be due to limited ly-@xmath34 penetration depth through the ice
, we also photolysed a @xmath1910 nm think co@xmath0 ice while monitoring the lo phonon mode ; the penetration depth of 10.2 ev photons through co@xmath0 ice has been previously measured in our laboratory as 50 nm @xcite .
figure [ fig : photothin ] shows that photolysis of the thin ice also results in a split lo phonon mode , with one peak that can be ascribed to pure co@xmath0 and another to a co : co@xmath0 ice mixture , demonstrating that the second peak can not be due to a limited photon penetration depth .
furthermore , the absorbance at the pure co@xmath0 phonon mode frequency ( 2381 @xmath5 ) plateaus after @xmath1920 minutes of irradiation , while the production of co and destruction of co@xmath0 remains linear through the irradiation period .
the plateau at the pure co@xmath0 lo frequency thus indicates that an environment exists in the ice in which the co@xmath0 is not mixed with the photoproducts and this hypothesis is further discussed in section [ sec : effects ] .
the lo - to splitting depends on long range dipole interactions between co@xmath0 molecules and therefore is perturbed by the addition of other species into the ice lattice .
the magnitude of the lo - to splitting depends on the ice physical structure and bonding , the optical properties of the ice and the transition dipole moment of the co@xmath0 normal modes .
thus , the lo redshift upon addition of co , h@xmath0o and o@xmath0 into the co@xmath0 ice lattice is a combination of how each of these species changes structure , binding energies , dielectric constant and transition moments of the co@xmath0 molecules within the lattice . according to equation [ eq:1 ]
, the transition dipole moment governs the lo - to splitting in polycrystalline films .
the co@xmath0 asymmetric stretch has a large transition dipole matrix element , consistent with the large lo - to splitting in pure co@xmath0 ice .
one might expect that differences in the redshifts of the co@xmath0 lo frequency correspond to differences in the effects of the transition dipole elements of co , h@xmath0o and o@xmath0 on that of the co@xmath0 lo mode .
co , while exhibiting only a small dipole moment , has a very large transition dipole moment that could perturb the effective co@xmath0 transition element . while h@xmath0o has a large transition dipole ,
o@xmath0 has @xmath4=0 for all internuclear separations and therefore its transition dipole moment is zero everywhere . if the changes in the co@xmath0 transition dipole moment were the only explanation for the lo redshift we would expect that o@xmath0 would not shift the lo frequency and that co and h@xmath0o would have similar lo redshifts , which is not the case in our measurements .
we therefore consider other effects on the lo - to splitting .
the addition of co , o@xmath0 and h@xmath0o affects the co@xmath0 ice physical structure .
changes in the crystal structure can shift the lo mode to lower frequencies as shown for different co@xmath0 ice nanostructures @xcite .
our data suggest that co has the greatest effect on the longitudinal phonon frequency in co@xmath0 ice . despite the substantially lower binding energy to co@xmath0 compared to h@xmath0o
, co may produce the observed large shift because it is able fit well into the co@xmath0 ice lattice so that each co@xmath0 molecules in the lattice has dipole interactions with more co molecules than in the case of h@xmath0o ; however , detailed structural simulations are required to constrain the exact mechanism .
splitting of the lo phonon into more than one band indicates that the mixing is not homogeneous and distinct co@xmath0 bulk environments are present in the ice .
this may be caused by co@xmath0 segregation away from the defect ice component .
figure [ fig : spectra ] shows splitting of the lo mode in the co@xmath0:o@xmath0 ice mixtures and in the co@xmath0:co mixtures with high co abundance . in the h@xmath0o mixtures ,
the lo mode at the pure co@xmath0 frequency is absent , this suggests that the co@xmath0 is mixed in the h@xmath0o ice lattice such that pockets of pure co@xmath0 do not occur .
this is expected as co@xmath0 diffusion through h@xmath0o is slow at 28 k @xcite while diffusion and segregation of co and o@xmath0 may be possible , though further experiments are needed to support this hypothesis .
we see the clearest splitting of the lo mode in the photolysis of co@xmath0 ice , where one component stays at the pure co@xmath0 lo position .
as demonstrated by the thin ice photolysis experiment , this is not due to shielding of the uv radiation for the deepest ice layers .
instead , pockets of pure co@xmath0 must either be maintained or produced during the photolysis .
only @xmath1920@xmath23 of the co@xmath0 ice is lost during the photolysis period .
it is possible that some of the leftover co@xmath0 is not mixed with the photoproducts and maintained as pure co@xmath0 ice .
pure co@xmath0 could also be produced during the photolysis by several mechanisms .
excited co@xmath0 that is not photolysed carries excess energy that may allow it to segregate away from the photoproducts to form pure co@xmath0 pockets .
co@xmath0 is also reformed during the photolysis by co + o recombination and co@xmath35 dissociation .
another possible explanation is the desorption of photoproducts that are formed in the surface layers , leaving behind pure co@xmath0 .
this hypothesis could be investigated further by studying the molecules desorbing from the ice during the photolysis , by using isotopically labeled layers or by capping the co@xmath0 ice with an inert gas matrix .
the sensitivity of the co@xmath0 lo phonon mode to ice mixing may be utilized in laboratory studies to provide information on bulk diffusion , physical mixing properties and ice phase properties such as crystallization .
we have shown that the lo mode is particularly sensitive to the addition of ice defect molecules added either by deposition of gas mixtures or _ in situ _ by photolysis of the ice . mixing
may also occur through diffusion of molecules in the ice .
diffusion is proposed to underpin much of the chemistry in interstellar ices , but diffusion barriers and mechanisms are highly uncertain @xcite .
bulk diffusion may also regulate the morphology of interstellar ices .
diffusion of co out of co@xmath0 ice may cause co distillation from co - rich ices protostellar regions , thus explaining the presence of pure co@xmath0 in some such lines of sight @xcite .
we have conducted preliminary experiments to check whether the co@xmath0 lo mode could trace diffusion and desorption processes .
the co@xmath0 ice mixtures were slowly heated passed the desorption temperature of co or o@xmath0 and the changes in lo mode were monitored .
as the defect molecule desorbs , the lo phonon mode moves back toward the pure co@xmath0 frequency .
the process is not completely reversible as some defect molecules remain trapped and desorb with the co@xmath0 ice around 75 k. this suggests that phonon mode spectroscopy could be used to characterize diffusion in co@xmath0 ices .
furthermore , phonon modes are not unique to co@xmath0 ice , and optical phonons in other ices may be used to measure diffusion coefficients for a range of astronomically relevant species . phonon mode monitoring may also be used to characterize ice crystallization .
the lyddane - sachs - teller relationship ( equation [ eq:2 ] ) shows that the lo phonon mode should shift abruptly upon a phase change in the ice due to a rapid change in the static dielectric constant .
optical phonons in ices may therefore provide a useful measure of crystallization kinetics .
the observed lo - to splitting for ice deposited at different temperatures may be used to accurately deduce phase change temperatures and the corresponding index of refraction for ices around the phase change boundary .
longitudinal optical phonon modes may contribute to interstellar ice spectra for several important ice constituents if light from the background source is sufficiently polarized . in laboratory reflection absorption
spectra p - polarized light , parallel to the surface normal , can interact with longitudinal phonons in the ice . in astronomical environments
longitudinal phonons can be excited by p - polarized light or by the component of the unpolarized light wave that is parallel to the surface normal . because the interstellar ices are not confined to slabs , as in the case of laboratory experiments , the fraction of p - polarization and s - polarization is generally equal for randomly orientated dust particles .
unpolarized light will predominantly produce the to mode , though it can also excite lo phonons when the incoming ir wave is at oblique angles to the ice surface as seen in our laboratory spectra ( figure [ fig : anglesco2 ] ) .
the best sources for observing longitudinal phonons in ices are therefore those with the highest linear polarization fraction .
linear polarization arises from absorption and scattering from irregular dust grains which will absorb light from the background star preferentially along one axis .
dust grains can align with a local magnetic field and produce elliptically polarized light with linear and circular components @xcite .
the degree of linear polarization in star forming regions is often very large ranging from 20 - 70% and can extend over regions as large as 1 pc @xcite .
a good source to characterize interstellar co@xmath0 lo phonon modes would be , for example iras 05329 - 0505 , for which @xcite measure the linear polarization fraction as 70% in the k@xmath36 band .
optical phonon modes have been previously identified in laboratory spectra of h@xmath0o and co @xcite .
water ice phonons are considered carriers of far infrared bands at 62 @xmath4 m ( longitudinal acoustic ) and 44 @xmath4 m ( transverse optical ) bands @xcite , which can be observed with far - ir telescopes such as sofia .
these bands are sensitive to the ice structure with the 62 @xmath4 m appearing at the crystallization phase transition @xcite .
spectra of co bearing ices in 39 ysos observed using the vlt have been fit using the optical phonon modes of @xmath34-co @xcite .
here , the lo mode was able to accurately reproduce the blue wing of the co band for all of the observed sources .
lo phonon modes thus needs to be taken into account when identifying the carriers of observed ice spectral features . in the case of co@xmath0 ,
identifying and characterizing the lo phonon mode in observed spectra could also provide important information about the ice composition , morphology and thermal history . because the lo phonon mode is sensitive to defects in the co@xmath0 ice lattice
, its frequency can be used to provide information about abundance of other species mixed with solid co@xmath0 . in pristine co@xmath0 ice
the lo mode will appear at 4.2 @xmath4 m , if light from the background source is sufficiently polarized . for co@xmath0 ice in apolar and polar phases
the lo phonon mode would be present at a redshifted frequency that is dependent on the fractional abundance of other ice components .
a lack of lo - to splitting in all ice species provides constraints on the source ( polarization and dust properties ) , while a lack of lo - to splitting in co@xmath0 when present in co would constrain the co@xmath0 ice to be dilute , in which case the lo mode is convoluted with the to mode . to fit observational spectra , the co@xmath0 lo phonon spectra reported here need to be adjusted to account for grain size and shape effects .
the standard method for treating grain shape and size variations has been to use a continuous distribution of ellipsoids ( cde ) ( e.g. @xcite ) . @xcite
have shown that the size and shape of grains affect the co@xmath0 absorption profile causing typical shifts and broadenings of 5 @xmath5 that can dominate over variations due to ice matrix effects .
an in depth analysis of grain geometric effects on the co@xmath0 lo phonon mode will be needed to before it can be used to constrain ice mixing in observational spectra .
analysis of lo - to splitting should be especially fruitful on future ice spectra acquired with the jwst , since the analysis requires both high sensitivity and spectral resolution .
we have shown that longitudinal optical phonons in co@xmath0 ices are particularly sensitive to the presence of other ice species in the lattice .
the frequency of the co@xmath0 lo mode redshifts linearly as the concentration of the ice mixing molecule is increased .
similarly , photolysis of co@xmath0 ices results in a linear redshift of the lo phonon mode with formation of the major photoproduct , co. these spectral signatures will be useful for constraining ice mixing both in laboratory experiments and in observations of astronomical ices .
+ i. r. c. acknowledges full fellowship support from the department of chemistry , university of virginia .
i. r. c. would like to acknowledge instrument design and machining by matthew reish and helpful conversations with rob garrod , eric herbst , patrick king and jennifer bergner . i. r. c would like to acknowledge jeffrey shabanowitz for providing equipment used in producing these data .
this work was conducted in the laboratory of the late john t. yates jr .
to whom we would like to thank for his experimental techniques , motivation and mentorship leading to this study . | co@xmath0 ice is an important reservoir of carbon and oxygen in star and planet forming regions .
together with water and co , co@xmath0 sets the physical and chemical characteristics of interstellar icy grain mantles , including desorption and diffusion energies for other ice constituents .
a detailed understanding of co@xmath0 ice spectroscopy is a prerequisite to characterize co@xmath0 interactions with other volatiles both in interstellar ices and in laboratory experiments of interstellar ice analogs .
we report laboratory spectra of the co@xmath0 longitudinal optical ( lo ) phonon mode in pure co@xmath0 ice and in co@xmath0 ice mixtures with h@xmath0o , co , o@xmath0 components .
we show that the lo phonon mode position is sensitive to the mixing ratio of various ice components of astronomical interest . in the era of jwst
, this characteristic could be used to constrain interstellar ice compositions and morphologies . more immediately ,
lo phonon mode spectroscopy provides a sensitive probe of ice mixing in the laboratory and should thus enable diffusion measurements with higher precision than has been previously possible . |
the author thanks c. blume and m. gazdzicki for fruitful and stimulating discussions . this work was supported by gsi , dfg and bmbf .
this work used computational resources provided by the center for scientific computing at frankfurt ( csc ) .
10 s. v. afanasiev _ et al . _ [ the na49 collaboration ] , phys .
c * 66 * ( 2002 ) 054902 [ arxiv : nucl - ex/0205002 ] .
m. gazdzicki _ et al . _
[ na49 collaboration ] , j. phys .
g * 30 * ( 2004 ) s701 [ arxiv : nucl - ex/0403023 ] . c. blume , j. phys .
g : nucl . part .
31 , s57 ( 2005 ) c. roland [ na49 collaboration ] , j. phys .
g * 31 * ( 2005 ) s1075 .
p. koch , b. muller and j. rafelski , phys .
* 142 * ( 1986 ) 167 .
m. bleicher , s. jeon and v. koch , phys .
c * 62 * ( 2000 ) 061902 [ arxiv : hep - ph/0006201 ] .
e. v. shuryak and m. a. stephanov , phys .
c * 63 * ( 2001 ) 064903 [ arxiv : hep - ph/0010100 ] .
h. heiselberg and a. d. jackson , phys .
c * 63 * ( 2001 ) 064904 [ arxiv : nucl - th/0006021 ] .
b. muller , nucl .
a * 702 * ( 2002 ) 281 [ arxiv : nucl - th/0111008 ] .
m. gazdzicki , m. i. gorenstein and s. mrowczynski , phys .
b * 585 * , 115 ( 2004 ) [ arxiv : hep - ph/0304052 ] . m. i. gorenstein , m. gazdzicki and o. s. zozulya , phys .
b * 585 * , 237 ( 2004 ) [ arxiv : hep - ph/0309142 ] .
m. gazdzicki and m. i. gorenstein , acta phys . polon .
b * 30 * , 2705 ( 1999 ) .
m. i. gorenstein , m. gazdzicki and k. a. bugaev , phys .
b * 567 * , 175 ( 2003 ) [ arxiv : hep - ph/0303041 ] .
y. hama , f. grassi , o. socolowski , t. kodama , m. gazdzicki and m. gorenstein , acta phys .
b * 35 * ( 2004 ) 179 .
h. weber , c. ernst , m. bleicher _
et al . _ ,
phys . lett .
b * 442 * , 443 ( 1998 ) . c. m. hung and e. v. shuryak , phys .
lett . * 75 * , 4003 ( 1995 ) [ arxiv : hep - ph/9412360 ] .
d. h. rischke , y. pursun , j. a. maruhn , h. stoecker and w. greiner , heavy ion phys . * 1 * , 309 ( 1995 ) [ arxiv : nucl - th/9505014 ] .
j. brachmann , a. dumitru , h. stoecker and w. greiner , eur .
j. a * 8 * , 549 ( 2000 ) [ arxiv : nucl - th/9912014 ] .
r. hagedorn , nuov .
suppl . 3 , 147 ( 1965 )
; j. rafelski and r. hagedorn , bielefeld symp . , ed .
h. satz , pp . 253 | we analyze longitudinal pion spectra from @xmath0 gev to @xmath1 gev within landau s hydrodynamical model . from the measured data on the widths of the pion rapidity spectra
, we extract the sound velocity @xmath2 in the early stage of the reactions .
it is found that the sound velocity has a local minimum ( indicating a softest point in the equation of state , eos ) at @xmath3 gev .
this softening of the eos is compatible with the assumption of the formation of a mixed phase at the onset of deconfinement . over the last years
, a wealth of detailed data in the @xmath4 gev energy regime has become available .
the systematic study of these data revealed surprising ( non - monotonous ) structures in various observables around @xmath5 gev beam energy .
most notable irregular structures in that energy regime include , * the sharp maximum in the k@xmath6 ratio @xcite , * a step in the transverse momentum excitation function ( as seen through @xmath7 ) @xcite , * an apparent change in the pion per participant ratio @xcite and * increased ratio fluctuations ( due to missing data at low energies it is unknown if this is a local maximum or an ongoing increase of the fluctuations ) @xcite
. it has been speculated , that these observation hint towards the onset of deconfinement already at @xmath5 gev beam energy .
indeed , increased strangeness production @xcite and enhanced fluctuations have long been predicted as a sign of qgp formation @xcite within different frameworks and observables .
the suggestion of an enhanced strangeness to entropy ratio ( @xmath8 ) as indicator for the onset of qgp formation was especially advocated in @xcite .
also the high and approximately constant @xmath9 inverse slopes of the @xmath10 spectra above @xmath11 gev - the step - was also found to be consistent with the assumption of a parton @xmath12 hadron phase transition at low sps energies @xcite . surprisingly , transport simulations ( supplemented by recent lattice qcd ( lqcd ) calculations ) have also suggested that partonic degrees of freedom might already lead to visible effects at @xmath11 gev @xcite .
finally , the comparison of the thermodynamic parameters @xmath13 and @xmath14 extracted from the transport models in the central overlap region @xcite with the experimental systematics on chemical freeze - out configurations @xcite in the @xmath15 plane do also suggest that a first glimpse on a deconfined state might be possible around @xmath16 gev . in this letter
, we explore whether similar irregularities are also present in the excitation function of longitudinal observables , namely rapidity distributions . here
we will employ landau s hydrodynamical model @xcite .
this model entered the focus again after the most remarkable observation that the rapidity distributions at all investigated energies can be well described by a single gaussian at each energy .
the energy dependence of the width can also be reasonably described by the same model . for recent applications of landau s model to relativistic hadron - hadron and nucleus - nucleus interactions
the reader is referred to @xcite ( and refs .
therein ) .
the main physics assumptions of landau s picture are as follows : the collision of two lorentz - contracted nuclei leads to full thermalization in a volume of size @xmath17 .
this justifies the use of thermodynamics and establishes the system size and energy dependence .
usually , a simple equation of state @xmath18 with @xmath19 ( @xmath20 denotes the speed of sound ) is assumed . for simplicity ,
chemical potentials are not taken into account . from these assumptions
follows a universal formula for the distribution of the produced entropy , determined mainly by the initial lorentz contraction and gaussian rapidity spectrum for newly produced particles . under the condition
that @xmath20 is independent of temperature , the rapidity density is given by @xcite : @xmath21 with @xmath22 where @xmath23 is a normalisation factor and @xmath24 is the proton mass .
the model relates the observed particle multiplicity and distribution in a simple and direct way to the parameters of the qcd matter under consideration .
let us now analyze the available experimental data on rapidity distributions of negatively charged pions in terms of the landau model .
fig .
[ rapwidth ] shows the measured root mean square @xmath25 of the rapidity distribution of negatively charged pions in central pb+pb ( au+au ) reactions as a function of the beam rapidity .
the dotted line indicates the landau model predictions with the commonly used constant sound velocity @xmath19 .
the full line shows a linear fit through the data points , while the data points @xcite are depicted by full symbols .
at a first glance the energy dependence looks structureless .
the data seem to follow a linear dependence on the beam rapidity @xmath26 without any irregularities .
however , the general trend of the rapidity widths is also well reproduced by landau s model with an equation of state with a fixed speed of sound .
nevertheless , there seem to be systematic deviations . at low ags energies and at rhic , the experimental points are generally underpredicted by eq .
( [ eq2 ] ) , while in the sps energy regime landau s model overpredicts the widths of the rapidity distributions .
exactly these deviations from the simple landau picture do allow to gain information on the equation of state of the matter produced in the early stage of the reaction . by inverting eq .
( [ eq2 ] ) we can express the speed of sound @xmath2 in the medium as a function of the measured width of the rapidity distribution : @xmath27 ^ 2 + 1}\quad .
\label{eq3}\ ] ] let us now investigate the energy dependence of the sound velocities extracted from the data .
figure [ cs2 ] shows the speed of sound as a function of beam energy for central pb+pb ( au+au ) reactions as obtained from the data using eq .
( [ eq3 ] ) . the sound velocities exhibit a clear minimum ( usually called the softest point ) around a beam energy of @xmath5 gev .
a localized softening of the equation of state is a long predicted signal for the mixed phase at the transition energy from hadronic to partonic matter @xcite .
therefore , we conclude that the measured data on the rapidity widths of negatively charged pions are indeed compatible with the assumption of the onset of deconfinement at the lower sps energy range .
however , presently we can not rule out that also an increased resonance contribution may be the cause of the softening @xcite . in conclusion
, we have explored the excitation functions of the rapidity widths of negatively charged pions in pb+pb ( au+au ) collisions . *
the rapidity spectra of pions produced in central nucleus - nucleus reactions at all investigated energies can be well described by single gaussians .
* the energy dependence of the width of the pion rapidity distribution follows the prediction of landau s hydrodynamical model if a variation of the sound velocity is taken into account .
* the speed of sound excitation function extracted from the data has a pronounced minimum ( softest point ) at @xmath3 gev .
* this softest point might be due to the formation of a mixed phase indicating the onset of deconfinement at this energy .
further explorations of this energy domain is needed and can be done at the future fair facility and by cern - sps and bnl - rhic experiments . |
parity violation ( pv ) was discovered in 1956 in nuclear beta - decay by c.s .
wu @xcite , following a suggestion of t.d .
lee and c. n. yang @xcite . in 1960 ,
zeldovich @xcite pointed out that pv should lead to parity - odd ( p - odd ) terms also in electron - hadron interactions .
these are now considered as a manifestation of the electroweak interaction , whose properties are dictated by the standard model ( sm ) .
several p - odd observables have since been studied , in two types of pv experiments , namely in atomic physics @xcite ( at very low energy and momentum transfer ) and in electron scattering ( at relatively high energies and non - zero momentum transfers ) . at first
, these experiments were aiming at testing the sm and measuring the weinberg angle .
a pioneering experiment was performed at slac on a deuterium target @xcite , followed 10 years later by experiments at mainz on @xmath21be @xcite and bates on @xmath22c @xcite .
their determination of the weinberg angle were confirmed later on , within their stated accuracy of 10% , by high energy experiments . since @xmath23 is now known to three decimal places [ @xmath24 @xcite , the emphasis of e - p scattering today , is to make use of the sm to learn about the internal structure of the nucleon . until recently
, it has been assumed that the nucleon was only made of u and d valence or sea quarks , but there are indications that the nucleon carries also hidden strangeness : * the sigma - term ( deduced from the pion - nucleon scattering length ) is very different from the theoretical value calculated within the chiral perturbation theory ( which is a realization of the sm at low energy ) , indicating that @xmath25 of the nucleon mass might be carried out by strange quarks @xcite , * experiments of polarized deep - inelastic - scattering ( dis ) of leptons show that up to 10 - 20@xmath26 of the nucleon spin could be carried by strange quarks @xcite , * elastic scattering of neutrinos and anti - neutrinos by protons can only be explained by taking into account strange quarks in the nucleon @xcite , * a natural explanation of the strong violation of the ozi - rule in @xmath27 annihilation @xcite and of @xmath28- production @xcite or @xmath29-meson production @xcite in nucleon - nucleon interactions takes into account a nucleon ( antinucleon ) strange sea .
these experiments are sensitive to various aspects of nucleon structure : for example , the sigma - term and the results of @xmath30 or @xmath31 experiments are sensitive to the scalar part of the hadronic current , polarized dis and elastic scattering of neutrinos ( or anti - neutrinos ) by protons are sensitive to the vector - axial current . in this respect , pv in electron - nucleon scattering seems the most attractive way of measuring the _ strange _ vector current , thanks to a clean theoretical interpretation through the sm . the sample collaboration at mit - bates , has measured pv asymmetries in @xmath32 elastic scattering at @xmath2= 0.1 @xmath33 ( @xmath34 , where @xmath35 is the four - momentum transfer squared ) and large angle @xcite , which allowed them to obtain the first experimental determination of the weak magnetic form - factor of the proton . from this measurement and the knowledge of the proton and neutron electromagnetic form - factors ,
one could extract a strange magnetic form - factor = @xmath36 .
note that most calculations based on qcd or quark models predict negative values for ( see e.g. @xcite ) .
another experiment , done by the happex collaboration at jefferson lab,@xcite has done a measurement at 2= 0.48 @xmath33 and small scattering angle @xmath37 where the sensitivity to the weak electric form factor is enhanced . here
the measured asymmetry @xmath38 is consistent with the sm prediction in the absence of components in the nucleon sea . from this asymmetry
, one can deduce the following contribution to the strange form - factor : @xmath39 compatible with zero within the error bars .
these results have stimulated a strong interest and many predictions , for both and , have been published , whether within quark models @xcite , chiral perturbation theories @xcite or lattice qcd calculations @xcite .
these calculations predict that while is essentially constant as a function of 2 , may vary rapidly .
they also indicate that there might be some cancellation between and which are predicted of different signs .
therefore new e - p experiments are being set up in order to check these predictions : at @xmath40 ( gev / c)@xmath41 at mainz@xcite , at @xmath2 = 0.1 ( gev / c)@xmath41 and forward angles by the happex collaboration in order to do a rosenbluth separation of and in combination with the sample results , and finally a full separation of and in the momentum transfer range @xmath2 = 0.12 - 1.0 ( gev / c)@xmath41 is foreseen by the 0 collaboration at jefferson lab @xcite . it should be stressed that , due to their present high level of precision and their firm theoretical basis , pv experiments can also be used to answer other important physics questions , besides the existence of a possible component in the nucleon .
let s mention for example : @xmath42 search and test for new physics beyond the sm @xcite .
such effects could manifest themselves through , e.g. extra @xmath43-bosons ( heavier than the standard 0 ) or leptoquarks indicating that there are substructures common to both leptons and quarks .
the use of the 0 experimental set - up for such studies is being discussed @xcite , @xmath42 study of the axial part of the hadronic weak neutral current : it has been shown that , for the special case of a @xmath44 , @xmath45 spin - isospin transition ( @xmath9-excitation ) , there is a particular sensitivity to , the axial - vector transition form - factor which could be determined independently of pcac and free of uncertainties from extrapolation of low energy theorems @xcite , @xmath42 measurement of the neutron charge form - factor @xcite .
the reactions @xmath46 and @xmath47 are of practical interest for experimentalists as they may contaminate the elastic peak .
it is therefore important to determine their own asymmetries since , if they are much larger than or , even , of different sign from the elastic one , they might be a source of errors or large uncertainties . in 3-body reactions
, besides the weak pv asymmetries , there are also strong ( parity - conserving ) interactions , due to the so - called @xmath48 response function @xcite , which are generally much larger ( of the order of @xmath49 instead of @xmath50 ) than pv asymmetries but which cancel in inclusive reactions or when detectors have an azimuthal symmetry . pion production has been studied previously @xcite in quasi 2-body models with stable @xmath9 isobar .
a more complete calculation including background ( born ) terms with pseudoscalar @xmath4 coupling with the @xmath9 treated as a rarita - schwinger field with phenomenological @xmath4 electromagnetic transition currents can be found in @xcite . in the present study , we calculate pv asymmetries in inclusive @xmath51 electroproduction , starting from threshold up to the @xmath9-region in an approach similar to the one of ref .
@xcite , but differing in the following aspects : * the main improvement consists in including @xmath12- and @xmath11-exchange in the @xmath52-channel for @xmath53 ( where @xmath54 is a virtual photon or boson ) , * we use a different parametrization for the @xmath9 contribution , and slightly different values of mass and width , * crossing symmetries are treated differently ( and less accurately ) than in ref @xcite , * we use a pseudoscalar @xmath55 interaction in order to identify possible off - mass - shell effects .
two remarkable results were found in the calculations of ref.@xcite , for which no explanation or discussion was given : * * the full ( background + resonance ) asymmetry a does not depend on the total energy of the hadronic system , although the separate terms show strong ( and opposite ) variations , * * @xmath56 is , to a good approximation , a linear function of 2 .
the present work agrees with these features .
moreover it gives a physical explanation to them as it shows that there is a specific parametrization of the asymmetry which allows to separate the main ( isovector ) contribution in a model independent way .
this contribution only depends on the fermi constant @xmath57 , the fine structure constant @xmath58 and @xmath23
. therefore its dependence on the kinematics can be predicted exactly .
small corrections to the main contribution , due to the isoscalar part of the neutral vector and axial currents can then be calculated and their physical importance assessed .
we shall consider here the processes @xmath60 , where @xmath61 is a nucleon ( @xmath62 or @xmath63 ) and @xmath64 is a pion ( @xmath0 or @xmath1 ) .
we take into account two standard mechanisms , @xmath65 and @xmath66 boson exchanges ( fig .
1 ) , predicted by the sm .
the matrix element can be written in the following form : @xmath67 @xmath68 @xmath69 where @xmath57 is the fermi constant of the weak interaction , @xmath70 is the electromagnetic current for @xmath71 , @xmath72 and @xmath73 are the vector and vector - axial parts of the neutral weak current for @xmath74 .
the four - vector @xmath75 and @xmath76 are the vector and vector - axial parts of the neutral weak current of a point - like electron : @xmath77 @xmath78 where @xmath79 is the four - momentum of the initial ( final ) electron .
the notation for the particle four momenta is explained in fig .
1 . note that the formula for @xmath80 , eq .
( [ eq : eq1 ] ) , is valid in the so - called local limit , where @xmath81 in the standard model the constants @xmath82 and @xmath83 are determined by the following expressions : @xmath84 where @xmath85 is the weinberg angle .
then : @xmath86 from eq .
( [ eq : eq1 ] ) it appears that p - odd effects in @xmath87 result from the interference between @xmath88 and @xmath89 , which is the parity - violating part of @xmath80 : @xmath90 the p - odd asymmetry in the scattering of longitudinally polarized electrons can be written as : @xmath91 with two different contributions to @xmath92 : @xmath93 where @xmath94 is proportional to @xmath95 : @xmath96 @xmath97 @xmath98 and the overline in eq .
( [ eq : eq9 ] ) stands for the sum over the final nucleon polarizations and the average over the polarizations of the initial nucleon in the process @xmath71 .
the quantities @xmath99 and @xmath100 in eq .
( [ eq : eq7 ] ) characterize the interference of the electromagnetic hadronic current @xmath101 with the vector and axial parts of the weak neutral current : @xmath102 @xmath103 @xmath104 @xmath105 and @xmath106 where @xmath107 is the usual antisymmetric tensor . in the product of the tensors @xmath108 and @xmath109 ,
only the antisymmetric part of @xmath109 contributes , whereas the quantity @xmath99 , eq . ( [ eq : eq11 ] ) , is determined by the symmetrical part of the tensor @xmath110 .
according to eq .
( [ eq : eq4 ] ) , we can neglect the @xmath100 contribution ( the second p - odd contribution , which is induced by the axial part of the neutral weak current , is more model dependent and it will be the object of a detailed analysis in a subsequent paper ) . in this approximation ,
the p - odd asymmetry is solely determined by the vector part of the hadronic neutral weak current : @xmath111 in order to calculate the ratio @xmath112 , we shall use the isotopic structure of the vector neutral current , which holds in the sm when neglecting the contributions of the isoscalar quarks ( @xmath113 , @xmath114 , ... ) : @xmath115 where @xmath116 and @xmath117 and @xmath118 are the isoscalar and isovector components of the electromagnetic hadronic current . considering the specific isotopic structure of @xmath72 , eq .
( [ eq : eq17 ] ) , the asymmetry @xmath119 for any process @xmath120 can be written as : @xmath121 , \label{eq : eq19}\ ] ] where the quantity @xmath122 results from the interference of the isoscalar component @xmath117 of the electromagnetic current with the full electromagnetic current in @xmath70 i.e. : @xmath123 one can see from eq .
( [ eq : eq19 ] ) that the isovector part of the electromagnetic current induces a definite contribution to the p - odd asymmetry @xmath119 , which is model independent and can be predicted in terms of the fundamental constants @xmath57 , @xmath58 and @xmath124 .
note that this contribution depends only on the variable @xmath125 .
therefore , for reactions such as @xmath126 , @xmath127 , where the electromagnetic current is pure isovector ( and therefore @xmath128 ) , the asymmetry can be predicted exactly : @xmath129 , \label{eq : eq21}\ ] ] in agreement with ref.@xcite and neglecting the small contributions from the axial hadronic current , which is not considered here ( note that @xmath130 ) . in particular , for the reaction @xmath131 this model - independent estimate of @xmath119 together with the possibility of a precise measurement of the p - odd asymmetry , open new ways to look for new physics @xcite and to study effects due to the axial current . in the next section
, we will show that the quantity @xmath122 , in the near - threshold region for @xmath87 , as well as in the region of the @xmath9 excitation , can be considered as a small correction to the main isovector contribution .
therefore , the uncertainty in the estimate of @xmath122 will affect very little the results .
note that there is a model independent relation between the isoscalar components of the electromagnetic currents , for the considered processes @xmath132 and @xmath133 , which holds for any interaction mechanism : @xmath134 > from eq .
( [ eq : eq19 ] ) it appears that the inclusive asymmetry @xmath119 depends on the variables @xmath135 and @xmath3 only through the correction @xmath122 : @xmath136 taking into account the longitudinal and transversal polarizations of the virtual @xmath137 and @xmath138-boson , the following representation for the correction @xmath122 can be written ( in case of a single channel : @xmath46 or @xmath139 ) : @xmath140 @xmath141 where @xmath142 and @xmath143 are the total cross sections of virtual photon absorption in @xmath71 : @xmath144 @xmath145 @xmath146 being the element of solid angle of the produced pion ( in the cms of the process @xmath71 ) .
we use here a coordinate system in which the z - axis is along the three momentum of the virtual photon , and @xmath147 , @xmath148 and @xmath149 are the space components of the hadronic electromagnetic current .
the interference contributions @xmath150 and @xmath151 are defined as follows : @xmath152 @xmath153 , \label{eq : eq24}\ ] ] where @xmath154 are the space components of the isoscalar part of the hadronic electromagnetic current .
the lines above the products of the components of the electromagnetic currents mean the sum over the polarizations of the final nucleons and the average over the polarizations of the initial nucleons .
the inclusive asymmetry for @xmath155 with the contribution of two channels @xmath156 and @xmath157 in the final state , is determined by the following expressions : @xmath158,\ ] ] @xmath159 with @xmath160 therefore , the p - odd inclusive asymmetry @xmath119 for @xmath161 is determined by a set of four total cross sections : @xmath162 for each @xmath132 and @xmath163 processes ( 8 in total ) , as functions of two independent kinematical variables @xmath125 and @xmath3 .
the polarization parameter @xmath164 , @xmath165 , which represents the linear polarization of the virtual photon , contains the dependence on the kinematical conditions of the electrons in the initial and final states ( i.e. initial energy and scattering angle ) . in the present calculation we shall use the following parametrization of the spin structure of the matrix element for @xmath71 , in terms of six standard contributions : @xmath166 @xmath167 @xmath168 where @xmath169 and @xmath170 are the two - component spinors of the initial and final nucleons ,
@xmath171 is the three - vector of the virtual photon polarization , @xmath172 and @xmath173 are the unit vectors along the 3-momentum of the @xmath7 and @xmath64 in the cms of the @xmath174- reaction .
the complex scalar amplitudes @xmath175 , which are functions of three independent kinematical variables , @xmath176 , can be related to the usual set of amplitudes , @xmath177 , when the operator @xmath178 is used instead of @xmath179 : @xmath180 @xmath181 @xmath182 @xmath183 the results of averaging over the polarization states of the initial nucleon and summing over the polarizations of the produced nucleon gives : @xmath184 @xmath185 @xmath186 @xmath187 to calculate the cross sections @xmath151 and @xmath150 , in eqs .
( [ eq : eq27 ] ) the following substitutions are made : @xmath188 @xmath189 where @xmath190 are the scalar amplitudes , describing the isoscalar part of the hadronic electromagnetic current @xmath191 .
we use here the standard approach for the calculation of the electromagnetic current for the @xmath193 processes , which describes satisfactorily well the existing photo- and electro - production data , in the region of @xmath3 starting from threshold , @xmath194 , up to @xmath195 1.3 gev ( the @xmath9 excitation region ) .
this approach takes into account the following three contributions : * born terms in the @xmath113 , @xmath52 and @xmath196 channels , * vector meson ( @xmath12 and @xmath11 ) exchanges in the @xmath52-channel , * @xmath9-isobar excitation in the @xmath113 channel . using the isotopic structure of the strong vertices on the diagrams ( fig .
2 ) , the scalar amplitudes for each @xmath71 process can be written as : @xmath197 , \label{eq : eq29}\ ] ] where @xmath198 characterize the contributions of the different feynmann diagrams to the scalar amplitudes @xmath175 , @xmath199 . the energies @xmath135 and @xmath200 of the initial and final nucleons
are determined by the following formulae : @xmath201 the isotopic numerical coefficients @xmath202 for the two processes @xmath133 and @xmath132 are shown in table 1 . the scalar amplitudes for the isoscalar part of the electromagnetic current can be written as follows : @xmath203 where @xmath204 and @xmath205 are the amplitudes for the isoscalar part of the @xmath113- and @xmath196- channel born diagrams .
one can see now that , in the framework of the considered approach , the main contributions to @xmath206 have an isovector nature : * @xmath9-excitation in @xmath1 and @xmath0 production , * @xmath1-exchange for @xmath1 production , * @xmath12-exchange for @xmath0 production , * contact term for @xmath1 production ( in the case of a pseudovector @xmath55-interaction ) , * @xmath207 born contributions .
therefore , the isoscalar electromagnetic current can only contain the following contributions : * @xmath208exchange for @xmath0 and @xmath1-production , * the isoscalar part of the @xmath207-diagrams .
however these isoscalar contributions are small in comparison with the corresponding isovector ones . indeed , the @xmath11-exchange term is smaller than the @xmath12 -exchange term , due to the following reasons : * @xmath209 : suppression at electromagnetic vertices ; * @xmath210 : suppression at the strong vertex . in the same way , the isoscalar born contribution due to the nucleon magnetic moment , for example , is smaller than the isovector contribution : latexmath:[\[\displaystyle\frac{|\mu_p+\mu_n|}{|\mu_p-\mu_n|}=\displaystyle\frac{|1.79 - 1 .91 @xmath122 can be considered a small correction to the model - independant prediction of eq .
( [ eq : eq21 ] ) .
let us briefly discuss now the properties of the suggested model , for the @xmath212 processes . using a pseudoscalar @xmath55-interaction
, we can write the relativistic invariant expression for the matrix element of the @xmath213 reaction in the following form : @xmath214 @xmath215 @xmath216 @xmath217 where @xmath113 , @xmath52 , and @xmath196 are the standard mandelstam variables : @xmath218 @xmath35 , @xmath219 , @xmath220 and @xmath221 are the four - momenta of @xmath7 , @xmath64 , initial and final nucleons , @xmath222 is the four - vector of the virtual photon polarization , @xmath223 is the @xmath55 coupling constant ( for a pseudoscalar interaction ) , @xmath224 and @xmath225 ( @xmath226 and @xmath227 ) are the dirac and pauli electromagnetic form factors of the proton ( neutron ) .
the electromagnetic form factor of the nucleon is usually parametrized in form of a @xmath125-dependence of the electric ( @xmath228 ) and magnetic ( @xmath229 ) nucleonic form factors : @xmath230 @xmath231 a simple dipole dependence of @xmath232 , @xmath233 and @xmath234 : @xmath235 ^ 2},\ ] ] with @xmath236 has been considered a good parametrization of the existing experimental data , in a wide region of space - like momentum transfer while @xmath237 . however a very recent direct measurement @xcite of the ratio @xmath238 shows some deviation of @xmath232 from a dipole behavior , in the region @xmath239 ( gev / c)@xmath41 .
this high precision experiment is based on the measurement of the polarization of the final protons in @xmath240 , in the elastic scattering of longitudinally polarized electrons @xcite .
this effect should be taken into account in future calculations , as well as the fact that @xmath241 deviates from zero , at least in the region @xmath242 .
the last direct measurement of @xmath241 , in @xmath243 @xcite confirms some previous estimates of the neutron form factor based on the saclay @xmath244 elastic scattering data @xcite . in the vector dominance model ( vdm ) approach , the pion electromagnetic form factor @xmath245 is described by : @xmath246 where @xmath247 is the @xmath11-meson mass .
note that the electromagnetic current for the reaction @xmath248 , corresponding to the sum of the born diagrams in the @xmath113 and @xmath196-channels , is conserved for any form factors @xmath249 and @xmath250 in the whole kinematical region .
this is not the case for the reaction @xmath132 ( [ eq : eq26 ] ) , as one can show that the divergence of the corresponding electromagnetic current ( in the born approximation , for the sum of the @xmath113 , @xmath52 , and @xmath196-contributions ) is proportional to the following combination of the electromagnetic form factors : @xmath251 the simplest way to conserve the hadronic electromagnetic current , is to extend to all values of @xmath125 the following relation : @xmath252 which is in general valid only for @xmath253 .
however existing data on pion and nucleon form factors are in contradiction with relation ( [ eq : eq30 ] ) . a possible way to avoid
this difficulty is to renormalize the matrix element @xmath254 in the following way : @xmath255 the electromagnetic current , corresponding to the new born matrix element @xmath256 is conserved for any form factor .
note that such a procedure changes only @xmath257 , without any effect on the transversal cross - sections @xmath142 and @xmath258 . in our considerations
we shall use the procedure ( [ eq : eq31 ] ) , while relation ( [ eq : eq30 ] ) was taken in ref.@xcite .
the scalar amplitudes @xmath175 , corresponding to different diagrams of the born mechanism , are given in the appendix .
the matrix element @xmath259 , corresponding to vector meson exchange in the @xmath13channel can be written in the following form : @xmath260 @xmath261 u(p_1)\ ] ] where @xmath262 is the electromagnetic form factor for the @xmath263-vertex , @xmath264 is the vector meson mass , @xmath265 and @xmath266 are the `` strong '' form factors for the @xmath267 vertex ( with a virtual v - meson ) . in principle the `` static '' values of these form factors ( i.e. for @xmath268 ) , are related to the @xmath269 and @xmath270 coupling constants : @xmath271 an estimate for the @xmath269 coupling constants , based on the bonn potential @xcite , gives : @xmath272 the @xmath270 coupling constants can be estimated from pion photoproduction data @xcite : @xmath273 note that the constants @xmath274 can be identified in vdm , with the values of the isoscalar and isovector anomalous magnetic moment of the nucleon .
the vdm allows to write the following parametrization for the @xmath125 dependence of the electromagnetic form factor for the @xmath275 vertex : @xmath276 where @xmath264 is the mass of the @xmath11 or @xmath12 vector meson .
the @xmath277 coupling constant can be fixed by the width of the radiative decay @xmath278 , through the following formula : @xmath279 the numerical estimate , is based on the following values @xcite : @xmath280
@xmath281 @xmath282 @xmath283 @xmath284 the following relation holds for the hadronic electromagnetic current , : @xmath285 therefore any violation of this relation is an indication of the presence of an isotensor component of the electromagnetic current , which is absent , however , at the quark level .
so , a precise experiment with the simultaneous determination of the two coupling constants for @xmath286 and @xmath287 would be very important .
it would not only constitute a test of the isotopic properties of the hadronic electromagnetic current , but also have application in the calculation of the meson exchange current contributions ( mec ) to the deuteron electromagnetic form factors .
such contributions are considered to be very important at high momentum transfer .
note in this respect that the @xmath52-dependence of the form factors @xmath288 and @xmath289 is also important , for the correct calculation of mec , in case of elastic electron - deuteron scattering .
moreover , the relative sign of the @xmath290exchange and born contributions to the @xmath212 processes , is generally not known .
so , we shall consider here both relative signs .
the determination of this sign is also important for mec calculations , because in both cases ( @xmath291 and @xmath212 ) the set of constants and form factors are the same .
finally we stress that the electromagnetic current , corresponding to vector meson exchange in the processes @xmath292 is automatically conserved , independently of the parametrization of the strong form factors @xmath293 and @xmath294 and the electromagnetic form factor @xmath262 .
this contribution can be analyzed in a relativistic framework @xcite , considering a virtual @xmath9 as a rarita - schwinger field with spin 3/2 but in this approach it is difficult to treat off - shell effects .
first of all , this means that @xmath295exchange may contain contributions from a state with spin 1/2 as well as antibaryonic terms with negative p - parity and s=1/2 and 3/2 .
therefore the description of the @xmath295isobar , with @xmath296 , especially in the @xmath10channel is not straightforward . to avoid these complications , we choose here a direct parametrization of the @xmath297 contribution .
note that the cms for @xmath298 is the optimal frame , because the three - momentum of the @xmath9 is zero , so that the @xmath299 can be described by a two - component spinor , with a vector index , @xmath300 , which satisfies the following auxiliary condition : @xmath301 typical for a pure spin 3/2 state . using this condition , it is possible to find the following expression for the @xmath9-density matrix : @xmath302 with the normalization condition : @xmath303 . in this formalism
the @xmath304-vertex can be parametrized as follows : @xmath305 where @xmath306 is the 2-component spinor of the nucleon in the decay @xmath307 , @xmath308 is the unit vector along the pion three momentum , in the @xmath9 rest frame , and the constant @xmath309 characterizes the width of the strong decay @xmath310 .
taking into account the conservation of the total angular momentum and of the p - parity in the electromagnetic decay @xmath311 with production of m1 photons , the following expression can be written for the matrix element : @xmath312 where @xmath313 is the constant of the magnetic dipole radiation ( or the magnetic moment for the transition @xmath311 ) , @xmath171 and @xmath172 are the photon polarization three - vector and unit momentum vector along the three - momentum of @xmath137 , respectively . in the general case ,
the transition @xmath314 must be described by three different form factors , corresponding to the absorption of @xmath315 , @xmath316 ( transversal ) and @xmath317 ( longitudinal ) virtual photons .
but the existing experimental data about pion photo and electro - production of pions on the nucleons ( in the @xmath9 resonance region ) indicate that the @xmath315 term is dominant @xcite , therefore in our analysis we will consider only this form factor . using expressions ( [ eq : eq34 ] ) and ( [ eq : eq35 ] ) for both the @xmath318 and @xmath319 vertices
, one can write the matrix element of the @xmath295contribution in the @xmath10channel ( fig .
2e ) as follows : @xmath320 where @xmath321 and @xmath322 are the mass and width of @xmath9 and @xmath323 is proportional to the magnetic form factor of the @xmath314 transition . the following @xmath9 contributions to the scalar amplitudes , @xmath324 can be derived : @xmath325 @xmath326 @xmath327 @xmath328 where we use the notation : @xmath329 the role of the factor @xmath330 is to correctly describe the threshold behavior of the @xmath315 amplitude for @xmath331 , according to the p - wave nature of the produced pion .
we shall use the following formula for the @xmath125-dependence of the transition electromagnetic form factors : @xmath332 the factor @xmath333 , with @xmath334 gev@xmath41 , is included in order to take into account a steeper decreasing of @xmath323 in comparison with the dipole behavior of the elastic electromagnetic form factors of the nucleons @xcite .
+ the normalization constant @xmath335 can be found according to the following procedure .
( [ eq : eq36 ] ) , let us calculate first the differential cross section for @xmath0-photoproduction : @xmath336 at @xmath337 , where the @xmath9-excitation in the @xmath10channel is the main mechanism .
so , our parametrization of the @xmath9-contribution describes correctly the angular dependence @xmath338 , typical for the magnetic excitation of a @xmath339 state in @xmath340 .
therefore , the total cross section can be written as : @xmath341 where : @xmath342 @xmath343 we can approximate with a good accuracy @xmath344 by a single @xmath295resonance contribution . for a numerical estimate of
@xmath335 we use @xmath345 cm@xmath41 .
note again that this procedure can not determine the sign of @xmath335 .
however for the @xmath347 reaction , there is a strong interference between the pion diagram and the @xmath295contribution .
the comparison of the calculations using different signs with the experimental @xmath348-dependence in the resonance region allows to fix the corresponding relative sign .
two remarks should be done about this procedure : * there is no ambiguity concerning off - mass shell effects for the @xmath9-contribution , at least in the @xmath10channel , * this special contribution is gauge invariant .
we neglect in our consideration the @xmath9-exchange in the @xmath349channel . the main reason to include this contribution is to have the crossing symmetry of the model .
this is in principle an important property of the photoproduction amplitude , in particular in connection with the dispersion relation approach .
however in the framework of phenomenological approaches , this symmetry is strongly violated .
for example , the @xmath10channel @xmath295 contribution induces an amplitude which is mostly complex ( with a typical breit - wigner behavior ) , whereas the @xmath196-channel contribution results in a real amplitude .
the inclusion of different form factors for the @xmath10 and @xmath349channel violates the crossing symmetry , which is important for the born contributions .
this appears clearly for the reaction @xmath163 , because here the crossing symmetry is correlated with the gauge invariance of the electromagnetic interaction , and the violation of the crossing symmetry has for direct consequence the violation of the current conservation . on the contrary , for the @xmath9-contribution
, this important correlation is absent .
having determined all the parameters of the model , it is possible in principle to calculate all observables for the processes @xmath87 ( on proton and neutron targets ) in the kinematical region from threshold to the @xmath9-resonance region ( @xmath350 mev ) , for any value of the pion production angle , @xmath351 , ( in the cms of the @xmath4-system , ) and of the four momentum transfer @xmath125 . in order to test the present model , we used existing experimental data on the angular dependence of the differential cross sections for both the @xmath163 and @xmath347 reactions .
this comparison allowed to fix empirically the relative sign of the different contributions : born , @xmath9-excitation ( in the @xmath10channel ) and vector meson exchange ( in the @xmath13 channel ) .
the relative sign of all three diagrams for the born approximation in the case of the process @xmath347 are fixed by gauge invariance , but it is necessary to find the relative signs between the born amplitudes , on one side , and the @xmath9-isobar and vector meson exchange contributions , on another side .
the @xmath347 reaction is more sensitive to the signs of the @xmath9-contribution and @xmath352-exchange .
then the data about the differential cross sections for @xmath163 allow a further check and give a constrain for the @xmath353exchange amplitude .
note that the sign of the @xmath208exchange contribution relative to the born contribution ( in both the @xmath163 and @xmath354 reactions ) has to be the same as the relative sign of meson exchange currents ( due to the @xmath355 meson - exchange mechanism in the calculation of the electromagnetic form factors of the deuteron ) with respect to the amplitude in the impulse approximation for elastic @xmath356scattering .
this represents an important link between very different physical problems . in order to obtain a good description of the experimental data for @xmath357 and @xmath347 we introduced small corrections to the different contributions . for the reaction @xmath163 a form factor
was added to the born contribution . the @xmath349channel nucleon contribution for @xmath347
can be neglected without violating gauge invariance , because its magnetic content satisfies alone the current conservation condition . as a matter of fact this contribution
has a diverging behavior at large angle , which is typically corrected by introducing an _ ad - hoc _ form factor .
we choose to replace this contribution with a somewhat simplified phenomenological ( s - wave like ) contribution : @xmath358 where @xmath359 is a parameter which is adjusted in order to reproduce at best the @xmath0 photoproduction data .
we did not attempt to reproduce with a good accuracy the threshold behavior of the @xmath163 and @xmath347 amplitudes . a precise description of this behavior , in particular for the process @xmath163 ,
can be obtained , for example , in the framework of the chiral perturbative theory approach @xcite . for inclusive calculations ,
a qualitative description of the data in the threshold region is sufficient .
the quality of our model is shown in fig .
3 , where we present the comparison of our predictions with the experimental data on the differential cross sections for the @xmath163 and @xmath347 reactions , in the kinematical region where our model can be considered a reasonable approach . indeed
the unpolarized differential cross sections are well described .
we did not apply the model to polarization observables . in particular different @xmath360odd observables , such as , for example , the target asymmetry or the polarization of the final nucleons , are very sensitive to the relative phases of the different contributions . a good description requires a very precise treatment of the unitarity condition as well as of t - invariance of the hadron electromagnetic interaction , which are not so important for the differential or total cross section .
therefore , after having determined the relative signs of the different contributions , our model can be generalized to pion electroproduction .
our aim is the calculation of the inclusive p - odd asymmetry @xmath119 , in @xmath361 , for the sum of two possible channels , @xmath362 and @xmath363 .
one can see , from eq .
( [ eq : eq25 ] ) , that such asymmetry is determined by the following ratios of inclusive cross sections : @xmath364 for both channels , @xmath133 and @xmath132 , and @xmath365 which characterizes the relative role of the two channels .
the 2-dimensional plots of these ratios as functions of @xmath125 and @xmath3 are shown in fig . 4 and 5 , for the reactions @xmath133 and @xmath132 , respectively .
for @xmath0-electroproduction , both @xmath366 and @xmath367 are small corrections to @xmath119 . in the considered kinematical region , they are positive and tend to decrease in the region of the @xmath9 resonance , due to the dominance of the isovector resonance contribution .
the behavior of all these ratios in the threshold region can be improved , as we discussed above . in the case of the @xmath132 reaction ,
the corresponding corrections are also small , especially @xmath366 .
note that @xmath367 is negative in the whole region of @xmath125 and @xmath3 .
combining these results it is possible to calculate the resulting asymmetry @xmath119 for the sum of both channels , again in a 2-dimensional representation ( fig .
the dependence on the detailed electron kinematics for @xmath361 ( energies of the initial and final electron and electron scattering angle ) is contained in the single parameter @xmath164 , for which we used three different values : @xmath368 and 1 . in order to extract the strong @xmath125 dependence of @xmath119 , the `` reduced '' asymmetry @xmath369 is shown . in this picture
one can see that the behavior of @xmath119 versus @xmath125 and @xmath3 , in the region @xmath370 gev and in a wide region of momentum transfer @xmath125 , is smooth everywhere and negative ( note the @xmath371 factor in the formula ) .
such a behavior results from the isovector nature of the electroproduction processes which we have considered . the role of the different contributions is illustrated in figs . 7 , 8 and 9 . in fig .
7 ( fig . 8)
the ratio of the cross sections @xmath366 and @xmath367 is reported as a function of w , for two fixed values of @xmath372 , ( a ) @xmath373 and ( b ) @xmath374 ( gev / c)@xmath41 , for the reaction @xmath133 ( @xmath132 ) .
the @xmath9 contribution ( dashed - dotted line ) vanishes , while the born terms ( dotted line ) give the largest contribution at forward angles .
the contribution given by the vector meson ( @xmath11 and @xmath12 ) exchange diagrams is not so essential here .
the different contributions to the total asymmetry a are shown in fig .
9 is a projection of fig .
6 showing the resulting reduced asymmetries @xmath369 as a function of @xmath3 at a fixed value of the virtual photon polarization @xmath375 and for two values of the momentum transfer ( a ) @xmath373 ( gev / c)@xmath41 ; ( b ) @xmath374 ( gev / c)@xmath41 .
the @xmath9-contribution only is constant as a function of @xmath3 due to its isovector dominance , the vector - meson exchange gives a rather small contribution at low @xmath3 ( below 1.2 @xmath376 ) and it is negligible above .
the full calculation gives values of a varying smoothly from @xmath377 at @xmath3=1.1 gev ( close to the elastic region ) to @xmath378 at @xmath3=1.25 gev , in the region of the @xmath9 at @xmath379 ( gev / c)@xmath41 .
one purpose of the present paper is to give an estimate of a possible contamination of the elastic peak by @xmath64-production and to assess its effect on the measured asymmetry .
although a detailed comparison of elastic and inelastic channels can only be made for specific geometries at the same incident energies and scattering angles ( or equivalently , same @xmath125 and @xmath164 ) , a rough estimate can be made using inclusive cross - sections .
weak asymmetries have been calculated in the standard weinberg - salam model assuming no strangeness in the nucleon .
the results depend on the electromagnetic form factors for protons and neutrons , and give therefore different predictions depending on which values are taken .
calculations of ref.@xcite give @xmath380 , at @xmath372=0.1 ( gev / c)@xmath41 up to @xmath381 , at @xmath372=0.3 ( gev / c)@xmath41 for @xmath382= 2 gev .
the sample collaboration @xcite has measured @xmath383 at @xmath372=0.1 ( gev / c)@xmath41 and backward angle .
weak asymmetry calculations have been done for the @xmath20 @xcite , pva4-mainz @xcite and happex @xcite experimental conditions , predicting values ranging from @xmath384 at @xmath372=0.1 ( gev / c)@xmath41 to @xmath385 at @xmath372=0.5 ( gev / c)@xmath41 @xcite .
as one can see all calculations agree to predict negative asymmetries except one @xcite where small positive values have been obtained in an extension of the sm when taking right - handed doublets of @xmath52 and @xmath386 quarks into account ( in this early paper , @xmath387 is taken equal to 1/3 ) .
note that one can also find in this paper a simple explanation of the negative sign of the asymmetry in terms of @xmath5 coupling to quarks and leptons . now comparing elastic scattering and inclusive @xmath64-production ( fig .
9 ) , we see that they are both negative and of the same order of magnitude .
moreover @xmath56 is smaller in the region @xmath3=1.1 gev ( close to elastic scattering ) and larger in the @xmath297 region .
therefore we can conclude that a small admixture of @xmath64-production events in the region of the elastic peak , is not going to produce a large uncertainty in the elastic pv asymmetry .
specific cases can be computed using fig . 6 or from the corresponding numerical values available from the authors .
we have derived the dependence of the p - odd asymmetry for inclusive scattering of longitudinally polarized electrons by unpolarized protons with production of neutral and positive pions @xmath388 , with @xmath362 or @xmath363 . using the known isotopic properties of the electromagnetic current for the @xmath133 and @xmath132 processes and the vector part of the hadronic weak neutral current for the @xmath389 and @xmath390 processes , we derived an original expression for the inclusive asymmetry @xmath119 . without approximations , it is possible to disentangle the main contribution to @xmath119 , which depends only on @xmath125 .
the calculation of @xmath119 is then reduced to the analysis of specific isoscalar contributions to the electromagnetic currents .
such contributions appear , then , as corrections to the main term , which can be calculated exactly .
we have calculated the amplitudes for @xmath212 , taking into account three standard contributions : born+vector meson exchange+ @xmath9-excitation .
all the necessary parameters : interaction constants and different electromagnetic form factors are taken from other sources .
small adjustements in this basic approach were done in order to obtain a good description of the differential cross section data for the @xmath163 and @xmath347 reactions .
the model gives the vector part of the hadronic weak neutral current which is the main contribution to p - odd effects in @xmath391 .
the reduced weak asymmetry @xmath392 varies very little as a function of the two basic kinematical variables , @xmath125 and @xmath3 . in our approach
this appears naturally from the fact that the isoscalar content of the electromagnetic current for @xmath393 is very small in the considered kinematical region .
it is of the same sign and size as the @xmath394 elastic pv asymmetry , and will , therefore , not much contribute to the experimental uncertainty of the former .
a straightforward extension of the present model would open a way to use p - odd observables in elastic and inelastic electron - proton scattering for the study of the axial contributions .
we plan to discuss this in a forthcoming paper .
@xmath415 @xmath416\left ( -1+\displaystyle\frac{\vec k\cdot\vec q}{(e_1+m)(e_2+m ) } \right ) + ( 1+\kappa_v)(w+m)\left ( \displaystyle\frac{1}{e_1+m}+ \displaystyle\frac{1}{e_2+m}\right ) \right \},\ ] ] @xmath417 @xmath418\right.\ ] ] @xmath419 \displaystyle\frac{\vec k^2\vec q^2-(\vec k\cdot\vec q)^2}{(e_1+m)(e_2+m ) } \right \},\ ] ] @xmath420 @xmath421 \displaystyle\frac{(\vec k\cdot\vec q)}{(e_1+m)(e_2+m ) } \right \},\ ] ] @xmath422 @xmath423,\ ] ] @xmath424,\ ] ] @xmath425,\ ] ] where @xmath426 the isoscalar amplitudes are : @xmath435 where the contributions @xmath204 and @xmath205 are determined by the corresponding formulas , with the following substitutions : @xmath436 @xmath437 with @xmath438 in the framework of the considered model , for the process of neutral pion electroproduction , @xmath439 , the corresponding hadronic electromagnetic current is conserved:@xmath440 for any form factor in @xmath441,@xmath442,@xmath443 , and @xmath444-vertices . in case of charged pion electroproduction , @xmath445 , a special contribution
must be added to the matrix element : @xmath446 which results in additional contributions to the scalar amplitudes : @xmath447 : @xmath448 @xmath449/k^2\ ] ] @xmath450/k^2\ ] ] c.s .
et al . _ ,
rev . * 105 * , 1413 ( 1957 ) .
t.d . lee and c.n .
yang , phys . rev . *
104 * , 254 ( 1956 ) .
zeldovich , jetp * 9 * , 682 ( 1959 ) .
c.s . wood _
et al_. , science * 274 * , 1759 ( 1997 ) ; d. chauvat _
et al_. eur .
j. d * 1 * , 169 ( 1998 ) ; e.n .
et al_. , phys .
lett * 70 * , 2383 ( 1993 ) . a recent review of the field can also be found in m .- a .
bouchiat and c. bouchiat , rep .
. phys . * 60 * , 1351 ( 1997 ) .
et al_. , phys .
b * 77 * , 347 ( 1978 ) ; phys . lett .
b * 84 * , 24 ( 1979 )
. w. heil _ et al_. , nucl
b327 * , 1 ( 1979 ) .
souder , phys .
65 * , 694 ( 1990 ) .
codata recommended values of the fundamental physical constants , http://physics.nist.gov/cgi-bin/cuu .
t. p. cheng , phys .
d * 13 * , 2161 ( 1976 ) .
donoghue and c.k .
nappi , phys . lett . *
168b * , 105 ( 1976 ) .
j. gasser , h. leutwyler and m. e. sainio , phys .
b * 253 * , 252 ( 1991 ) .
j. ashman _
et al_. , phys .
b * 206 * , 364 ( 1988 ) ; nucl .
b328 * , 1 ( 1989 ) .
d. adams _ et al_. phys .
b * 329 * , 399 ( 1994 ) ; phys .
b * 339 * , 332 ( 1994 ) .
et al_. , phys .
lett . * 74 * , 346 ( 1995 ) ; phys .
lett . * 75 * , 25 ( 1995 ) .
et al_. , phys .
b * 302 * , 533 ( 1993 ) ; phys .
b * 329 * , 399 ( 1994 ) .
l. a. ahrens _ et al_. , phys .
d * 35 * , 785 ( 1987 ) .
garvey , w.c .
louis and d.h .
white , phys .
c * 48 * , 761 ( 1993 ) .
j. ellis , m. karliner , d.e .
kharzeev and m.g .
sapozhnikov , phys .
b * 353 * , 319 ( 1995 ) .
rekalo , j. arvieux and e. tomasi - gustafsson , z. phys .
a * 357 * , 133 ( 1997 ) .
rekalo , j. arvieux and e. tomasi - gustafsson , phys .
c * 56 * , 2238 ( 1997 ) .
rekalo , j. arvieux and e. tomasi - gustafsson , phys .
c * 55 * , 2630 ( 1997 ) .
d.t . spayde _ et al_. phys .
lett . * 84 * , 1106 ( 2000 ) .
r. jaffe _ et al_. , phys .
b * 229 * , 275 ( 1989 ) .
musolf and m. burkardt , z. phys .
c * 61 * , 433 ( 1994 ) .
h. weigel _
et al_. , phys .
b * 353 * , 20 ( 1995 ) .
hammer _ et al_. , phys .
b * 367 * , 323 ( 1996 ) .
kim , t. watabe and k. goeke , nucl .
a616 * , 606 ( 1997 ) .
d. leinweber , phys .
d * 55 * , 5115 ( 1996 ) .
m. musolf and h. ito , phys .
c * 55 * , 3066 ( 1997 ) .
mckeown , in parity - violation in atoms and polarized electron scattering , eds . b. frois and m .- a .
bouchiat , world scientific , 423 ( 1999 ) .
s.j . dong , k.f .
liu and a.g .
williams , phys .
d * 58 * , 074504 ( 1998 ) .
et al_. , phys .
. lett . * 82 * , 1096 ( 1999 ) .
p. achenbach _
et al_. , mami proposal a4/1 - 93 , d. v. harrach , spokesperson .
jefferson laboratory proposal e-91 - 017 ( 1991 ) , d.h .
beck , spokesperson , http://www.npl.uiuc.edu / exp / g0/. j.l .
rosner , comments nucl .
phys . , * 22 * , 205 ( 1998 ) . m.t . peskin and t. takeuchi , phys
* 65 * , 964 ( 1996 ) .
marciano and j.l .
rosner , phys .
lett . * 65 * , 2963 ( 1996 ) .
et al_. , phys .
rep . * 239 * , 1 ( 1994 ) .
n. mukhopadhyay _ et al_. , nucl
a633 * , 481 ( 1998 ) .
ramsey - musolf , phys .
c * 60 * , 045204 ( 1999 ) ; phys .
rev.c * 60 * , 045205 ( 1999 ) .
j. erler and p. langacker , hep - ph/9910315 , ( 1999 ) .
r. carlini , private communication .
s.p . wells _
et al_. , measurement of the parity violating asymmetry in the @xmath451 transition , jefferson laboratory proposal e-97 - 104 ( 1997 ) . g. feinberg , phys . rev .
d * 12 * , 3575 ( 1975 ) .
walecka , nucl . phys . *
a285 * , 349 ( 1977 ) .
w.m . alberico and a. molinari , in parity - violation in atoms and polarized electron scattering , eds . b. frois and m .- a .
bouchiat , world scientific , 1999 .
a. picklesimer and j.w .
van orden , phys .
c * 40 * , 290 ( 1989 ) .
nath , k.m .
schiller , m. krestzschmar , phys .
d * 25 * , 2300 ( 1982 ) .
d.r.t . jones and s.t .
petcov , phys .
b * 91 * , 137 ( 1980 ) .
w . hammer and d.
drechsel , z. phys .
a * 353 * , 321 ( 1995 ) .
r.n . cahn and f.j .
gilman , phys .
d * 17 * , 313 ( 1978 ) m. k. jones _ et al_. phys .
. lett . * 84 * , 1398 ( 2000 ) .
a. akhiezer and m. p. rekalo , dokl .
nauk ussr , * 180 * , 1081 ( 1968 ) and sov . j. part
* 3 * , 277 ( 1974 ) .
m. meyerhoff _ et al_. , phys .
b * 327 * , 201 ( 1994 ) ; + t. eden _
et al_. phys .
c * 50 * , r1749 ( 1995 ) ; + m. ostrick
_ et al_. phys .
lett . * 83 * , 276 ( 1995 ) ; + d. rohe _
et al_. , phys .
. lett . * 83 * , 4257 ( 1999 ) .
s. platchkov _
et al_. nucl .
phys * a510 * , 740 ( 1990 ) .
m. vanderhaeghen , k. heyde , j. ryckebusch and m. waroquier , nucl . phys . *
a595 * , 219 ( 1995 ) .
r. machleidt , k. holinde and ch .
elster , phys . rep . * 149 * , 1 ( 1987 ) . c. caso _ et al_. , eur .
j. c * 3 * , 1 ( 1998 ) . v. v. frolov _
et al_. , phys .
* 82 * , 45 ( 1999 ) .
v. bernard , n. kaiser , ulf - g .
meissner , phys .
b * 378 * , 337 ( 1996 ) .
g. fischer , h. fischer , g. von holtey , h. kmpgen , g. knop , p. schulz , and h. wessels , z. phys . *
245 * , 225 ( 1971 ) .
g. fischer , g. von holtey , g. knop and j. stmpfig , z. phys .
* 253 * , 38 ( 1972 ) .
h. genzel , e. hilger , g. knop , h. kemen and r. wedemeyer z. phys .
* 268 * , 43 ( 1974 ) .
j. napolitano , phys .
c * 43 * , 1473 ( 1991 ) .
beck , phys .
d * 39 * , 3248 ( 1989 ) . | parity violating asymmetry in inclusive scattering of longitudinally polarized electrons by unpolarized protons with @xmath0 or @xmath1 meson production , is calculated as a function of the momentum transfer squared @xmath2 and the total energy @xmath3 of the @xmath4-system .
this asymmetry , which is induced by the interference of the one - photon exchange amplitude with the parity - odd part of the @xmath5-exchange amplitude , is calculated for the @xmath6 processes ( @xmath7 is a virtual photon and @xmath8 a virtual z - boson ) considering the @xmath9-contribution in the @xmath10channel , the standard born contributions and vector meson ( @xmath11 and @xmath12 ) exchanges in the @xmath13channel . taking into account the known isotopic properties of the hadron electromagnetic and neutral currents , we show that the p - odd term is the sum of two contributions .
the main term is model independent and it can be calculated exactly in terms of fundamental constants .
it is found to be linear in @xmath2 .
the second term is a relatively small correction which is determined by the isoscalar component of the electromagnetic current . near threshold and in the @xmath9-region ,
this isoscalar part is much smaller ( in absolute value ) than the isovector one : its contribution to the asymmetry depend on the polarization state ( longitudinal or transverse ) of the virtual photon .
2@xmath2 1g@xmath14 2g@xmath15 1z@xmath16 2z@xmath17 1s@xmath18 2s@xmath19 0@xmath20 0@xmath5 |
when it comes to the ultimate physical limit of a computer , two questions arise immediately : how much information can it store , and how fast can it run @xcite ?
the former is related to the famous boltzmann s entropy formula that restricts the number of ( qu)bits available to a physical system @xcite .
the latter concerns the speed of quantum evolution , and is limited by the remarkable time - energy uncertainty relation : @xmath0 with the two bounds usually referred to as mandelstam - tamm @xcite and margolus - levitin @xcite types , respectively .
this fundamental relation determines that it takes at least @xmath1 , the so - called quantum speed limit ( qsl ) time , to evolve to an orthogonal state for a quantum system with energy spread @xmath2 or average energy @xmath3 .
this qsl time has been proved to be tight @xcite , can be extended to nonorthogonal state cases @xcite , and plays an important role in a great deal of quantum tasks such as quantum computation @xcite , quantum communication @xcite , quantum optimal control optimal control , and quantum metrology @xcite . due to the inevitable coupling to the surrounding environment , quantum systems are usually treated as open @xcite .
significant experimental progress has been made in controlling the dynamics of open systems @xcite . how to detect the limit to quantum non - unitary evolution of open systems now becomes paramountly important . in recent years
, efforts have been made towards the description of qsl time in the evolution of open systems @xcite .
a unified lower bound , including both mandelstam - tamm and margolus - levitin types , has been derived by deffner and lutz @xcite . by von neumann trace inequality and cauchy - schwarz inequality , the qsl time between an initial open system state @xmath4 and its target state @xmath5 , governed by the time - dependent non - unitary equation @xmath6 with @xmath7 a super - operator , is given by @xmath8where @xmath9 /e_{\tau } ^{p } , $ ] @xmath10 and @xmath11 denotes the schatten @xmath12-norm , @xmath13 are the singular values of @xmath14 , and @xmath15 denotes the bures angle between the initial pure state @xmath16 and the target state @xmath5 .
to study the environmental effects on qsl time , one method is to evaluate the characteristic of the intrinsic speed of the quantum evolution , i.e. , given a driving time , how fast can a quantum system evolve .
interestingly , it is discovered that non - markovian effect can speed up the quantum evolution with a damped jaynes - cummings model for an atom resonantly coupled to a leaky single mode cavity : qsl decreases when the non - markovian effect becomes stronger @xcite .
several questions , however , naturally arise : ( i ) is the speedup phenomenon a rather general feature which also exists in other physical models ? ( ii ) is non - markovianity the only key factor for speeding up quantum evolution ?
( iii ) what is the mechanism for quantum acceleration in memory environments ? to address these questions ,
in this paper , we consider a two - level atom embedded in a photonic crystal cavity ( pcc ) with the periodic dielectric structures forming photonic band gaps ( pbg ) @xcite . this structured reservoir has been widely used to form atom - photon bound states @xcite , strong localization of light @xcite , as well as entanglement preservation @xcite . in this
setting , the transition from no - speedup to speedup of quantum evolution is observed when the atomic frequency is approaching and going inside the band edge .
qsl time is found to be related to two quantities : non - markovianity during the driving time and excited population at @xmath17 .
we illustrate that it is the competition between the two quantities that ultimately determines the speedup of quantum evolution .
finally , a possible experimental realization for our illustrated phenomena by a single nitrogen - vacancy ( n - v ) center embedded in a two - dimensional pcc is discussed .
the work is organized as follows . in sec .
ii , we introduce the physical model with a two - level atom embedded in a pcc and derive the relevant qsl time in sec . iii . in sec .
iv , we explore the mechanism for the speedup of quantum evolution and suggest a pcc - n - v based experimental proposal in sec .
v. finally , we discuss and summarize the results in sec . vi .
a two - level atom ( open system ) embedded in a pcc is coupled to the radiation field ( environment ) with the hamiltonian ( @xmath18 ) @xcite @xmath19where @xmath20 is the resonant transition frequency of the atom between the excited and the ground states , @xmath21 are the pauli raising and lowering operators , @xmath22 and @xmath23 are , respectively , the frequency , the annihilation and creation operators of the @xmath24th mode of the reservoir with the real coupling constant @xmath25 . here
@xmath26 is the free space permittivity , @xmath27 and @xmath28 refer to the normalized volume and the unit polarization vector of the radiation field , respectively , and @xmath29 is the dipole moment of the atom .
the master equation for the reduced density matrix of the atom is @xmath30 with @xmath31 + \gamma _ { t}\left ( { \sigma}_{+ } { \sigma}_{-}\rho _ { t}+\rho _ { t } { % \sigma}_{+ } { \sigma}_{-}-2 { \sigma}_{-}\rho _ { t } { \sigma}_{+}\right ) , \]]where @xmath32=im@xmath33 and @xmath34=re@xmath35 are time - dependent lamb shift and decay rate respectively , and @xmath36 is the decoherence function depending on certain reservoir structures @xcite .
the reduced density matrix of the atom with an initial state @xmath37 ( in matrix form ) can be evaluated as book - open @xmath38with @xmath39 the quantum map . in this work
, we consider an ideal photonic crystal with isotropic photon dispersion relation approximated by @xmath22 @xmath40 near the band edge @xcite , where @xmath41 is the upper band - edge frequency , and @xmath42 with @xmath43 being a specific wave vector with respect to the point - group symmetry of the pcc .
the laplace transform of @xmath44 is @xmath45^{-1}$ ] with @xmath46 and @xmath47 @xcite .
the decoherence function @xmath44 can then be calculated by the standard inverse laplace transform as @xmath48where @xmath49 @xmath50 are the three parameters of the equation @xmath51 and erf(@xmath52 ) is the error function .
in this section , we evaluate the intrinsic speed for the evolution between the initial state @xmath53 and the final state @xmath5 , with @xmath17 the actual driving time . for convenience and without loss of generality ,
the initial state is set to be the excited state @xmath54 @xcite .
it is readily checked that the maximum in eq .
( [ qsl time ] ) is @xmath55 for @xmath56 . in the light of eqs .
( [ qsl time ] ) and ( [ density m ] ) , the qsl time of the above model can be derived as @xmath57with @xmath58 denoting the population of excited states at time @xmath59 . in unit of @xmath60 ( blue solid curves ) for an atom embedded in an ideal isotropic photonic crystal cavity as a function of @xmath61 the detuning between the atom transition frequency and the upper band - edge frequency , for different driving time @xmath62 and @xmath63 ( in unit of @xmath60 ) ( red dashed lines ) in @xmath64 , respectively.,width=326 ] in fig . 1
, we depict the qsl time @xmath65 in unit of @xmath66 ( blue solid curves ) together with different actual driving time @xmath17 ( red dashed lines ) as functions of @xmath67 where the transition from no - speedup to speedup is clearly shown .
when the atom transition frequency goes far outside the band gap ( e.g. , @xmath68 ) , the qsl time is actually the driving time .
however , a remarkable phenomenon appears when the atom transition frequency is approaching and going inside the band edge ( @xmath69 ) , the qsl time @xmath65 will begin to decrease , implying the intrinsic speedup of quantum evolution .
obviously , the deeper the atomic transition frequency lies in the band edge , the smaller the qsl time will be in a general trend . on the other hand , if the driving time is not very long , e.g. , in fig .
1(a ) , the speedup phenomenon only occurs inside the band edge ( @xmath70 ) , which is quite different from fig .
1(b)@xmath71(d ) , where the speedup region lies even outside the band edge .
how can we explain the above phenomena ?
can we claim that the reason for the speedup in the above model is solely due to the memory ( non - markovian ) effect of the reservoir ? to answer these questions , in the following we first describe the non - markovian behavior of the evolution of this model .
in the definition of non - markovianity [ eq .
( [ blp ] ) ] versus the driving time @xmath72 in the case of @xmath73 .
the red ( dark gray ) and green ( light gray ) curves represent the integrals with the optimal state pair @xmath74 and other 2000 randomly generated pairs , respectively.,width=326 ] the non - markovianity measure we employ here is based on the total amount of information , quantified by trace distance @xmath75 of a pair of evolved quantum states @xmath76 , flowing back from the environment .
the gradient of trace distance @xmath77 represents the information flow , with positive value indicating information flowing back to the system .
here @xmath78}$ ] denotes the dynamical map .
the non - markovianity is defined as the total backflow of information @xcite @xmath79with the maximization over all initial state pairs @xmath80 .
there exists no general analytical method to find the optimal initial state pair @xmath80 @xcite .
therefore , we will use numerical calculations instead .
[ red ( dark gray ) curves ] and excited population @xmath81 [ light blue ( light gray ) curves ] and as a function of @xmath82 in the case of @xmath83 ( in unit of @xmath84 ) in @xmath64 respectively@xmath85,width=326 ] in fig .
2 , the integrals @xmath86 in the definition of @xmath87 of 2000 random initial state pairs @xmath80 [ green ( light gray ) curves ] are generated with @xmath88 as an example .
clearly all these pairs yield smaller values than that of the state pair @xmath89 [ red ( dark gray ) curve ] .
it is easy to check that for this optimal state pairs , the trace distance of the evolved states can be written as @xmath90 @xmath91 ( @xmath92 denotes the optimal trace distance in the following ) .
based on this , we take the non - markovianity as @xmath93to associate the non - markovianity with the qsl time , we rewrite eq .
( nonm ) as @xmath94 , \label{nonm2}\]]consequently , the qsl time is reduced to @xmath95clearly , the qsl time is related to two quantities : the non - markovianity @xmath87 within the driving time and the atomic excited population @xmath96 . as an illustration , the two quantities @xmath97 [ red ( dark gray ) curves ] and @xmath96 [ light blue ( light gray ) curves ] , as functions of @xmath67 are drawn in fig .
3 . equation ( [ qsl - monm ] ) implies that the transition point from no - speedup to speedup of quantum evolution is just the point when the markovian environment becomes non - markovian , for @xmath98 when @xmath99 it is therefore easy to account for the phenomenon that the speedup only takes place within the bandgap edge when @xmath100 [ fig .
1(a ) ] , for the transition point from markovian to non - markovian environment just occurs inside the bandgap edge when the driving time is short [ fig .
3(a ) ] . equation ( [ qsl - monm ] ) also provides us with a route to explore the mechanism for the intrinsic speedup of quantum evolution . for illustration
, we consider the driving time @xmath101 as an example [ fig .
when the atom transition frequency is far outside the bandgap edge frequency ( e.g. , @xmath102 ) , the population will not be trapped , i.e. , @xmath103 and the non - markovianity also approaches zero , therefore , @xmath104 .
when the atomic transition frequency is around the bandgap edge ( e.g. , @xmath105 $ ] ) , @xmath106 is dependent on both the population @xmath96 and non - markovianity @xmath107 for the atomic excited population is trapped , i.e. , @xmath108 and the non - markovianity is strong .
finally , if the atomic transition frequency is deeply inside the bandgap edge ( e.g. , @xmath109 ) , the non - markovianity becomes small , and @xmath65 is mainly dependent on @xmath96 . summarizing , it is the competition between @xmath96 and @xmath110 that takes responsibility for the intrinsic speedup of quantum evolution .
we note that our conclusion is not only restricted to the above physical model , it can also be used to explain the phenomenon illustrated in ref .
@xcite , where the optimal trace distance for non - markovianity of a two - level atom resonantly coupled to a leaky single model cavity is found to be @xmath111 @xcite . since the monotonicity of @xmath112 and @xmath113 is the same , for simplicity , we still use @xmath113 as the trace distance ( not optimal ) for non - markovianity @xmath114 .
therefore eq .
( [ qsl - monm ] ) still holds with @xmath97 replaced by @xmath115 it is easy to check that no population will be trapped , i.e. , @xmath116 when @xmath117 thus @xmath118which is inversely proportional to non - markovianity @xmath114 if @xmath17 is longer enough . in this case , when @xmath119 ( markovian ) , @xmath120 consequently , the non - markovian effect becomes the unique reason for speeding up quantum evolution in this model .
and @xmath121 are employed as the two - level system coupled to the modes of the pcc with coupling constant g@xmath122,width=326 ] in this section , we briefly discuss a possible experimental realization for testing the above phenomena using a composite isotropic planar pcc - n - v system [ shown in fig .
4(a ) ] @xcite , where an external magnetic field * b * is applied along the [ 111 ] axis of the n - v center to lift the degeneracy between the states @xmath123 in our case , the ground state @xmath124 and the excited state @xmath125 are employed as the two - level open system coupled to the pcc with @xmath126 circular polarization mode , where @xmath127 refers to the orbital state with 0,@xmath128 orbital angular momentum projection nv - level , nv - ywl . in experiment ,
the initial state of n - v center can be prepared by 532 nm laser pulse with high fidelity @xcite , and the finial state can be tested with standard tomography technique book qic . to control the transition frequency in order to observe the transition from no - speedup to speedup phenomenon
, we may shift the frequency @xmath20 with stark effect . for the relevant experimental parameters ,
the optical transition dipole moment and the typical transition frequency of the n - v center are observed to be @xmath129 @xcite and @xmath130 @xcite , which lead to @xmath131 we can tune the energy levels by the stark shift with @xmath132 nv - stark@xmath85 therefore , @xmath133 which is large enough to obverse the no - speedup to speedup transition ( e.g. , in fig .
1(b ) , @xmath134 ) . on the other hand ,
numerous experiments have successively demonstrated the strong coupling between n - v centers and the modes of gallium phosphide pcc @xcite , silicon nitride pcc @xcite , and pcc in monocrystalline diamond @xcite , respectively .
these advances imply that this composite system is a suitable platform to test the results .
though our present work mainly focuses on a specific open system model , further study on more complicated non - markovian systems characterized by other non - markovian master equations @xcite , e.g. , post - markovian master equation @xcite , will be of great interest and importance . on the other hand ,
as the non - markovian effect plays a significant role in the speedup of quantum evolution , one interesting question then arises : will non - markovianity also be associated with some other speed limits ? in quantum many - body systems , for instance , there exists a maximum speed of information propagation in discrete quantum systems with local interactions .
this speed limit , known as lieb - robinson bound @xcite , has been extensively studied @xcite and was observed in experiment with a one - dimensional ultracold gas of bosonic atoms @xcite . since quantum systems
are always subject to decoherence and dissipation in practice , recent studies of lieb - robinson bound have been extended to classical markovian dynamics @xcite as well as quantum markovian @xcite and non - markovian @xcite conditions .
it is desirable to further investigate the effects of noise and memory environments on the lieb - robinson velocity bound , and the interplay between qsl and the lieb - robinson bound . in summary , for a model of two - level atom embedded in a pbg structured reservoir , the transition from no - speedup to speedup of quantum evolution has been observed . in particular , we have established a linkage between the qsl time and non - markovianity as well as the population of excited states for a given driving time .
we have found that it is the competition between the two quantities that finally determines the speedup of quantum evolution under noise .
the phenomenon we illustrated in this work is finally analyzed by real pcc - n - v based experimental data .
our approach may be of both theoretical and experimental interests in exploring the ultimate limits to quantum computers in memory environments .
z .- y.x . thanks dr .
sebastian deffner for helpful comments .
this work was supported by nnsfc under grant nos .
11204196 , 11274351 , 11074184 , 11375259 , 61134008 , srfdphe under grant no . 20123201120004 , and ncmis - cas under grant no .
y029152k51 .
l. mandelstam and i. tamm , j. phys .
( ussr ) * 9 * , 249 ( 1945 ) ; g.n . fleming , nuovo cimento a * 16 * , 232 ( 1973 ) ; j. anandan and y. aharonov , phys . rev .
lett . * 65 * , 1697 ( 1990 ) ; l. vaidman , am . j. phys .
* 60 * , 182 ( 1992 ) .
p. pfeifer , phys .
70 * , 3365 ( 1993 ) ; p. pfeifer and j. frhlich , rev .
* 67 * , 759 ( 1995 ) ; s. luo , z. wang , and q. zhang , j. phys . a * 35 * , 5935 ( 2002 ) ; v. giovannetti , s. lloyd , and l. maccone , phys .
a * 67 * , 052109 ( 2003 ) ; s. luo , phys .
d * 189 * , 1 ( 2004 ) ; s. luo and z. zhang , lett .
phys . * 71 * , 1 ( 2005 ) ; s. luo , j. phys .
a * 38 * , 1991 ( 2005 ) ; p. jones and p. kok , phys .
a * 82 * , 022107 ( 2010 ) ; m. zwierz , phys . rev .
a * 86 * , 016101 ( 2012 ) ; s. deffner and e. lutz , j. phys .
a * 46 * , 335302 ( 2013 ) .
et al . _ , nature commun . * 2 * , 570 ( 2011 ) ; k. h. madsen _
et al . _ , phys .
lett . * 106 * , 233601 ( 2011 ) ; u. hoeppe _
et al . _ ,
lett . * 108 * , 043603 ( 2012 ) ; f. reinhard _
et al . _ ,
lett . * 108 * , 200402 ( 2012 ) .
b. bellomo _
a * 78 * , 060302(r ) ( 2008 ) ; f .- q . wang , z .- m . zhang , and r .- s .
liang , phys .
a * 78 * , 042320 ( 2008 ) ; m. al - amri _ et al .
a * 80 * , 022314 ( 2009 ) ; b. bellomo _
et al . _ ,
t**140 * * , 014014 ( 2010 ) ; q .-
et al . _ ,
a * 81 * , 052330 ( 2010 ) ; r. lo franco _ et al .
b * 27 * , 1245053 ( 2013 ) . | memory ( non - markovian ) effect is found to be able to accelerate quantum evolution [ s. deffner and e. lutz , phys . rev . lett . * 111 * , 010402 ( 2013 ) ] . in this work , for an atom in a structured reservoir , we show that the mechanism for the speedup is not only related to non - markovianity but also to the population of excited states under a given driving time . in other words , it is the competition between non - markovianity and population of excited states that ultimately determines the acceleration of quantum evolution in memory environments .
a potential experimental realization for verifying the above phenomena is discussed by using a nitrogen - vacancy center embedded in a planar photonic crystal cavity under current technologies . |
azimuthal correlations in photon - photon collisions have been studied , in the past , in a number of papers where , in particular , the single - tag configuration was considered @xcite .
one paper was also devoted to the study of those correlations in a double - tag configuration with both electrons being tagged at small angle @xcite .
let us mention , in addition , a paper@xcite where the authors investigated azimuthal correlations in pair production in a no - tag configuration , using the acoplanarity of the produced particles in the lab frame .
the purpose of the present paper is to show how the potential of azimuthal correlations , as they can be derived from the general helicity formula for @xmath0 collisions in the case of 2-body or quasi 2-body reactions , can be exploited , in various experimental configurations , in order to extract a maximum of physical information from measurements of those reactions . for each of those configurations ,
we define the corresponding experimental constraints to be applied . in section ii
we write down the general helicity formula as a sum of 13 terms involving each a different dependence on azimuthal angles . in section
iii we apply that formula to four particular experimental configurations where the analysis of azimuthal correlations should allow one to determine a number of structure functions @xmath1 .
our treatment of the first two is rather trivial ; that of the third and fourth one is more sophisticated , since it involves the use of `` azimuthal selection '' , as we shall define it .
since the fourth configuration appears to be the most promising one and has not been considered elsewhere , we compute , in section iv , a number of corresponding applications to particular processes and dynamic models .
section v contains a brief discussion and conclusion .
kinematic constraints to be applied in configurations 2 and 4 are being computed in an appendix .
the general helicity formula for two - body or quasi two - body reactions induced by the photon - photon interaction , i.e. for processes of the type @xmath2 ( where @xmath3 may be a system of particles instead of a single one ) , as shown by fig . 1 , has been in the literature for a long time @xcite . while at the start this formula contains @xmath4 terms , since the helicity matrix of either photon is composed of @xmath5 elements , that number is considerably reduced by applying first principles , namely hermiticity , parity conservation and rotational invariance .
gathering together terms which have the same behaviour with respect to azimuthal angles , we get a 13-term formula : @xmath6 with : @xmath7 here the quantities @xmath1 ( @xmath8 ) are linear combinations of elements of the helicity tensor associated with the process @xmath9 ; these quantities , which are typical structure functions ( @xmath10 is the diagonal structure function , while all others may be called `` interference structure functions '' ) , are given as follows : @xmath11 where the tensor elements @xmath12 are defined as @xmath13 calling @xmath14 resp .
@xmath15 the helicities of @xmath16 resp . @xmath17
; the symbol @xmath18 indicates summation over the spin states of @xmath19 and @xmath3 .
the elements @xmath20 depend on @xmath21 and @xmath22 , @xmath23 ( where @xmath24 and @xmath25 are the respective four - momenta of @xmath16 and @xmath26 ) , and calling @xmath22 the polar angle of @xmath19 with respect to @xmath16 in the @xmath27 center - of - mass frame .
let us remark that hermiticity entails the relation @xmath28 , while parity conservation and rotational invariance entail @xmath29 .
in addition we use the following notations : @xmath30 is the azimuthal angle of @xmath31 with respect to @xmath32 in the @xmath27 c. m. frame with the @xmath33 axis oriented along the three - momentum of the photon @xmath16 ; @xmath34 is the azimuthal angle of @xmath19 with respect to @xmath32 in the same frame .
the polarization parameters @xmath35 and @xmath36 of , respectively , the photons @xmath16 and @xmath17 are given as follows : @xmath37 where @xmath38 is the total energy squared in the overall center - of - mass frame , while @xmath39 and @xmath40 are defined as : @xmath41 , @xmath42 , calling @xmath43 the respective four - momenta of @xmath44 .
notice that in ( 5 ) we have neglected the electron mass ; actually one has @xmath35
( @xmath45 when @xmath46 ( @xmath47 ) reaches its minimal value .
the lorentz - invariant phase space is given by @xmath48 where @xmath49 is the number of final particles , and @xmath50 are the respective four - momenta of initial and final particles , while the superscript @xmath51 indicates the energy component .
( if @xmath52 is a system of particles , an integration over all corresponding 3-momenta , except one , is implicitly included . )
finally @xmath53 is given by @xmath54 where @xmath55 is the azimuthal angle of @xmath56 with respect to @xmath57 in the lab frame with the @xmath33 axis oriented along the three - momentum of the incident electron @xmath58 .
we here assume a measurement where both outgoing electrons are tagged , and a so - called `` unfolding '' procedure is used in order to extrapolate the acceptance of the central detector to @xmath60 .
this allows one to integrate formula ( 2 ) over @xmath61 between @xmath51 and @xmath62 , so that one obtains the three - term formula ( see ref .
@xcite ) : @xmath63 actually formula ( 8) can hardly be exploited in that form for an azimuthal - correlation study , for the following reason : since @xmath64 involves longitudinal ( helicity 0 ) components of both photons , it can be shown by general arguments ( see section iii of ref . @xcite ) to stay non - negligible only when both outgoing electrons are tagged at large angles ( i.e. @xmath65 ) @xcite .
but in that case @xmath66 depends on @xmath30 ( as can be inferred from formula ( 7 ) , since @xmath55 is correlated with @xmath30 ) , and therefore the functions @xmath1 , as well as their coefficients , also depend on @xmath30 .
it is then obvious that it would become much more complicated to extract the structure functions from the analysis of the @xmath30 distribution ( all the more as the integrated cross sections are expected to be very small in that kinematic situation ) .
one is thus led to consider the case where at least one of the outgoing electrons ( for instance @xmath56 ) is tagged at small angle ( i.e. @xmath67 ) .
we now use formula ( 7 ) and in addition the relation @xmath68 that relation shows that - except for a small range near @xmath69 , and for marginal ranges of @xmath39 and @xmath40 that one can suppress by setting upper limits on those variables - one is allowed to identify @xmath30 with @xmath55 . we thus get : @xmath70 still one may consider two different experimental situations : ( a ) both electrons are tagged at small angle @xmath71 ; ( b ) @xmath56 is tagged at small angle , @xmath72 at large angle ( @xmath73 ) . in both cases
@xmath66 tends to become independent of @xmath30 , i.e. : @xmath74 while at the same time the condition for neglect of longitudinal contributions ( ref .
@xcite , formula ( 3.5 ) ) , i.e. @xmath75 , is satisfied , so that @xmath64 goes to zero .
one is then left with the two - term formula @xmath76 where in case ( a ) one has @xmath77 with the helicity - tensor elements depending only on @xmath78 and @xmath22 , while in case ( b ) one gets @xmath79 with the helicity - tensor elements depending on @xmath78 , @xmath46 and @xmath22 .
@xmath10 , being composed only of diagonal elements of the helicity - tensor , is to be identified , apart from kinematic factors , with the differential cross section @xmath80 in case ( a ) , resp .
@xmath81 in case ( b ) , for the reaction @xmath82 ; for instance , in the case where ( @xmath83 ) is a pair of charged particles , one gets @xmath84 where @xmath85 is defined as : @xmath86 . that cross section
can obviously be determined as well in the no - tag resp . the single - tag mode .
double tagging , with the constraints defined in order to allow for an azimuthal - correlation analysis , provides the possibility of extracting an additional structure function ( @xmath87 ) ; the price to pay is , of course , a lowering of the yield obtained .
notice that , at @xmath88 , formula ( 5 ) becomes considerably simplified , as one gets @xmath89 finally let us remark that , since @xmath90 , as shown by formula ( 9 ) , and since one may assume the electron - tagging systems to be cylindrically symmetric , there should be no distorsion , due to the apparatus , of the @xmath30 distribution . in that configuration the untagged ( or antitagged ) electron , e.g. @xmath31 ,
is predominantly emitted very close to @xmath91 , so that @xmath92 becomes essentially negligible as compared with @xmath93 , which has the obvious consequences that : ( i ) @xmath66 becomes independent of @xmath30 ( see again formula ( 10 ) ) ; ( ii ) @xmath90 ( formula ( 9 ) ) and therefore , assuming here again the electron tagging system to be cylindrically symmetric , there should be no distorsion of the @xmath30 distribution due to the apparatus .
we may thus integrate formula ( 2 ) over @xmath30 between @xmath51 and @xmath62 , so that we obtain the three - term formula : @xmath94 with @xmath95 again @xmath96 are given by formula ( 16 ) . in ref .
@xcite formula ( 18 ) has been applied to muon and pion pair production , using two different models for the latter . however , in that paper , the problem of experimental constraints to be applied in order to suppress the distorsion of the @xmath61 distribution , induced by the limited acceptance of the central detector , was left aside .
actually one should here assume that the latter has an `` almost @xmath97 '' acceptance .
yet a kinematic study shows that , even if the acceptance cuts of the central detector remain very small , the cuts induced by them in the polar emission angle of the particles produced in the @xmath27 c. m. frame may be large , not confined to the margins of phase space , and azimuth - dependent ( as was already noticed in ref .
@xcite ) . therefore additional constraints should be imposed in order to minimize the latter cuts .
this problem , which appears as well in configuration 4 , is treated in the appendix of this paper .
notice that in the single - tag case @xmath66 can only be determined by measuring all particles produced . when this is not possible ( in the case of multi - particle final - states ) , the fact that @xmath98 is an additional source of complications .
let us mention that a measurement of this type , involving muon pair production , has recently been performed by the l3 collaboration at lep ( cern ) @xcite .
we here assume that the electron - tagging angles are small enough to ensure that + @xmath99 , so that in formula ( 2 ) we may neglect all longitudinal helicity - tensor elements , i.e. those with at least one 0 subscript .
we are thus left with the five - term formula ( see ref .
@xcite , eq .
( 5.33 ) ) : @xmath100 with @xmath101 where the helicity - tensor elements depend on @xmath78 and @xmath22 ; @xmath102 , @xmath103 are again given by formula ( 16 ) .
at this point we note that , if the whole range @xmath104 , @xmath105 is available , one may use the orthogonality of the functions \{@xmath106 } in that range and thus extract @xmath107 by projecting formula ( 19 ) on those functions .
however it seems preferable , in order to get more information , to apply what we call `` azimuthal selection '' , i.e. to select individual distributions with respect to the azimuthal angles involved , as follows : let us assume , for instance , that we are interested in the distribution with respect to the azimuthal angle @xmath108 .
switching from our system of azimuthal variables @xmath109 to the system @xmath110 , we write : @xmath111 integrating over @xmath30 between @xmath51 and @xmath62 , we then get : @xmath112 similarly we can select three other azimuthal - angle distributions : @xmath113 here we have again made use of the fact that the @xmath30-dependence of @xmath66 can be neglected according to formula ( 10 ) .
in addition we have implicitly assumed that there is no distorsion of the @xmath114 distribution due to the apparatus ( i.e. @xmath30 varies between @xmath51 and @xmath115 whatever the values of the other variables within the phase space considered ) ; that assumption is justified , here again , by the fact that @xmath116 ( formula ( 9 ) ) , and that we may suppose both electrons to be tagged in a cylindrically symmetric way .
finally we have implicitly assumed that there is no distorsion of the @xmath61 distribution .
this assumption , as already discussed in ref .
@xcite , requires a somewhat more stringent condition to be imposed on @xmath117 and @xmath118 , namely : @xmath119 ; that means that the transverse momenta of both photons in the lab frame can be neglected with respect to the transverse momentum of @xmath120 in the @xmath27 c. m. frame . under that condition
the @xmath121 collision axis tends to be identical with the colliding - beam axis ; consequently one gets @xmath122 , and it thus becomes sufficient to assume that the central detector is , as well , cylindrically symmetric .
the condition here defined implies of course that an appropriate lower limit is assigned to @xmath123 .
as one sees from formulas ( 21)-(24 ) , azimuthal selection should allow one , in the configu- ration considered , to extract four azimuthal - angle distributions , and correspondingly five structure functions , from the data obtained in a single measurement .
it must however be remarked that because of left - right symmetry the @xmath61 and @xmath124 distributions , and correspondingly the structure functions @xmath125 and @xmath126 , should be identical .
thus , we get in fact three independent azimuthal correlations , allowing for the determination of four independent structure functions . in ref .
@xcite formulas ( 22)-(25 ) have been applied to lepton and pion pair production .
it is to be mentioned that an experiment of that type is presently being prepared at the low - energy electron - positron collider da@xmath127ne at frascati @xcite . as in case ( b ) of configuration 1 , we here assume that the electron @xmath32 is tagged at large angle ( @xmath128 ) , while @xmath31 is assumed to be tagged at small angle ( @xmath129 ) . thus @xmath66 becomes practically indendent of @xmath114 , according to formula ( 10 ) , while at the same time all helicity - tensor elements with at least one @xmath51 helicity subscript for the right - hand photon tend to vanish .
formula ( 2 ) is then reduced to an eight - term formula : @xmath130 with @xmath131 where the helicity - tensor elements depend on @xmath132 and @xmath22 ; @xmath133 are again given by formula ( 16 ) .
azimuthal selection , applied in the same way as in iii.c , makes it possible to derive from formula ( 26 ) six independent azimuthal correlations , from which eight independent structure functions can be extracted .
those correlations are the following : @xmath134 here again we have assumed that there is no distorsion of the @xmath30-distribution due to the apparatus , since @xmath90 ( formula ( 9 ) ) and we may consider that both electrons are tagged in a cylindrically symmetric way . on the other hand ( as in configuration 2 ) , given the limited acceptance of the central detector , ensuring the absence of distorsion of the @xmath61 distribution is a critical problem that will be treated in the appendix .
we shall now consider practical applications of the 8-term formula ( 26 ) computed in iii.d .
the processes considered are , here again , @xmath135 production of muon and pion pairs . in that case one gets from formula ( 6 ) , assuming @xmath136 : @xmath137 equating @xmath55 with @xmath114 ( see ( 9 ) ) , and taking account of ( 10 ) ( with @xmath138 ) , this expression transforms into : @xmath139 returning to formula ( 1 ) , and using ( 16 ) and again ( 10 ) , we obtain @xmath140 for our computations we have fixed a number of limits on the various integration parameters chosen .
those limits are the following : * @xmath141 ( corresponding approximately to a minimal tagging angle of @xmath142 mrad for @xmath56 ) , and @xmath143 ( in order to justify the neglect of @xmath30-dependence of @xmath66 in formula ( 10 ) , and at the same time to motivate the neglect of longitudinal polarization of the photon @xmath17 ) . * @xmath144 ( again because of formula ( 10 ) ) , and @xmath145 ( in order to ensure a significant contribution of the longitudinal polarization of @xmath16 ) . * @xmath146 ( see appendix ) .
* once again because of formula ( 10 ) , we set : @xmath147 ; @xmath148 , or equivalently + @xmath149 .
* @xmath150 and @xmath151 ( see appendix ) ; these limits , combined with those defined in ( iv ) , also induce further limitations of @xmath117 and @xmath66 . * finally , because of the particular sensitivity of polar - angle ranges close to @xmath51 resp .
@xmath59 with regard to possible cuts located there , we set : @xmath152 . using formula ( 36 ) and replacing @xmath153 , resp .
@xmath154 , by their expressions given in formulas ( 28)-(33 ) , and then integrating over the variables @xmath155 and @xmath66 , we have obtained , for each particular process and model considered , the six angular distributions @xmath156 , where @xmath127 is to be identified respectively with @xmath157 . in figures
2 - 5 we are showing , for @xmath127 ranging between @xmath158 and @xmath59 ( notice that , in the @xmath159 range between @xmath59 and @xmath62 , symmetric values are obtained ) , the normalized distributions @xmath160 , defined in such a way that @xmath161 .
let us remark that , since the structure functions @xmath162 and @xmath163 are changing sign with @xmath164 , the distributions in @xmath165 have been computed by integrating only over half of the @xmath164 range , i.e. @xmath166 . fig .
2 shows the azimuthal correlations obtained for muon pair production , assuming a beam energy of @xmath167 gev .
we have checked that the shapes of those curves remain practically unchanged when one goes over to a much higher beam energy ( @xmath168 gev ) .
3 and 4 show , again for @xmath169 gev , the corresponding correlations obtained for @xmath170 production , as computed respectively in two different models : the perturbative - qcd model proposed by brodsky and lepage @xcite , as extended by gunion et al .
@xcite for @xmath171 collisions , with the pion wave function given by chernyak and zhitnitsky @xcite ; and the finite - size model of poppe @xcite , where the amplitude of the born - term calculation is multiplied by an overall form factor .
notice that here the lower limit introduced for @xmath66 is @xmath172 gev .
5 shows the analogous curves obtained for @xmath173 production at the same beam energy , using again the brodsky - lepage model with the chernyak - zhitnitsky wave function .
the values of integrated cross sections ( obtained by replacing @xmath174 by @xmath10 in ( 36 ) ) are given in table 1 , where they are compared with those of the `` theoretical background '' ( see ref .
@xcite ) , i.e. of the contribution of the feynman diagram shown in fig . 6 , computed within the same kinematic limits . for @xmath175 production , we here use a simple vdm model involving only @xmath176-exchange .
it is seen that this background remains insignificant as compared to the signal ( in the case of @xmath177 production , it is of course strictly zero ) . as regards the interference term between the diagrams of figs . 1 and 6 , it is reduced to zero if , instead of identifying particle @xmath120 with the muon ( resp .
pion ) of either positive or negative charge , we average over those two options .
notice that in the figures of table 1 we have included a factor of 2 in order to account for the possibility of performing a symmetric measurement ( @xmath32 tagged at large angle , @xmath56 at small angle ; and conversely ) .
finally it is to be emphasized that , if one uses a monte - carlo program taking account of all experimental acceptances , one would certainly be allowed to suppress some of the severe kinematic restrictions here introduced .
azimuthal correlations are , to a large extent , a new approach to the study of photon - photon collisions .
till now practically all experimental efforts in two - photon physics have been concentrated on measuring @xmath178 cross sections or the `` photon structure function '' @xmath179 ( i.e. the @xmath180 total hadronic cross section ) , which correspond to the diagonal structure function @xmath10 in the formalism presented here .
azimuthal correlations should allow one to determine , depending on the configuration considered , one , two , three or seven additional independent structure functions .
it should be emphasized that in principle @xmath10 does not contain a larger amount of physical information than the others .
its special status arises only from the fact that it is more easily determined ; in addition , due to the schwarz inequality , it is larger than ( or at least equal to ) any of the interference structure functions . in that sense
it appears as the _ primus inter pares _ , but not more . in other words : azimuthal correlations should allow one to multiply the physical information , obtained in a number of two - photon processes , by a factor of 2 , 3 , 4 or 8 . as can be seen for instance by comparing figs . 3 and 4 , two different dynamic models for a given process can lead to very different shapes of the azimuthal - correlation curves , implying that the values of the structure functions involved are very different as well .
thus azimuthal correlations should provide one with a powerful tool for checking dynamic models .
actually we may safely state that no model will survive this kind of check if it is not entirely realistic .
there is however a price to pay for this achievement : there are more or less stringent experimental constraints that should be satisfied . in configurations 1 and 3 ,
what is required is the possibility of tagging the outgoing electrons ( and of measuring their azimuthal angles ) at small scattering angles . in configuration 2 , an `` almost @xmath60 ''
central detector appears necessary . in configuration 4
both requirements are combined , and in addition , since the measured cross section is sharply reduced by significant cuts in most of the parameters to be measured , a very high machine luminosity is required . however , as already noticed above , the use of a monte - carlo program should allow one to relax to some extent the constraints here defined and thus to increase the integrated cross section . in any case , and whatever the particular @xmath178 process and the configuration considered , azimuthal - correlation measurements in muon pair production under the same conditions should always be used as a test , so as to check the validity of the approximations applied .
notice that in general it would be more problematic to use electron pairs for that purpose , given the complications due to exchange between scattered and produced electrons .
let us finally remark that , for checking perturbative qcd , it appears particularly interesting to look for azimuthal correlations in the two - photon production of quark pairs ( notice that , in the quark - parton model , the azimuthal - correlation curves predicted for the production of light - quark pairs are the same as for @xmath181 production ) .
measuring azimuthal angles of jets is probably a difficult , but not impossible , task for experimentalists .
another option would be to investigate azimuthal correlations in the inclusive production of one hadron ( plus anything ) .
the authors are grateful to dr .
a. courau for illuminating discussions and precious advice .
one of us ( o. p. ) would like to thank the theory group of the `` laboratoire de physique corpusculaire , collge de france '' for the warm hospitality extended to him .
our problem is how to minimize azimuth - dependent cuts in @xmath182 , induced by the limited acceptance of the central detector . for simplicity
, we shall stick to the case of production of particle - antiparticle pairs ( @xmath183 ) .
then we can use formulas ( a10 ) - ( a12 ) of the appendix of ref .
we shall however rewrite those formulas , using slightly different notations ( and in addition , as regards formula ( a12 ) , exchanging the variables pertaining to the left- and right - hand vertex of fig .
we thus get with @xmath185 , the lab energy of particle @xmath120 , given by @xmath186^{1/2}}{2(e_x^2 -\tilde{p}_x^2)}\ ] ] where we have defined @xmath187\end{aligned}\ ] ] and where in addition we have used the following definitions : @xmath188 are the lab scattering angles of @xmath189 , while @xmath190 is the lab polar emission angle of particle @xmath120 with respect to @xmath32 ; @xmath191 ; @xmath192 .
notice that in formula ( a2 ) only solution + is valid when @xmath193 .
it is to be remarked that @xmath182 is implicitly depending on @xmath194 through its explicit dependence on @xmath55 and @xmath195 ( see formulas ( a9 ) , ( a13 ) of ref .
@xcite , showing the relations between azimuthal angles in the lab and the c. m. frame ) .
on the other hand we assume @xmath199 , which entails ( according to ( a3 ) , setting @xmath200 ) that @xmath201 .
therefore only solution + is to be considered in ( a2 ) ; in addition it becomes possible to make a convergent series expansion , in @xmath202 , of the radical in ( a2 ) .
that formula thus becomes @xmath203 where we shall retain only the first term on the right - hand side . substituting that expression of @xmath185 into ( a5 )
, we get , taking account of ( a3 ) , ( a4 ) with @xmath204 : we now assume that the acceptance of the central detector is given by : @xmath206 , with @xmath207 1 rad .
we then compute the cuts in @xmath208 induced by the forward and backward cut in @xmath190 , neglecting terms in @xmath209 . setting @xmath211 $ ] , i.e. @xmath212
, one is led to @xmath213 . for simplicity
( and since values of @xmath214 much smaller than @xmath215 are largely predominating ) , we shall approximate the above - defined upper limit of @xmath216 by requiring only @xmath217 .
99 v. e. balakin , v. m. budnev , i. f. ginzburg , jetp lett . * 11 * , 388 ( 1970 ) . v. m. budnev , i. f. ginzburg , phys . lett . *
b37 * , 320 ( 1971 ) .
a. higuchi , s. matsuda , and j. kodaira , phys .
d * 24 * , 1191 ( 1981 ) . c. peterson , f. m. zerwas , and t. f. walsh , nucl .
* b229 * , 301 ( 1983 ) . f. le diberder , lpnhe - paris internal report 84 - 05 , p. 67
( unpublished ) .
s. ong and p. kessler , mod .
lett . a * 2 * , 683 ( 1987 ) .
s. ong , p. kessler , and a. courau , mod .
lett . a * 4 * , 909 ( 1989 ) . v. l. chernyak and v. g. serbo , nucl
. phys . * b67 * , 464 ( 1973 ) . c. e. carlson and w. k. tung , phys .
d * 4 * , 2873 ( 1971 ) .
r. w. brown and i. j. muzinich , phys .
d * 4 * , 1496 ( 1971 ) . c. j. brown and d. h. lyth , nucl .
b53 * , 323 ( 1973 ) .
v. m. budnev , i. f. ginzburg , g. v. medelin , and v. g. serbo , phys .
* 15c * , 181 ( 1975 ) . c. carimalo , p. kessler , and j. parisi , proceedings of the international conference on photon - photon interactions , lake tahoe 1979 , ed .
j. f. gunion ( california university at davis , 1979 ) , p. 28 . c. carimalo , p. kessler , and j. parisi , phys .
d * 20 * , 1057 ( 1979 ) .
this may not be verified in some particular cases , e.g. when transverse - photon interactions are inhibited by a selection rule ( as the landau - yang rule for spin 1 resonance production ) or possibly if a certain inner scale comes into the picture ( see .
e. leonardi , invited talk given at _
photon 95 _ , sheffield , u.k .
, april 1995 , to be published in the proceedings .
+ j. h. field , invited talk given at the same conference .
g. alexander et al . , nuovo cimento * 107a * , 837 ( 1994 ) .
s. bellucci , a. courau , and s. ong , `` azimuthal correlations in @xmath223 at da@xmath127ne '' , preprint lal 94 - 73 , lnf 94/078(p ) , lpc 94 - 52 , to be published in the da@xmath127ne physics handbook ii ( 1995 ) . s. j. brodsky and p. lepage , phys .
d * 24 * , 1808 ( 1981 ) .
j. f. gunion , d. miller , and k. sparks , phys .
d * 33 * , 689 ( 1986 ) . v. l. chernyak and a. zhitnitsky , nucl . phys . * b201 * , 492 ( 1982 ) .
m. poppe , int . journ . of mod .
phys . a * 1 * , 545 ( 1986 ) .
j. parisi , proceedings of the international colloquium on photon - photon collisions in electron - positron storage rings , paris 1973 , journal de physique * 35 * , suppl c-3 , 52 ( 1974 ) .
cddd @xmath224 & @xmath225 & @xmath226 & @xmath227 & ( gev ) & @xmath228 ) & @xmath229)@xmath230 & 5 . &
& 88.6 @xmath231 & 5 . & 142 . & 0.215@xmath231 & 5 . & 249 . &
0.215 @xmath232 & 5 . & 20 . &
@xmath230 & 100 . & 222 . & 0.283 | using the general helicity formula for @xmath0 collisions , we are showing that it should be possible to determine a number of independent `` structure functions '' , i.e. linear combinations of elements of the two - photon helicity tensor , through azimuthal correlations in two - body or quasi two - body reactions induced by the photon - photon interaction , provided certain experimental conditions are satisfied .
numerical results of our computations are presented for some particular processes and dynamic models . |
after the discovery of the higgs boson - or several higgs bosons in extensions beyond the standard model ( sm ) - we have to probe its properties in order to establish the higgs mechanism @xcite as responsible for the creation of particle masses without violating gauge symmetry . in the sm the higgs mechanism is implemented by adding one isodoublet complex scalar field which leads after electroweak symmetry breaking to one single spin zero cp - even higgs particle @xcite . in extensions beyond the sm
the higgs sector can be non - minimal , as _
e.g. _ in the minimal supersymmetric standard model ( mssm ) , which contains five physical higgs states , _
i.e. _ two cp - even @xmath1 and @xmath2 , one cp - odd @xmath3 and two charged @xmath4 higgs bosons @xcite . to establish the higgs mechanism
experimentally we have to determine the higgs spin , its behavior under charge and parity transformations , the couplings to gauge bosons and fermions , and finally the trilinear and quartic higgs self - interactions must be measured to reconstruct the higgs potential itself . to fulfill this program the large hadron collider ( lhc ) analyses @xcite
must be complemented by the high - precision measurements at a future international linear @xmath0 collider ( ilc ) @xcite . + we address in this contribution the determination of the higgs cp quantum numbers .
to do so in an unambiguous way is somewhat problematic @xcite .
observables sensitive to the higgs spin - parity such as angular correlations in higgs decays into @xmath5 pairs @xcite or in higgs production with or through these states @xcite only project out the cp - even component of the @xmath6 coupling , even in the presence of cp violation .
in addition , the purely pseudoscalar @xmath7 coupling is zero at tree - level and is generated only through tiny loop corrections . in the higgs couplings to fermions , however , the cp - even and cp - odd components can have the same magnitude . here
, the heaviest fermion discovered so far , the top quark , plays a special role .
the @xmath8 coupling is largest being proportional to the top quark mass due to the higgs mechanism . at a future ilc the higgs boson
can therefore be produced with sufficient rate in associated production with a @xmath9 pair , @xmath10 @xcite .
we propose a simple and straightforward way to determine the cp nature of a sm - like higgs boson in this process , in an unambigious way , where we exploit that the cross section as well as the top quark polarization behave in a radically different way for cp - even and cp - odd higgs production .
the diagrams which contribute in the sm to the process @xmath11 are shown in fig.[fig : diagrams ] . the bulk of the cross section is generated when the higgs is r0.6 radiated off the heavy top quarks @xcite , whereas the higgs produced in association with a @xmath12 boson which then splits into a @xmath9 pair provides only a very small contribution , amounting to a few percent for @xmath13 tev detailed simulations have shown the cross section to be measurable with an accuracy of order 10% @xcite .
we will discuss the case of a sm - like mixed cp higgs state @xmath14 and use the general form of the @xmath15 coupling @xmath16 where the coefficients @xmath17 and @xmath18 are assumed to be real and @xmath19 . in the sm we have @xmath20 and for a purely pseudoscalar higgs boson @xmath21 . in the pseudoscalar case we take @xmath22 , consistent with a convenient normalization @xmath23 chosen for the general case of a higgs boson with an indefinite cp quantum number .
zero value for the product @xmath24 will hence signal cp violation in the higgs sector . for the @xmath25 coupling
, we will use the form , @xmath26 and for the numerical analysis we chose @xmath27 @xcite as @xmath28 in the case of a cp - even ( odd ) higgs boson .
we will thus have only one free parameter @xmath18 .
however , this simple parametrization for a sm like higgs need not be true in _ e.g. _ a general 2hdm , where @xmath29 and @xmath30 are three independent parameters .
we have calculated the cross section for the production of a mixed cp higgs state including the polarization dependence of the final state top quarks .
the lengthy result has been checked to agree with ref .
@xcite for the unpolarized total cross section . in fig .
[ fig : totalcxn ] left panel we show the production cross section for a purely scalar ( @xmath2 with @xmath31 ) and a pseudoscalar ( @xmath3 with @xmath22 ) higgs as a function of the c.m .
energy and for two mass values @xmath32 and @xmath33 gev .
as can be inferred from the figure , the threshold rise of the cross section in the scalar and the pseudoscalar case is very different .
furthermore , for the same strength of the @xmath34 couplings , there is an order of magnitude difference between the @xmath2 and @xmath3 cross sections at moderate energies . only at very high energies , @xmath35 tev ,
the chiral limit is reached and the two cross sections become equal , r0.5 up to the small contribution due to the diagram including the @xmath25 coupling .
these two features hence provide an extremely powerful tool to discriminate the cp properties of the spin zero particle produced together with the top quark pair .
the difference in the threshold behavior of the @xmath3 and @xmath2 case is strong enough so that the cross section measurement at only two different c.m .
energies allows a clear determination of the cp properties of the @xmath14 state .
_ e.g. _ for @xmath36 gev , the ratio of the cross sections at @xmath37 and 500 gev is @xmath38 and @xmath39 for the scalar and pseudoscalar case , respectively .
finally , taking the ratio makes the conclusion robust with respect to the effect of the top yukawa coupling , the higher order radiative corrections @xcite or systematic errors in the measurement .
we also studied the @xmath18 dependence of the total @xmath15 production process at a given energy and for fixed @xmath40 .
being a cp even quantity it only depends on @xmath41 .
[ fig : totalcxn ] ( right ) shows the result for unpolarized and polarized @xmath42 beams . for the latter
, we used the standard ilc values @xmath43 and @xmath44 , which double the total rate .
since the top quark , due to its large decay width @xmath45 gev , decays much before hadronization , its spin information is translated to the decay distribution before contamination through strong interaction effects .
furthermore , the lepton angular distribution of the decay @xmath46 is independent of any non - standard effects in the decay vertex , so that it is a pure probe of the physics of the top quark production process @xcite .
the net polarization of the top quark therefore provides an interesting tool for the probe of @xmath18 , see also ref.@xcite . in fig .
[ fig : toppol ] ( left ) we show as a function of @xmath47 for @xmath36 and @xmath33 gev in the @xmath48 and @xmath49 case the expected degree of @xmath50-quark polarization @xmath51 , given by @xmath52 as can be inferred from the figure , the degree of top polarization is again strikingly different for the cp even and cp odd case and shows a very different threshold dependence .
in addition , since @xmath51 is constructed as a ratio of cross sections , the insights gained from this variable are not affected by a possibly model dependent normalization of the overall @xmath53 coupling strength , higher order corrections etc .
@xmath51 is a parity odd quantity and receives contributions from the interferences between the @xmath54 and all @xmath12 exchange diagrams , with the one stemming from the diagram involving the @xmath25 vertex being small .
the parity violation effect for the emission of a ( pseudo)scalar is controlled by the ( vector ) axial - vector @xmath55 coupling ( @xmath56 ) @xmath57 , where @xmath58 denotes the top isospin and r0.5 @xmath59 the electric charge .
hence the ratios of the @xmath51 values away from threshold are expected to be given by the @xmath60 , which indeed is confirmed by both figs .
[ fig : toppol ] at @xmath37 gev .
we investigate how the behavior of the cross section and the measurement of the top polarization , which both are clear discriminators between a scalar and pseudoscalar higgs state , can be used to get information on the cp mixing , _
i.e. _ the value of @xmath18 . ignoring systematical errors , the sensitivity of an observable @xmath61 to the parameter @xmath18 at @xmath62 is @xmath63 , if @xmath64 , where @xmath65 is the statistical fluctuation in @xmath66 at an integrated luminosity @xmath67 . for the cross section @xmath68 and the polarization @xmath51 , the statistical fluctuation at a level of confidence @xmath69
are given by @xmath70 and @xmath71 .
[ fig : sens ] ( left ) shows the sensitivity @xmath63 from the cross section measurement for r0.5 @xmath72 gev at @xmath37 gev with @xmath73 fb@xmath74 . for polarized @xmath42 beams it varies from 0.25 for @xmath48 to 0.01 for @xmath49 , a rather precise determination obtained from a simple measurement .
the top polarization is less sensitive to @xmath18 , see fig .
[ fig : sens ] ( right ) . both @xmath68 and @xmath51
are cp even quantities , they can not depend linearly on @xmath18 and hence not probe cp violation directly
. observables depending directly on the sine of the azimuthal angle ( @xmath14 ) are linear in @xmath18 .
the up - down asymmetry @xmath75 of the antitop quark production with respect to the top - electron plane ( @xmath76 ) is an example of such an observable : @xmath77 with @xmath78 , where @xmath79 is the @xmath80 momentum in the @xmath81-higgs rest frame .
[ fig : udasy ] shows the asymmetry @xmath75 for a higgs boson of 120 gev and a c.m .
energy of @xmath37 gev as a function of @xmath18 .
it can reach values of order 5% .
the non - zero value of the asymmetry arises from the channel which involves the @xmath25 coupling .
we have shown that the total cross section and the top polarization asymmetry for associated higgs production with top quark pairs in @xmath0 collisions provide a very simple and unambiguous determination of the cp quantum numbers of a sm - like higgs particle . exploiting the up - down asymmetry of the anti - top with respect to the top - electron plane we further have a direct probe of cp violation at hand .
99 slides : + ` http://ilcagenda.linearcollider.org/contributiondisplay.py?contribid=193&sessionid=85&confid=1296 ` p.w .
higgs , phys .
* 13 * 508 ( 1964 ) ; _ ibid .
_ phys . rev . *
145 * 1156 ( 1966 ) ; f. englert and r. brout , phys .
* 13 * 321 ( 1964 ) ; g.s .
guralnik , c.r .
hagen and t. kibble , phys .
* 13 * 585 ( 1965 ) .
j. gunion , h. haber , g. kane , and s. dawson , _ the higgs hunter s guide _ , addison
wesley , reading ( usa ) , 1990 ; for recent reviews , see a. djouadi , arxiv : hep - ph/0503172 and arxiv : hep - ph/0503173 to appear in physics reports ; m. gomez - bock et al . , arxiv : hep - ph/0509077 .
k. hagiwara and m. stong , z. phys . * c62 * 99 ( 1994 ) ; t. plehn , d. rainwater and d. zeppenfeld , phys .
. lett . * 88 * 051801 ( 2002 ) ; d. miller _
et al . _ ,
* b505 * 149 ( 2001 ) ; t. han and j. jiang , phys . rev . * d63 * 096007 ( 2001 ) ; s. biswal _ et al . _ ,
phys . rev .
* d73 * 035001 ( 2006 ) .
s. dittmaier _
et al . _ ,
* b441 * 383 ( 1998 ) ; s. dawson and l. reina , phys .
* d57 * 5851 ( 1998 ) , * d59 * 054012 ( 1999 ) and * d60 * 015003 ( 1999 ) ; s. dittmaier _ et al . _ , phys . lett . *
b478 * 247 ( 2000 ) ; g. blanger _
et al . _ ,
b571 * 163 ( 2003 ) ; a. denner _
et al . _ ,
b575 * 290 ( 2003 ) and nucl . phys . *
b680 * 85 ( 2004 ) . | one of the main endeavors at future high - energy colliders is the search for the higgs boson(s ) and , once found , the probe of the fundamental properties . in particular
, the charge conjugation and parity ( cp ) quantum numbers have to be determined .
we show that these are unambiguously accessible at future @xmath0 colliders through the measurement of the total cross section and the top polarization in associated higgs production with top quark pairs . |
in the next few years , the sm description of the charged current interactions through the cabibbo kobayashi maskawa ( ckm ) matrix @xcite , and , in particular , the nature of cp violation , will be subject to tests of unprecedented precision .
the final objective is to over - constrain the ckm matrix and , thus , probe any effects due to new physics .
two important tests will be the determination of @xmath7 from the cp violating asymmetry in @xmath8 , where @xmath9 , and the search for @xmath10 mixing .
interesting constraints will also arise from the decay chains @xmath11 .
the idea here is based on the fact that @xmath12 if the final state @xmath13 is common to @xmath14 and @xmath15 , then these two amplitudes interfere and one probes essentially the weak phase @xmath16 .
the phenomenological factor @xmath17 @xcite accounts for the fact that the @xmath18 decay is color suppressed , while the @xmath19 decay is not . inspired by a triangular construction due to gronau and london @xcite , gronau and wyler ( gw ) proposed a method to extract @xmath20 which uses the decay chains @xmath21 , where @xmath22 is a cp eigenstate @xcite .
atwood , dunietz , and soni ( ads ) have modified this method , using only final states @xmath13 which are not cp eigenstates @xcite .
recently , soffer has stressed the experimental advantages of combining the two strategies into a single analysis , while pointing out the complications due to the discrete ambiguities inherent in these methods and other measurements of direct cp violation @xcite .
the nice feature of these decays is that they involve only tree level diagrams and , thus , are not subject to penguin pollution .
however , one must consider what effects the mixing in the @xmath2 system might have on the @xmath11 decay chains @xcite , especially if these measurements are used to uncover new physics .
otherwise , new physics effects in the @xmath2 system could be misidentified as new physics in the @xmath23 system .
in fact , meca and silva @xcite have used @xmath24 to argue that these effects could be of order @xmath25 in @xmath26 decays , and they may be as large as @xmath27 in @xmath28 or @xmath29 decays @xcite . the main objective of our article is to study in detail the effect of @xmath2 mixing on the measurement of @xmath0 in @xmath30 decays .
if @xmath2 mixing is observed , it must be incorporated into the @xmath30 analysis .
if only upper limits on mixing are known , their effect should be included as a systematic error .
we also mention briefly how the effect of @xmath2 mixing may , under certain conditions , be detected in the measurement of @xmath0 . in section [ sec :
notation ] we establish our notation . in section [ sec : decays ] we present the complete expressions for the @xmath31 decay rates in the presence of @xmath2 mixing . in section [ sec : thebigone ] , we start by studying the influence of @xmath2 mixing on the gw and ads methods separately , concentrating on some regions of parameter space . we show that these effects depend on the specific value of @xmath0 and that they might affect the extraction of @xmath20 by as much as @xmath32 , even for values of @xmath3 and/or @xmath4 around @xmath5 .
then , we combine the gw and ads methods into a realistic experimental analysis , performing a scan over parameter space to discuss the impact that the mixing effects have on such experiments .
we also show how to include the mixing effects in the analysis .
we draw our conclusions in section [ sec : conclusions ] . for completeness , the formulae relevant for the study of the @xmath2 system are included in appendix [ app : a ] .
appendix [ app : b ] contains a comparison between cp - even and cp - odd corrections to the extraction of @xmath20 .
appendix [ app : c ] presents an analysis of the measurements of the strong phases in the @xmath33 decays that enter in the extraction of @xmath0 , and which may be performed at the tau - charm factories .
one might be surprised by the fact that eq .
( [ thetrick ] ) , is not invariant under the rephasing of the @xmath34 and @xmath35 quarks . in fact , the ratio measured experimentally in the gw and ads methods is rather @xmath36 this ratio depends on the weak phase in eq .
( [ thetrick ] ) , on the relative weak phase between the decay amplitudes @xmath37 and @xmath38 , and it has the correct rephasing - invariant properties .
the weak phase in eq .
( [ correct ] ) is essentially given by @xmath0 .
indeed , tree level , @xmath39-mediated @xmath33 decays only probe the weak phase in the first two families , @xmath40 , and this is also the weak phase that appears with @xmath0 in eq .
( [ thetrick ] ) . in the sm
, @xmath41 lies around @xmath42 radians @xcite and its presence is completely irrelevant .
new physics could , in principle , affect this result by altering @xmath41 or by allowing for new diagrams to drive the @xmath33 decays .
however , both @xmath41 @xcite and any additional contributions to @xmath33 decays @xcite are likely to remain small in most models of new physics .
we will neglect them henceforth . for simplicity
, we will use the parametrizations @xmath43 for the initial @xmath44 decays ; @xmath45 for the @xmath33 decays into non - cp eigenstates ; and @xmath46 for @xmath33 decays into a cp eigenstate @xmath22 with cp eigenvalue @xmath47 . in these parametrizations
, we have removed all irrelevant overall cp - even phases . however , the differences between cp - even phases in competing paths , @xmath48 and @xmath49 , are physically meaningful . @xmath50 and
@xmath51 are discussed in the next section . in the sm ,
the value of @xmath0 is already constrained by an overall analysis of the unitarity triangle . in recent reviews
ali and london find @xcite @xmath52 while buras
@xcite quotes @xmath53 .
the variations found in the literature are mostly due to different estimates for the theoretical errors and to the different methods used to combine the theoretical and experimental errors . for definiteness
, we will use eq .
( [ smboundgamma ] ) as a reasonable estimate for the allowed region in the sm .
however , we stress that @xmath0 is allowed to take any value in our analysis .
we are concerned not only with the effects that @xmath2 mixing may have on sm values of @xmath0 , but also with the possibility that such mixing effects may ` hide ' new - physics by bringing values of @xmath0 outside the sm allowed region into this region . using @xmath54 @xcite , eq .
( [ thetrick ] ) , and assuming factorization , we get @xmath55 notice that the parametrization in eq .
( [ param : b ] ) is valid for any @xmath56 decay , with @xmath57 , and higher resonances , but the exact values of @xmath58 , @xmath51 and @xmath48 will vary from one channel to the next .
for example , for the @xmath59 and @xmath60 decays we expect @xmath61 in addition , @xmath62 = ( 1.34 \pm 0.18 ) \times 10^{-2}$ ] , @xmath63 = ( 5.3 \pm 0.5 ) \times 10^{-3}$ ] @xcite , and @xmath64/ { \cal b}[b^+ \rightarrow \overline{d^0 } \pi^+ ] = 0.055 \pm 0.015 $ ] @xcite .
while the factor @xmath65 is roughly @xmath66 ( @xmath67 ) times larger in @xmath59 ( @xmath60 ) decays than it is in @xmath68 decays , the decay chains of interest to us will scale instead with the @xmath69 factors in the ckm - suppressed decays @xmath70 & \sim & \tilde{\epsilon}^2 \
{ \cal b}[b^+ \rightarrow \overline{d^0 } k^+ ] \sim 2.4 \times 10^{-6 } , \nonumber\\ { \cal b}[b^+ \rightarrow d^0 \pi^+ ] & \sim & \tilde{\epsilon}_\pi^2 \ { \cal b}[b^+ \rightarrow \overline{d^0 } \pi^+ ] \sim 1.0 \times 10^{-7 } , \nonumber\\ { \cal b}[b^+ \rightarrow d^0 \rho^+ ] & \sim & \tilde{\epsilon}_\pi^2 \ { \cal b}[b^+ \rightarrow \overline{d^0 } \rho^+ ] \sim 2.5 \times 10^{-7}. \label{thescale}\end{aligned}\ ] ] as a result , the @xmath68 decays are the best to extract @xmath20 .
similarly , for the decays used in the ads method , @xmath50 is the magnitude of the ratio of the amplitudes of the doubly cabibbo suppressed ( dcs ) decay @xmath71 to the cabibbo allowed ( ca ) decay @xmath72 .
this may be estimated from @xcite @xmath73 } { { \cal b}[d^0 \rightarrow k^- \pi^+ ] } } \sim \sqrt{0.0031 } \sim 0.06\ .
\label{estimate : epsilon}\ ] ] the parametrization in eq .
( [ param : non - cp ] ) is valid for any @xmath74 decay , but the exact values of @xmath75 , @xmath50 and @xmath49 will vary from one channel to the next . in particular , eq .
( [ param : non - cp ] ) reduces to eq .
( [ param : cp ] ) when @xmath76 and @xmath77 is the cp eigenvalue of @xmath22 . for the cp - even eigenstates , such as @xmath78 and @xmath79 , @xmath80 . for the cp - odd eigenstates , such as @xmath81 and @xmath82 , @xmath83 .
the mixing in the @xmath2 system may be parametrized by the mass difference divided by the average width ( @xmath84 ) , by the width difference divided by twice the average width ( @xmath85 ) , and by a cp - violating phase ( @xmath86 ) defined by @xmath87 .
assuming @xmath88 , @xmath89 at the @xmath90 c. l. @xcite .
this bound is likely to remain stable even when one allows for @xmath91 @xcite .
predictions for @xmath3 and @xmath4 within the sm vary considerably among the different authors @xcite , but it is probably safe to estimate an upper bound around a few times @xmath92 .
this uncertainty is due to the role played by su(3 ) breaking effects in the long - distance part of the calculation .
for example , bucella , lusignoli and pugliese @xcite have estimated the sm value for @xmath4 to lie around @xmath93 . however , this result might be subject to sizeable errors , since it comes about through a large cancellation between two individual contributions , each of order @xmath94 .
on the other hand , in the sm , the value of @xmath86 is related to @xmath41 and is guaranteed to be extremely small .
when one goes beyond the sm , one finds many models for which @xmath86 may be large and @xmath24 , while @xmath4 ( which is closely related to the decay rates , where one would hardly expect any large new physics contributions ) is likely to retain its ( rather uncertain ) sm value @xcite .
ultimately , these values will be determined experimentally at @xmath58-factories , in fixed target experiments , and at tau - charm factories .
we take the point of view that , until a determination of @xmath3 and @xmath4 is available , their upper bounds must be included as a systematic uncertainty in the experimental determination of @xmath0 from the @xmath1 decays .
the relevant point about the notation introduced here is that @xmath50 and @xmath51 are of order @xmath95 , while we will take @xmath96 as an illustrative example .
( this is of the order of the sensitivity expected in the near future @xcite . ) therefore , any effect proportional to @xmath97 , @xmath98 , @xmath99 , or @xmath100 is naturally of order @xmath25 @xcite , and might be larger depending on the exact value of the other parameters in the problem .
we should stress that the @xmath2 mixing effects might be of order unity , or even dominate , in the @xmath101 and @xmath102 decays , because @xmath103 @xcite . as a result ,
even sm values around @xmath104 , would lead to a @xmath25 effect in these channels @xcite .
the @xmath105 decay chain is shown in fig .
[ fig:1 ] .
the solid lines refer to decays and the dashed lines to the time evolution of the @xmath2 system .
the functions @xmath106 and @xmath107 are discussed in appendix [ app : a ] and describe the flavor preserving and flavor changing time evolutions , respectively .
the corresponding decay amplitude is obtained simply by adding the four possible decay paths @xmath108 & + & a(b^+ \rightarrow \overline{d^0 } k^+)\ \frac{p_d}{q_d}\ , g_-(t)\ a(d^0 \rightarrow f_d ) \nonumber\\*[1 mm ] & + & a(b^+ \rightarrow d^0 k^+)\ \frac{q_d}{p_d}\ , g_-(t)\ a(\overline{d^0 } \rightarrow f_d ) .
\label{master1}\end{aligned}\ ] ] the magnitude squared of this expression yields the time dependent decay rate . using the notation of section [ sec : notation ] , we find , @xmath109 = \nonumber\\ & & \hspace{8ex } \left [ \tilde{\epsilon}^2 + \epsilon^2 - 2 \epsilon \tilde{\epsilon } \cos ( \gamma + \delta_b - \delta_d ) \right ] \nonumber\\ & & \hspace{6ex } + \
|g_-(t)|^2\ \left [ 1 + \epsilon^2 \tilde{\epsilon}^2 - 2 \epsilon \tilde{\epsilon } \cos ( \gamma + 4 \theta_d + \delta_b + \delta_d ) \right ] \nonumber\\ & & \hspace{6ex } + \
2\ , \mbox{im } \left(g_+^\ast(t ) g_-(t ) \right)\ \left [ \tilde{\epsilon } ( 1 - \epsilon^2 ) \sin ( \gamma + 2 \theta_d + \delta_b ) - \epsilon ( 1 - \tilde{\epsilon}^2 ) \sin ( 2 \theta_d + \delta_d ) \right ] \nonumber\\ & & \hspace{6ex } + \
2\ , \mbox{re } \left(g_+^\ast(t ) g_-(t ) \right)\ \left [ \tilde{\epsilon } ( 1 + \epsilon^2 ) \cos ( \gamma + 2 \theta_d + \delta_b ) - \epsilon ( 1 + \tilde{\epsilon}^2 ) \cos ( 2 \theta_d + \delta_d ) \right ] .
\label{master}\end{aligned}\ ] ] to obtain the expression for the cp - conjugated decay rate , @xmath110 , we simply substitute @xmath111 and @xmath112 . as shown in appendix [ app : a ] , the time integrated decay rates may be obtained through the substitutions @xmath113 and @xmath114 .
these expressions are completely general ; no approximations were made ( except for the use of @xmath115 ) . in subsequent derivations
we will often simplify the expressions , using the fact that @xmath3 , @xmath4 , and @xmath51 are small when compared with one .
however the plots and estimates presented in this article were calculated using the complete formulae .
we will also expand in @xmath50 , except for decays into cp - even ( cp - odd ) eigenstates , where @xmath116 ( @xmath117 ) , @xmath118 , and @xmath119 .
( [ master ] ) exhibits the usual three types of cp - violating terms . ) , along with cp - conserving terms . ]
the first line on the rhs contains a term proportional to @xmath120 , which is due to direct cp violation in the @xmath1 decays ( and appears multiplied by @xmath121 in the full decay chain ) .
the last line contains a term proportional to @xmath122 , which is due to direct cp violation in the @xmath74 decays . in both cases ,
cp violation requires a non - zero strong phase . the usual cp violation in @xmath2 mixing
has already been neglected in eq .
( [ master ] ) due to the assumption that @xmath123 .
the third line on the rhs contains a term proportional to @xmath124 .
this term is due to the interference between the mixing in the @xmath2 system and its subsequent decay into the final state @xmath13 .
this term is not zero even if the strong phase @xmath49 vanishes , but it requires a nontrivial new phase in the mixing , @xmath86 .
we could name this a first - mix - then - decay type of interference cp violation .
in addition , eq . ( [ master ] ) contains a fourth type of cp - violating term .
this appears on the third line of eq .
( [ master ] ) as the term involving @xmath125 .
this term persists even if @xmath126 ( meaning that no strong phase is required ) , or / and @xmath88 ( meaning that no new phase is required in @xmath2 mixing ) .
in fact , in the limit @xmath127 , this term is proportional to @xmath128 and , therefore , it is large even within the sm .
this fourth type of cp violation was first pointed out by meca and silva @xcite , and it arises from the interference between the @xmath129 decays and the subsequent @xmath2 mixing . we might call this the first - decay - then - mix type of interference cp violation . its effect
is best highlighted by considering a final state @xmath13 which tags the flavor of the @xmath14 meson ( corresponding to @xmath130 in fig .
[ fig:1 ] ) .
in that case fig .
[ fig:1 ] has only two paths and there is still cp violation proportional to @xmath128 , meaning that it can be large even within the sm .
this example should help in clearing some frequent misconceptions found in the literature .
we see that * one may in principle use a single charged @xmath58 decay ( and its cp conjugate ) to probe a source of cp violation which does not require strong phases , as long as a neutral @xmath33 meson system is present in the intermediate state ; * one can have a large cp - violating phase involving the @xmath2 system , even within the sm .
it is true that , in the sm , the cp - violating phases present in the @xmath2 mixing and @xmath33 decays are very small , because they are proportional to @xmath41 .
however , there is a large cp - violating _ phase _ in the first - decay - then - mix type of interference cp violation involved in @xmath129 decays , even within the sm ; * the first - decay - then - mix type of interference cp violation can , in principle , be probed even when the @xmath33 meson is detected through its ( flavor - tagging ) semileptonic decay . in general
these effects require values of @xmath3 and/or @xmath4 around @xmath5 , in order to have an impact on @xmath131 decays
. nevertheless , as we shall see below , some values of the parameters are possible for which @xmath132 would give a @xmath32 effect , meaning that @xmath133 would still give a @xmath134 effect .
if there were no mixing in the @xmath2 system , then we would have @xmath135 and @xmath136 , _ cf .
_ appendix [ app : a ] .
this would leave only the uppermost and lowermost ( unmixed ) decay paths of fig .
[ fig:1 ] ; the mixed paths represented by the diagonal dashed time - evolution lines would be absent .
this is precisely what one assumes in both the gw and ads methods . in that case
, the decay amplitude reduces to the first two line of eq .
( [ master1 ] ) and the only relevant phase is that in eq .
( [ correct ] ) . as discussed above , the corresponding weak phase is essentially given by @xmath0 .
obviously , under the no - mixing approximation , these decays are completely insensitive to any cp - violating phase @xmath86 that might be present in @xmath2 mixing . in the gw method ,
one uses time - integrated decays rates into cp - eigenstates , given by @xmath137 % /(a_{\rm
cp}^2 b^2 ) & \propto & 1 + \tilde{\epsilon}^2 + 2 \tilde{\epsilon } \cos ( \gamma + \delta_b ) , \nonumber\\ \gamma \left [ b^- \rightarrow f_{\rm cp } k^- \right ] % /(a_{\rm
cp}^2 b^2 ) & \propto & 1 + \tilde{\epsilon}^2 + 2 \tilde{\epsilon } \cos ( \gamma - \delta_b ) .
\label{gwdecays}\end{aligned}\ ] ] gronau and wyler @xcite assume that @xmath138 , @xmath58 , and @xmath51 are known from the decay rates of @xmath139 , @xmath19 , and @xmath18 , respectively .
therefore , eqs .
( [ gwdecays ] ) determine @xmath140 from which one may extract , @xmath141 , \label{trigno - exercise } \\
\sin^2 \delta & = & \frac{1}{2 } \left [ 1 - c_+ c_- \mp \sqrt{(1-c_+^2)(1-c_-^2 ) } \right ] .
\label{trigno - exercise - delta}\end{aligned}\ ] ] as discussed below , this determines @xmath0 up to an eight - fold ambiguity @xcite .
unfortunately , the gw method is difficult to implement for two main reasons .
the first reason is due to the hierarchy between the two interfering amplitudes presented in eqs .
( [ thetrick ] ) and ( [ estimate : epsilonk ] ) .
this is easily seen by noting that the gw method hinges on extracting an interference of order @xmath51 from an overall rate of order one , as shown in eqs .
( [ gwdecays ] ) . since eqs .
( [ gwdecays ] ) can be visualized as two triangles in the complex amplitude plane , this problem is sometimes explained by pointing out that the two triangles are squashed .
the second reason arises from the fact that it is very difficult to measure @xmath142 $ ] ( and , thus , @xmath51 ) experimentally .
one could envision determining @xmath142 $ ] through the unmixed semileptonic decay chain , @xmath143 .
however , since full reconstruction is impossible in semileptonic decays , this process is subject to daunting combinatoric backgrounds .
therefore , one is led to probe @xmath142 $ ] through the subsequent decay of the @xmath14 into hadronic final states , such as @xmath144 . however , as pointed out by atwood , dunietz and soni @xcite , the two ( unmixed ) decay chains , @xmath145 and @xmath146 interfere at order one , since @xmath147 this means that it is very difficult to determine @xmath51 and , thus , to implement the gw method .
atwood , dunietz and soni @xcite , have turned this order one interference problem into an asset . in their method ,
one uses two final states @xmath13 for which @xmath148 is a dcs decay , while @xmath149 is cabibbo allowed .
examples include @xmath150 , @xmath151 , @xmath152 , etc .
then , there are four decay rates , @xmath153 % /(a_{i}^2 b^2 ) & \propto & \tilde{\epsilon}^2 + \epsilon_i^2 - 2 \tilde{\epsilon } \epsilon_i \cos ( \gamma + \delta_b - \delta_{di } ) , \nonumber\\ \gamma \left [ b^- \rightarrow \bar f_{di } k^- \right ] % /(a_{i}^2 b^2 ) & \propto & \tilde{\epsilon}^2 + \epsilon_i^2 - 2 \tilde{\epsilon } \epsilon_i \cos ( \gamma - \delta_b + \delta_{di } ) , \label{adsdecays}\end{aligned}\ ] ] which may be solved for the four unknowns : @xmath51 , @xmath0 , @xmath154 , and @xmath155 . for each final state , @xmath156 or @xmath157 , one can use the analogue of eq .
( [ trigno - exercise ] ) to obtain @xmath20 and @xmath158 .
these expressions , of course , depend on the unknown @xmath51 , which is determined ( up to discrete ambiguities ) by requiring that the expressions for @xmath20 found for @xmath156 and for @xmath157 match . here ,
although the interference terms contribute at order @xmath159 , the other terms are @xmath160 , and @xmath161 . as a result ,
the interference is of order one , and the corresponding triangles in the complex amplitude plane are no longer squashed .
soffer @xcite has proposed to maximize the analyzing power of the analysis by combining the gw and ads methods , allowing each of them to contribute the information it is most sensitive to . in this scheme ,
one measures @xmath51 , @xmath49 , @xmath48 , and @xmath0 by minimizing the function @xmath162 @xmath163 compares the measured integrated decay rates @xmath164 in both the gw and ads methods , with their theoretical expectations , @xmath165 .
@xmath165 are functions of @xmath51 , @xmath48 , @xmath0 , and @xmath49 , as given by eqs .
( [ gwdecays ] ) ( for @xmath166 ) and ( [ adsdecays ] ) ( for @xmath167 ) .
the comparison is done with respect to the measurement uncertainties , @xmath168 .
this analysis has several advantages over the individual gw and ads methods , and , therefore , it is most likely to be used in the actual experiment .
first , combining the relatively high - statistics gw modes with the small ads decay rates improves the measurement sensitivity , due to the addition of independent data .
second , a single ads mode is enough to measure all four unknowns , leaving other modes to add redundancy and statistics .
third , this analysis is useful in removing some discrete ambiguities , as discussed below .
we will now discuss the discrete ambiguities involved in the determination of @xmath0 .
we start by recalling that eq .
( [ trigno - exercise ] ) determines @xmath0 up to an eight - fold ambiguity @xcite .
a two - fold ambiguity arises from the fact that we know the signs of the cosines of @xmath169 and @xmath170 , but not the signs of the corresponding sine functions .
physically , this amounts to a confusion between @xmath0 and @xmath48 .
a further four - fold ambiguity arises from the fact that a determination of @xmath20 only determines the angle @xmath0 up to the four fold ambiguity @xmath0 , @xmath171 , @xmath172 , and @xmath173 .
another way to interpret this ambiguity is to notice that eqs .
( [ gwdecays ] ) are invariant under the three independent transformations @xcite @xmath174 these discrete ambiguities are present in both the gw and ads decay rates@xcite .
the @xmath175 symmetry is @xmath176 for the gw decay rates , but @xmath177 for the ads rates .
thus , by combining the two methods , the @xmath175 ambiguity can be resolved with a single ads mode , for which @xmath49 is large enough .
large values of @xmath49 are expected in the sm . by contrast , resolving the @xmath175 ambiguity in the gw method , requires that @xmath48 vary significantly from one @xmath58 decay mode to the other .
this is unlikely , because the experimental limits on cp - conserving phases in @xmath178 , @xmath179 , @xmath180 and @xmath181 @xcite suggest that the @xmath48 are small .
the @xmath182 and @xmath183 ambiguities are likely to degrade the value of the measurement of @xmath0 in a non - discrete manner .
this is due to the fact that , for @xmath0 within the currently allowed region , eq .
( [ smboundgamma ] ) , @xmath184 tends to be close enough to @xmath0 to make the two values indistinguishable within the experimental errors .
overall , this results in a broad dip in the @xmath163 of eq .
( [ eq : chi2 ] ) as a function of the measured value of @xmath0 , which is likely to substantially increase the measurement uncertainty @xcite .
it is therefore instructive to notice that the term @xmath185 of eq .
( [ master ] ) breaks the @xmath182 symmetry . as a result
, one could naively expect that , incorporating mixing into the analysis would resolve the @xmath182 ambiguity and hence the @xmath186 ambiguity , bringing about significant improvement over the no - mixing case .
unfortunately , the @xmath182-breaking term is too small to have a significant effect with foreseeable data sample sizes .
this term vanishes in the large gw modes , in which @xmath187 , and it accounts for about 10% of the rate in the ads modes when @xmath188 .
it is estimated @xcite that with 600 fb@xmath189 collected at a next - generation b factory , the number of events in the ads modes may be as large as 130 .
the @xmath182 violating term thus accounts for some 13 events , and its contribution to the @xmath163 is only about @xmath190 , even when background is neglected .
therefore , we conclude that 60 times more data are needed to effectively resolve the @xmath182 ambiguity at the @xmath191 level . in the presence of @xmath2 mixing
, the @xmath192 decay rates involved in the gw method are altered according to eq .
( [ master ] ) .
we find , @xmath137 % /(a_{\rm
cp}^2 b^2 ) & \propto & 1 + \tilde{\epsilon}^2 + 2 \eta_f\ , \tilde{\epsilon } \cos ( \gamma + \delta_b ) \nonumber\\ & & - \eta_f\ ,
x_d \sin(2 \theta_d ) - y_d \left [ 2 \tilde{\epsilon } \cos{(\gamma + 2 \theta_d + \delta_b ) } + \eta_f\ , \cos{(2 \theta_d ) } \right ] , \nonumber\\ \gamma \left [ b^- \rightarrow f_{\rm cp } k^- \right ] % /(a_{\rm
cp}^2 b^2 ) & \propto & 1 + \tilde{\epsilon}^2 + 2 \eta_f\ , \tilde{\epsilon } \cos ( \gamma - \delta_b ) \nonumber\\ & & + \eta_f\ ,
x_d \sin(2 \theta_d ) - y_d \left [ 2 \tilde{\epsilon } \cos{(\gamma + 2 \theta_d - \delta_b ) } + \eta_f\ , \cos{(2 \theta_d ) } \right ] , \label{gwwmix}\end{aligned}\ ] ] to linear order in @xmath3 and @xmath4 . recall that @xmath193 is the cp eigenvalue of the cp eigenstate @xmath22 . in these expressions
we have effectively dropped the last line of eq .
( [ master1 ] ) , because the corresponding decay chain is suppressed both by @xmath51 and by @xmath2 mixing .
moreover , since we are only keeping linear terms in @xmath3 and @xmath4 , the second line of eq .
( [ master ] ) does not contribute .
these approximations were made in order to simplify the expressions .
we stress that we have used the full expressions in our analysis and in all the computer simulations .
the crucial step in the gw method is the identification of the interference terms on the first lines of eqs .
( [ gwwmix ] ) : @xmath194 .
these are identified with @xmath195 in eqs .
( [ c+- ] ) , which are then used in eq .
( [ trigno - exercise ] ) . in the gw method
this interference is of order @xmath51 ( although it is buried in an overall decay rate of order one ) , while the leading mixing contributions are of order @xmath3 and @xmath4 .
taking @xmath96 , we expect the corrections to the determination of @xmath20 to be of order @xmath196 .
it is clear that the importance of the mixing terms depends on the exact values of @xmath0 and @xmath48 .
therefore , they could be much larger than the previous naive estimate might lead us to believe .
similar considerations apply to the ads method . as long as only upper bounds on @xmath2 mixing
are known , these effects constitute systematic uncertainties that must be added to any other experimental uncertainties .
if there is @xmath2 mixing , but this is ignored in the experimental analysis , then the @xmath195 extracted from the gw method become @xmath197 , \nonumber\\ c_{-w } & = & \cos ( \gamma - \delta_b ) + \frac{x_d}{2 \tilde{\epsilon } } \sin(2 \theta_d ) - \frac{y_d}{2 \tilde{\epsilon } } \left [ 2 \tilde{\epsilon } \eta_f\ , \cos{(\gamma + 2 \theta_d - \delta_b ) } + \cos{(2 \theta_d ) } \right ] , \label{c+-gwwrong}\end{aligned}\ ] ] to linear order in @xmath3 and @xmath4 .
the @xmath3 terms are only important if there is also a large new cp - violating phase @xmath86 in the @xmath2 mixing . in contrast , the second @xmath4 term is important even if @xmath86 is small .
this is closely related to the fact that the linear term in the time - dependent expression for the direct @xmath139 decays is proportional to @xmath198 @xcite . when eqs .
( [ c+-gwwrong ] ) are substituted for the @xmath199 and @xmath200 in eq .
( [ trigno - exercise ] ) , one obtains incorrect values for @xmath201 , where @xmath202 stands for the ` wrong ' value of @xmath0 . notice that the cp - odd term proportional to @xmath3 does not involve @xmath203 .
however , as shown in appendix [ app : b ] , when eqs .
( [ c+-gwwrong ] ) are substituted into eq .
( [ trigno - exercise ] ) , the cp - odd contributions induce an error in the determination of @xmath20 which is always proportional to @xmath204 . as a result ,
the corrections due to @xmath4 tend to be more important than those due to @xmath3 , in the small @xmath48 limit .
these properties are illustrated in figs .
[ fig:2a ] , [ fig:2b ] , and [ fig:2c ] , where we probe the effects due to @xmath205 , @xmath206 , and a combination of both , respectively .
we have taken @xmath207 , which was chosen to allow a comparison with the results in ref .
@xcite , and we have used the definition @xmath208 for the difference between the ` wrong ' value and the ` correct ' value of @xmath20 . in fig .
[ fig:2a ] we illustrate the effect of @xmath3 by taking @xmath209 , @xmath210 , @xmath211 , and @xmath212 . for a true value of @xmath213
, we find a correction of order @xmath214 ( 20% ) .
of the two possible signs in eq .
( [ trigno - exercise ] ) , the plus sign , corresponding to the solution plotted in the figure , gives the correct value of @xmath20 , in the limit of vanishing @xmath3 . in fig .
[ fig:2b ] we illustrate the effect of @xmath4 by taking @xmath215 , @xmath216 , @xmath217 , and @xmath218 .
the two curves shown correspond to the two possible signs in eq .
( [ trigno - exercise ] ) .
the solid ( dashed ) curve corresponds to the plus ( minus ) sign . for each value of @xmath0 ,
the line closest to the horizontal axis provides us with the correct value for the corrections due to mixing ( the other solution is included in the discrete ambiguities ) .
the correct value for @xmath20 is best approximated by taking the minus sign in eq .
( [ trigno - exercise ] ) for @xmath219 , while it is best approximated by taking the plus sign in eq .
( [ trigno - exercise ] ) for @xmath220 .
therefore , the maximal value @xmath221 occurs precisely for @xmath222 .
this is a correction of order 75% , with either sign .
this example illustrates the possibility that a value of @xmath0 outside the sm allowed range might yield a value of @xmath202 inside the sm allowed region , due to the @xmath2 mixing effects . in fig .
[ fig:2c ] we take @xmath209 , @xmath210 , @xmath217 , and @xmath212 .
the combination of the two effects brings the maximum of @xmath223 up to higher values of @xmath20 , but the overall correction gets suppressed due to a partial cancellation between the @xmath3 and @xmath4 terms .
for example , for @xmath224 , we find a correction of order @xmath225 ( 5% ) .
the exact form of this figure depends very sensitively on the precise value of @xmath86 .
as we take @xmath86 from @xmath226 to @xmath227 , the figures start by looking like fig [ fig:2b ] , and end up looking like fig [ fig:2a ] , because the @xmath3 ( @xmath4 ) effect is proportional to @xmath228 ( @xmath229 ) .
for some values of @xmath230 the @xmath3 and @xmath4 effects come in with the same sign and there is no partial cancellation ; the figures become similar to fig [ fig:2c ] , but with larger values of @xmath223 .
[ fig:2a ] , [ fig:2b ] , and [ fig:2c ] , share some important features : * the solution for @xmath20 is found by using either sign in the first of eq .
( [ trigno - exercise ] ) .
in fact , this is the origin of the @xmath231 discrete ambiguity in the gw method . *
the effects of @xmath4 on the gw method tend to be larger that those due to @xmath3 , when @xmath48 is very small .
as shown in appendix [ app : b ] , this is the result of a @xmath204 suppression imposed by the inversion procedure .
however , we should point out that this holds only under the assumption that @xmath51 has been miraculously measured somehow . as shown by meca and silva @xcite ,
if one were to measure @xmath51 by tagging the @xmath33 meson in the final state through its semileptonic decay , then the @xmath3 effect would come into the extraction of @xmath51 without any @xmath204 suppression , and would be as large as the @xmath4 effects . in any case , both effects are sizeable .
* the mixing effects may take values of @xmath0 which are outside ( inside ) the sm allowed region and yield values of @xmath202 which are inside ( outside ) that region . in the first case ,
the mixing effects hide the new physics . in the second case
, they give a signal for new physics when , in reality , there is none . *
we see from figs .
[ fig:2a ] , [ fig:2b ] , and [ fig:2c ] , that , in the presence of the mixing corrections , not all values of @xmath0 survive the square root used in the inversion procedure of eq .
( [ trigno - exercise ] ) . indeed , for many values of @xmath3 , @xmath4 , and @xmath86 there is a range of values of @xmath0 for which @xmath232 or @xmath233 .
the worst case occurs when both @xmath232 _ and _ @xmath233 . in that case
, the presence of the mixing will go undetected in the inversion procedure , unless one is independently checking whether indeed @xmath234 .
* in all cases @xmath235 when @xmath236 and @xmath48 is small
. this effect is explained in appendix [ app : b ] .
using eq .
( [ master ] ) we obtain the @xmath105 decay rates relevant for the ads method in the presence of @xmath2 mixing , as @xmath153 %
/(a_{i}^2 b^2 ) & \propto & \tilde{\epsilon}^2 + \epsilon_i^2 - 2 \epsilon_i \tilde{\epsilon } \cos ( \gamma + \delta_b - \delta_{di } ) \nonumber\\ & & - x_d \left [ \tilde{\epsilon } ( 1 - \epsilon_i^2 ) \sin ( \gamma + 2 \theta_d + \delta_b ) - \epsilon_i \sin ( 2 \theta_d + \delta_{di } ) \right ] \nonumber\\ & & -y_d \left [ \tilde{\epsilon } \cos ( \gamma + 2 \theta_d + \delta_b ) - \epsilon_i \cos ( 2 \theta_d + \delta_{di } ) \right ] , \nonumber\\ \gamma \left [ b^- \rightarrow \bar f_{di } k^- \right ] % /(a_{i}^2 b^2 ) & \propto & \tilde{\epsilon}^2 + \epsilon_i^2 - 2 \epsilon_i \tilde{\epsilon } \cos ( \gamma - \delta_b + \delta_{di } ) \nonumber\\ & & + x_d \left [ \tilde{\epsilon } ( 1 - \epsilon_i^2 ) \sin ( \gamma + 2 \theta_d - \delta_b ) - \epsilon_i \sin ( 2 \theta_d - \delta_{di } ) \right ] \nonumber\\ & & - y_d \left [ \tilde{\epsilon } \cos ( \gamma + 2 \theta_d - \delta_b ) - \epsilon_i \cos ( 2 \theta_d - \delta_{di } ) \right ] , \label{adswmix}\end{aligned}\ ] ] to linear order in @xmath3 and @xmath4 .
we have used here the same approximations discussed in connection with eqs .
( [ gwwmix ] ) , and we have also expanded in @xmath237 .
, which was kept in order to allow a clear comparison between eqs .
( [ adswmix ] ) , where @xmath238 , and eqs .
( [ gwwmix ] ) , where @xmath239 and the corresponding terms are absent . ]
the crucial step in the ads methods is the identification of the interference terms @xmath240 on the first lines of eqs .
( [ adswmix ] ) .
these terms are of order @xmath159 , while the mixing effects are of order @xmath241 , @xmath242 , @xmath243 , and @xmath244 . as a result , taking @xmath96 ,
a naive estimate predicts the mixing effects to perturb the extraction of @xmath20 at order @xmath25 , since this is the common estimate for @xmath245 , @xmath246 , @xmath247 , and @xmath248 .
if this effect is ignored in the analysis , then the @xmath195 become @xmath249 there are two new features in eqs .
( [ adswmix ] ) and ( [ c+-adswrong ] ) , which were not present in eqs .
( [ gwwmix ] ) and ( [ c+-gwwrong ] ) .
the first is the presence of a term proportional to @xmath250 . as a result
, the @xmath3 contributions no longer require the presence of a new cp - violating phase in the mixing @xmath86 .
the second new feature is the presence of @xmath251 .
these phases are expected to be large in the sm .
for some specific values of the parameters , the magnitude , and even the sign , of @xmath251 will dramatically enhance the mixing effects .
in particular , there are now effects of @xmath3 on @xmath201 which are not proportional to @xmath204 , but rather to @xmath252 or @xmath253 , both of which may be large .
this property is discussed in detail in appendix [ app : b ] .
the end result is that both @xmath3 and @xmath4 have a similar impact on the ads method . strictly speaking , in the ads method the mixing contributions perturb the extraction of @xmath20 both directly , as discussed above , and indirectly , through their effects in the determination of @xmath51 . in order to obtain a simple estimate and to allow comparison with the gw method
, we will also assume in this section that @xmath254 is given ( in which case the ads method would require only one final state ) .
following ref .
@xcite , we use @xmath207 , @xmath255 , and @xmath256 . the effects on @xmath223
tend to mimic those shown in figs .
[ fig:2a ] , [ fig:2b ] , and [ fig:2c ] , except that now the @xmath3 effects are present even when @xmath257 , and that all effects are typically much larger .
for example , setting @xmath210 , @xmath217 , and @xmath212 , would yield @xmath258 for the all range @xmath259 . here
, there is no cancellation between the @xmath3 and @xmath4 terms .
more interestingly , one can find values of the parameters which exhibit dramatic new features .
the solid ( dashed ) curves in figs .
[ fig:3a ] and [ fig:3b ] correspond to the plus ( minus ) sign in eq .
( [ trigno - exercise ] ) . for each value of @xmath0 ,
the line closest to the horizontal axis provides us with the correct value for the corrections due to mixing ( the other solution is included in the discrete ambiguities ) . in fig .
[ fig:3a ] , we have taken @xmath210 , @xmath211 , @xmath212 , and @xmath260 .
we see that the largest correction has moved into higher values of @xmath20 . introducing
a nonzero @xmath4 would move this maximum even closer to @xmath261 . in fig .
[ fig:3b ] , we have taken @xmath216 , @xmath217 , @xmath212 , and @xmath260 . now the corrections are large and positive for @xmath262 , they are negative for @xmath263 , and they are positive again for @xmath264 .
this peculiar effect is due to the fact that @xmath223 is minimized in different regions of @xmath20 by each of the two possible signs for @xmath201 in eq.([trigno - exercise ] ) .
the maximum deviation is @xmath265 and occurs for @xmath266 .
this maximum occurs for higher values of @xmath20 and , therefore , it is proportionally smaller ( 15% ) .
we can increase this maximum by changing @xmath3 to @xmath267 . in that case , the first @xmath268 portion of fig .
[ fig:3b ] is very suppressed , and it turns negative at @xmath269 .
the maximum deviation would occur at @xmath270 , but would be enhanced to @xmath271 ( a @xmath272 effect ) .
we have not shown ads figures with large effects ( in percentage ) because they do not show new features .
nevertheless , such possibilities do exist .
for example , if we take @xmath273 , @xmath210 , @xmath217 , and @xmath274 , we obtain a figure which is very similar to fig .
[ fig:2b ] , except that large deviations around @xmath275 extend over a much larger domain , going approximately from @xmath276 up to @xmath277 . in this case
, the largest deviation occurs at @xmath278 and is @xmath279 ( a @xmath280 effect ) . to most
closely simulate the actual experiment , we will now combine the ads and gw methods , using eq .
( [ eq : chi2 ] ) .
we have seen above that the gw and ads methods are affected differently by non - trivial mixing .
one may therefore expect that when the two methods are combined , the total effect will be smaller than in the worst - case scenario for either method . to calculate @xmath281 in this scheme , we vary @xmath49 , @xmath48 , @xmath0 , @xmath86 , @xmath3 and @xmath4 over their allowed ranges . for each set of input values , we calculate @xmath282 of eq .
( [ eq : chi2 ] ) using the correct expression , eq .
( [ master ] ) , but conduct the ` wrong ' analysis by using eqs .
( [ gwdecays ] ) and ( [ adsdecays ] ) to calculate @xmath283 and @xmath284 , respectively .
having thus neglected mixing in our analysis , we proceed to minimize eq .
( [ eq : chi2 ] ) to obtain a measurement of @xmath202 .
is taken from eq .
( [ estimate : epsilonk ] ) , but its ` experimental ' output value is then found ( together with the other observables ) when we minimize eq . ( [ eq : chi2 ] ) . ] the resulting distributions of @xmath285 are shown in fig .
[ delta - gamma ] . in fig .
[ delta - gamma]b we restrict @xmath48 to the more likely range @xmath286 . since the actual values of the phases are not known , it is not meaningful to analyze the distributions of fig .
[ delta - gamma ] in detail .
these distributions clearly indicate , however , that neglecting to account for @xmath287 mixing may significantly bias the measurement of @xmath0 . as we have seen before , the measurement of @xmath0 depends on the the precise values of the parameters in the @xmath33 system .
given the current bounds on @xmath3 and @xmath4 , we expect the measurements of @xmath50 and @xmath49 to be rather insensitive to @xmath2 mixing . the measured value of @xmath50 , _ c.f . _
( [ estimate : epsilon ] ) , is already used in the ads method . once @xmath49 is measured , _
_ appendix [ app : c ] , it can also be used as a known input parameter in the fit for @xmath0 .
we have stressed the fact that these decays also depend on @xmath3 , @xmath4 , and @xmath86 .
the more we know about these quantities , the better will be our bound on @xmath0 from the @xmath1 decays .
the decay chains @xmath11 provide a good opportunity to determine the ckm phase @xmath0 .
naturally , information about the @xmath33 decays must be included in the analysis , either as parameters to be determined from the overall fit or as fixed quantities known from other @xmath33 system experiments .
this is well appreciated for the dcs decay parameter @xmath50 and for the strong phase @xmath49 . in this article , we stress the fact that this is also true for the parameters involved in @xmath2 mixing .
this point can not be ( as it often is ) overlooked , because the extraction of @xmath0 hinges on measurements of small quantities . in the gw method
there is a small interference ; in the ads method the decay rates are small . as a result
, mild values of @xmath3 and/or @xmath4 can have an important effect in these methods .
we have shown that dramatic effects are indeed present when the gw and ads methods are used individually .
in general , combining these methods reduces the errors involved , but one does still find deviations of order @xmath288 .
these effects may simulate the presence of new physics by taking values of @xmath0 inside the sm allowed region into values of @xmath202 outside that region .
they may also obscure the presence of new physics by taking new physics values of @xmath0 outside the sm allowed region and yielding @xmath202 inside that region .
the importance of this error in the determination of @xmath0 is made more acute by the discrete ambiguities associated with the gw and ads methods . as a result * the determination of @xmath0 with the decays @xmath289 must be made in connection with the search for mixing in the @xmath2 mixing ; * any uncertainties due to lack of knowledge of @xmath3 , @xmath4 , and @xmath290 must be correctly included as systematic uncertainties to the extraction of @xmath20 .
* once @xmath3 , @xmath4 , and @xmath290 are determined experimentally , they can be included in the analysis as known parameters .
we are indebted to h. n. nelson for elucidating remarks on the cleo measurements of @xmath3 and @xmath4 .
it is a pleasure to thank y. grossman and h. r. quinn for several discussions , suggestions , and for reading this manuscript .
the work of j. p. s. is supported in part by fulbright , instituto cames , and by the portuguese fct , under grant praxis xxi / bpd/20129/99 and contract cern / s / fis/1214/98 . the work of a. s. is supported by department of energy contracts de - ac03 - 76sf00515 and de - fg03 - 93er40788 .
assuming cpt invariance , the mass eigenstates of the @xmath2 system are related to the flavor eigenstates by @xmath291 with @xmath292 and @xmath293 where @xmath294 ( @xmath295-heavy , @xmath296-light ) is positive by definition , @xmath297 , and @xmath298 is the off - diagonal matrix element in the effective time evolution in the @xmath2 space . consider a @xmath14 ( @xmath15 ) meson which is created at time @xmath299 and denote by @xmath300 ( @xmath301 ) the state that it evolves into after a time @xmath302 , measured in the rest frame of the meson @xmath33 .
one finds @xmath303 + g_+(t ) | \overline{d^0 } \rangle , \label{timeevolution}\end{aligned}\ ] ] where @xmath304 @xmath305 , and @xmath306 .
it is useful to trade the mass and width differences for the dimensionless parameters @xmath307 and @xmath308 , where @xmath309 is the average width .
we already know from studies of the direct @xmath33 decays , @xmath310 , that @xmath311 , at the @xmath90 c. l. @xcite .
therefore , we may expand the time - dependent functions as @xmath312 , \nonumber\\ g_-(t ) & \sim & e^{-imt}\ , e^{-\tau/2 } \left [ ( - i x_d - y_d ) \tau/2 + \cdots \right],\end{aligned}\ ] ] where @xmath313 is the ( proper ) time of the @xmath33 evolution , in units of the @xmath33 average width .
the time - dependent decay rates involve @xmath314 \sim e^{- \tau } \left [ 1 + ( y_d^2 - x_d^2 ) \tau^2/4 + \cdots \right ] , \label{gpquad2 } \\*[3 mm ] \frac{e^{- \gamma t}}{2 } \left [ \cosh \frac{\delta \gamma t}{2 } - \cos \left ( \delta m t \right ) \right ] \sim e^{- \tau } \left [ ( y_d^2 + x_d^2 ) \tau^2/4 + \cdots \right ] , \label{gmquad2 } \\*[3 mm ] g_+^\ast(t ) g_-(t ) & = & \frac{1}{4 } \left [ e^{- \gamma_h t } - e^{- \gamma_l t } - 2 i e^{- \gamma t } \sin(\delta m\ , t ) \right ] \sim e^{- \tau } \left [ ( - y_d - i x_d ) \tau/2 + \cdots \right ] .
\label{g+*g- } \\*[3mm]\end{aligned}\ ] ] the time integrated decay rates involve @xmath315 g_- & \equiv & \int_0^{+ \infty } \!|g_-(t)|^2 dt = \frac{1}{2 \gamma } \left ( \frac{1}{1 - y^2 } - \frac{1}{1 + x^2 } \right ) \sim \frac{1}{\gamma } \left ( \frac{y_d^2 + x_d^2}{2 } + \cdots \right ) , \label{g- } \\*[3 mm ] g_{+- } & \equiv & \int_0^{+ \infty } \!g_+^\ast(t ) g_-(t ) dt = \frac{1}{2 \gamma } \left ( \frac{-y}{1 - y^2 } + \frac{-ix}{1 + x^2 } \right ) \sim \frac{1}{\gamma } \left ( \frac{- y_d - i x_d}{2 } + \cdots \right ) . \label{g+-}\end{aligned}\ ] ]
we have seen in eqs .
( [ c+-gwwrong ] ) and ( [ c+-adswrong ] ) that the presence of mixing may affect @xmath316 by @xmath317 where @xmath318 and @xmath319 stand for cp - even and cp - odd corrections , respectively .
the strong phase is given by @xmath320 in the gw method , and by @xmath321 in the ads method .
for example , the corrections to the gw method described in eqs .
( [ c+-gwwrong ] ) , are @xmath322 , \nonumber\\ \delta_o & = & \ - \frac{x_d}{2 \tilde{\epsilon } } \sin(2 \theta_d ) + y_d \eta_f\ , \sin{(\gamma + 2 \theta_d ) } \sin{(\delta_b)}. \label{osdeltasgw}\end{aligned}\ ] ] notice that the cp - odd term proportional to @xmath3 does not involve @xmath203 .
in fact , that term is due to the interference cp violation present in the decay @xmath139 , and it is completely independent of the production mechanism of the neutral @xmath33 meson .
nevertheless , as we will now show , when eqs .
( [ c+-w ] ) are substituted into eq .
( [ trigno - exercise ] ) , the cp - odd contributions to the difference between the wrong value of @xmath201 ( obtained using @xmath323 ) and the correct value of @xmath20 ( obtained using @xmath316 ) , are always proportional to @xmath324 .
we start by noting that @xmath323 have a standard cp - even ( cp - odd ) component given by @xmath325 ( @xmath326 ) , in addition to the mixing component @xmath318 ( @xmath319 ) .
we will denote their sum by @xmath327 and write @xmath328 then @xmath329 \nonumber\\*[1 mm ] & = & \frac{1}{2 } \left [ 1 - c_e^2 + c_o^2 \pm \sqrt{(1 - c_e^2 + c_o^2)^2 - 4 c_o^2 } \right],\end{aligned}\ ] ] and the results depend only on @xmath330 and @xmath331 .
is cp - even ( its signs remains the same under a cp transformation ) .
therefore , cp - even ( @xmath332 ) and cp - odd ( @xmath333 ) contributions to this quantity can not interfere with one another and , moreover , @xmath333 can only contribute in the combination @xmath331 . ] consequently , the mixing contributions to @xmath330 and @xmath331 are either quadratic ( and , thus , much suppressed , although they could dominate for small values of @xmath316 ) , or linear , but appearing only in the combinations @xmath334 and @xmath335 .
now , in the gw method @xmath3 only shows up in @xmath319 , _ c.f . _
( [ osdeltasgw ] ) .
therefore , the mixing contribution to the gw extraction of @xmath201 proportional to @xmath336 is also proportional to @xmath204 and , thus , it vanishes in the @xmath337 limit .
we can also use this discussion to explain why @xmath201 is very similar to @xmath20 for @xmath338 , provided that @xmath203 is small .
indeed , in that case @xmath339 , @xmath340 , and both contributions become small .
the situation is altered in the ads method because there we have a new cp - even phase , @xmath251 , which is expected to be large .
using eqs .
( [ c+-adswrong ] ) we find the leading cp - even and cp - odd contributions to @xmath323 to be @xmath341 respectively . these are added to the standard contributions @xmath342 and @xmath343 , respectively .
the presence of a potentially large @xmath251 has two consequences .
firstly , the cp - odd quantity that exists even in the absence of mixing , is proportional to @xmath252 and can be large .
therefore , the fact that the mixing cp - odd contribution to @xmath201 linear in @xmath319 always appears multiplied by @xmath252 ceases to constitute a suppression factor .
secondly , there is now one @xmath3 contribution to @xmath318 which is proportional to @xmath253 .
this will interfere with the standard cp - even contribution , @xmath342 , as is also unsuppressed in the @xmath337 limit . , but it is proportional to @xmath344 . ]
we proceed to study the measurement of @xmath49 at a charm factory , operating at the @xmath345 resonance . for simplicity
, we will discuss this measurement in the context of the sm .
given the current bounds on @xmath3 and @xmath4 , we expect this measurement to be rather insensitive to @xmath2 mixing . the relation @xmath346
\label{eq : delta_d}\ ] ] is graphically represented in figure [ fig:3770-triangle ] , demonstrating how to obtain @xmath49 from the the cabibbo allowed @xmath33 decay amplitude , @xmath347 , the doubly cabibbo suppressed amplitude , @xmath348 , and their interference , @xmath349 . while @xmath347 and @xmath348 have been measured at cleo @xcite for the @xmath350 mode by using @xmath351 decays to tag the @xmath14 flavor , measuring @xmath352 requires producing @xmath2 pairs in a known coherent state .
it is therefore best to perform all three measurements at the charm factory , canceling many systematic errors in the construction of the triangles of figure [ fig:3770-triangle ] . to measure the amplitude @xmath353 ( @xmath354 ) , one of the @xmath345 daughters is tagged as a @xmath355 ( @xmath356 ) by observing it decay into a cp - odd ( cp - even ) state , such as @xmath357 ( @xmath358 ) .
the other daughter is then @xmath356 ( @xmath355 ) , and the fraction of the time that it is seen decaying into @xmath350 gives the interference amplitude @xmath359 .
we immediately find @xmath360 due to the low statistics tagging scheme of the @xmath361 measurement and the fact that @xmath362 , the error in @xmath363 is dominated by the @xmath361 measurement error .
hence @xmath364 where @xmath365 is the number of events observed in the @xmath352 channels , and we made use of @xmath366 . since the event is fully reconstructed , background is expected to be small , and its contribution to @xmath367 is neglected in this discussion .
@xmath365 is given by @xmath368 where @xmath369 is the number of @xmath370 events , @xmath371 are the cp - eigenstates used for tagging , and @xmath372 is the reconstruction efficiency of the state @xmath373 . taking @xcite @xmath374 , @xmath375 and @xmath376 , and @xmath377 , eq .
( [ eq : cos - delta - err2 ] ) becomes @xmath378 it is expected that in one year the charm factory will collect 10 @xmath379 , or @xmath380
@xcite , resulting in @xmath381 .
thus , @xmath363 can be measured to high precision , even in the presence of background and with relatively modest luminosity .
we note that the same measurement technique can be used with multi - body @xmath14 decays , in which @xmath363 varies over the available phase space . while the relative statistical error in every small region of phase space will be large
, its effect on the measurement of @xmath0 in @xmath382 will be proportionally small .
the total error in @xmath0 due to @xmath383 will be as small as in the two body @xmath350 mode , up to differences in @xmath14 branching fractions , reconstruction efficiencies , and backgrounds .
n. cabibbo , phys .
* 10 * , 531 ( 1963 ) ; m. kobayashi and t. maskawa , prog .
49 * , 652 ( 1973 ) .
cleo collaboration , b. barish _
et al . _ , report no .
cleo conf 97 - 01 , eps 339 .
m. gronau and d. london , phys . lett .
b * 253 * , 483 ( 1991 ) .
m. gronau and d. wyler , phys .
b * 265 * , 172 ( 1991 ) .
d. atwood , i. dunietz , and a. soni , phys .
lett . * 78 * , 3257 ( 1997 ) .
a. soffer , phys .
d * 60 * , 054032 ( 1999 ) . c. c. meca and j. p. silva , phys .
lett . * 81 * , 1377 ( 1998 ) .
see also ref .
@xcite a. amorim , m. g. santos , and j. p. silva , phys .
d * 59 * , 056001 ( 1999 ) .
r. aleksan , b. kayser , and d. london , phys .
lett . * 73 * , 18 ( 1994 ) .
see , for example , g. c. branco , l. lavoura , and j. p. silva , _ cp violation _ ( oxford university press , oxford , 1999 ) , and references therein .
s. bergmann and y. nir , j. high energy phys .
* 9909 * , 031 ( 1999 ) .
a. ali and d. london , eur .
j. c * 9 * , 687 ( 1999 ) .
a. j. buras , lectures given at the 14th lake louise winter institute ( 1999 ) , hep - ex/9905437 .
cleo collaboration , j. p. alexander _ et al .
_ , phys .
lett . * 77 * , 5000 ( 1996 ) . c. caso _ et al .
_ , european physical journal * c3 * , 1 ( 1998 ) , and also the url : http://pdg.lbl.gov .
cleo collaboration , m. athanas _ et al .
_ , phys .
lett . * 80 * , 5493 ( 1998 ) .
j. gronberg , ` @xmath14 mixing at cleo - ii ' , to appear in proc .
1999 division of particles and fields conference , los angeles , ca , u.s.a . ,
jan . 5 - 9 , 1999 . in @xcite the then - current value of 0.0077 was used for eq .
( [ estimate : epsilon ] ) .
h. n. nelson , hep - ex/9909028 , talk presented at 1999 chicago conference on kaon physics ( k 99 ) , chicago , ( 1999 ) .
the bounds obtained in this reference refer to new quantities @xmath384 and @xmath385 , where @xmath386 is the strong phase difference in the @xmath310 decays used in the study .
they find @xmath387 and @xmath388 at the @xmath90 c. l. the bound we quote in eq .
( [ xd - yd : limit ] ) results from the fact that @xmath389 . h. n. nelson , private communication . for
a review see , for example , h. n. nelson , hep - ex/9908021 .
f. bucella , m. lusignoli , and a. pugliese , phys .
b * 379 * , 249 ( 1996 ) .
for example , the prediction for babar can be found in _ the babar physics book _ , edited by p.
f. harrison and h. r. quinn ( slac , stanford , 1998 ) , page 798 .
there a sensitivity estimate for @xmath390 of @xmath391 is quoted for a year running at nominal luminosity .
h. yamamoto , cbx 94 - 14 , hutp-94/a006 ; h. n. nelson , private communication . t. liu , harvard university preprint number hutp-94/e021 , in _ proceedings of the charm 2000 workshop _ , fermilab - conf-94/190 , page 375 , 1994
. l. wolfenstein , phys .
lett . * 75 * , 2460 ( 1995 ) .
g. blaylock , a. seiden , and y. nir , phys .
b * 355 * , 555 ( 1995 ) .
t. e. browder and s. pakvasa , phys
. lett .
b * 383 * , 475 ( 1996 ) .
see also ref .
j. kirkby , in la thuile 1996 , results and perspectives in particle physics , 747 ( 1996 ) . | several methods have been devised to measure the weak phase @xmath0 using decays of the type @xmath1 , where it is assumed that there is no mixing in the @xmath2 system .
however , when using these methods to uncover new physics , one must entertain the real possibility that the measurements are affected by new physics effects in the @xmath2 system .
we show that even values of @xmath3 and/or @xmath4 around @xmath5 can have a significant impact in the measurement of @xmath6 .
we discuss the errors incurred in neglecting this effect , how the effect can be checked , and how to include it in the analysis . |
since the publication of the graph pebbling survey @xcite there has been a great deal of activity in the subject .
there are now over 50 papers by roughly 80 authors in the field ; the web page @xcite maintains a current list of these papers .
many researchers have asked for an updated survey thanks to the new york academy of sciences for providing the opportunity .
the following is based on the talk , `` _ everything you always wanted to know about graph pebbling but were afraid to ask _ '' , an obvious ripoff / homage to one of new york s favorite directors .
we begin by introducing relevant terminology and background on the subject . here , the term _ graph _ refers to a simple graph without loops or multiple edges . for the definitions of other graph theoretical terms
see any standard graph theory text such as @xcite .
the pebbling number of a disconnected graph will be seen to be undefined .
henceforth we will assume all graphs to be connected .
a configuration @xmath0 of pebbles on a graph @xmath1 can be thought of as a function @xmath2 .
the value @xmath3 equals the number of pebbles placed at vertex @xmath4 , and the _ size _ of the configuration is the number @xmath5 of pebbles placed in total on @xmath6 .
a pebbling step along an edge from @xmath7 to @xmath4 reduces by 2 the number of pebbles at @xmath7 and increases by 1 the number of pebbles at @xmath4 .
we say that a vertex @xmath8 can be _ reached _ by @xmath0 if one can repeatedly apply pebbling steps so that , in the resulting configuration @xmath9 , we have @xmath10 ( and @xmath11 for all @xmath4 ) . the _ pebbling number _ , @xmath12 , is defined to be the smallest integer @xmath13 so that any specified _ root _
vertex @xmath14 of @xmath6 can be reached by every configuration @xmath0 of size @xmath13 . a configuration that reaches every vertex
is called _ solvable _ , and _
unsolvable _ otherwise .
the origins of graph pebbling reside in combinatorial number theory and group theory .
a sequence of elements of a finite group @xmath15 is called a _ zero - sum sequence _ if it sums to the identity of @xmath15 . a simple pigeonhole argument ( on the sequence of partial sums )
proves the following theorem .
[ origin ] any sequence of @xmath16 elements of a finite group @xmath15 contains a zero - sum subsequence .
in fact , a subsequence of consecutive terms can be guaranteed by the pigeonhole argument .
furthermore , one can instead stipulate that the zero - sum subsequence has at most @xmath17 terms , where @xmath18 is the exponent of @xmath15 ( i.e. the maximum order of an element of @xmath15 ) , and this is best possible . initiated in 1956 by erds @xcite
, the study of zero - sum sequences has a long history with many important applications in number theory and group theory . in 1961
erds et al .
@xcite proved that every sequence of @xmath19 elements of a cyclic group @xmath15 contains a zero - sum subsequence of length exactly @xmath16 . in 1969 van emde boas and kruyswijk @xcite proved that any sequence of @xmath20 elements of a finite abelian group contains a zero - sum sequence . in 1994
alford et al .
@xcite used this result and modified erds s arguments to prove that there are infinitely many carmichael numbers .
much of the recent study has involved finding davenport s constant @xmath21 , defined to be the smallest @xmath22 such that every sequence of @xmath22 elements contains a zero - sum subsequence @xcite .
there are a wealth of results on this problem @xcite and its variations @xcite , as well as applications to factorization theory @xcite and to graph theory @xcite . in 1989
kleitman and lemke @xcite , and independently chung @xcite , proved the following theorem ( originally stated number - theoretically ) , strengthening theorem [ origin ] .
let @xmath23 denote the cyclic group of order @xmath24 , and let @xmath25 denote the order of an element @xmath26 in the group to which it belongs .
[ cyclic ] for every sequence @xmath27 of @xmath24 elements from @xmath23 there is a zero - sum subsequence @xmath28 such that @xmath29 . for a sequence @xmath30 the sum @xmath31
is known as the _ cross number _ of @xmath30 and is an important invariant in factorization theory . guaranteeing cross number at most 1 strengthens the extension of theorem [ origin ] that @xmath32 , and shows that equality holds if and only if every @xmath33 .
the concept of pebbling in graphs arose from an attempt by lagarias and saks to give an alternative ( and more natural and structural ) proof than that of kleitman and lemke ; it was chung who carried out their idea .
see also @xcite for another extension of this result .
kleitman and lemke then conjectured that theorem [ cyclic ] holds for all finite groups . for a subgroup @xmath34 of @xmath15 ,
call a sequence of elements of @xmath15 an @xmath34-_sum sequence _ if its elements sum to an element of @xmath34 .
in @xcite is proved the following theorem ( the methods of @xcite use graph pebbling ) .
[ subgroupconj ] let @xmath34 be a subgroup of a finite abelian group @xmath15 with @xmath35 .
for every sequence @xmath27 of @xmath24 elements from @xmath15 there is an @xmath34-sum subsequence @xmath28 such that @xmath36 .
the case @xmath37 here gives theorem [ cyclic ] for finite abelian groups , strengthening the van emde boas and kruyswijk result @xcite .
kleitman and lemke also conjectured that theorem [ subgroupconj ] holds for all finite groups , and verified their conjecture for all dihedral groups ( see @xcite ) . for other nonabelian groups ,
it has been shown recently to hold for the nonabelian solvable group of order 21 ( see @xcite ) .
it would be interesting to see whether graph pebbling methods can shed light on the davenport constant for finite abelian groups . in this regard ,
one of the most pressing questions is as follows .
write @xmath38 , where @xmath39
. then @xmath40 is the _ rank _ of @xmath15 .
it is natural to guess that @xmath41 from pigeonhole intuition .
this was conjectured in @xcite and is true by theorem [ cyclic ] for rank 1 groups .
moreover , it was proven in @xcite for rank 2 groups and @xmath42-groups as well .
however , it was proven in @xcite that the conjecture is false for some groups each rank at least 4 .
what remains open is the instance of rank 3 .
[ davenport ] if @xmath15 is a finite abelian group of rank @xmath43 then its davenport constant satisfies @xmath44 .
there are many known results regarding @xmath12 .
if one pebble is placed at each vertex other than the root vertex , @xmath14 , then no pebble can be moved to @xmath14 . also , if @xmath8 is at distance @xmath45 from @xmath14 , and @xmath46 pebbles are placed at @xmath8 , then no pebble can be moved to @xmath14 .
thus we have that @xmath47 .
graphs @xmath6 that satisfy @xmath48 are known as _
class 0 _ graphs , which include the complete graph @xmath49 , the @xmath45-dimensional cube @xmath50 @xcite , complete bipartite graphs @xmath51 @xcite , and many others .
we will say more about such graphs in section [ dcc ] .
any graph @xmath6 with a cut vertex @xmath52 has @xmath53 .
( indeed , let @xmath54 and @xmath55 , where @xmath56 and @xmath57 are two components of @xmath58 .
define the configuration @xmath0 by @xmath59 , @xmath60 and @xmath61 for every other vertex @xmath8 .
then @xmath62 and @xmath0 can not reach @xmath4 . )
the path @xmath63 , the cube @xmath50 @xcite , the petersen graph @xmath64 @xcite , the even cycle @xmath65 @xcite , and the line graph @xmath66 of the complete graph @xmath49 @xcite are examples of graphs @xmath6 that satisfy @xmath67 , while the odd cycle @xmath68 @xcite is an example of a graph not satisfying either lower bound . another standard result is the pebbling number of a tree , which is worked out in @xcite .
regarding upper bounds , it follows immediately from the pigeonhole principle that a graph @xmath6 on @xmath24 vertices with diameter @xmath45 has pebbling number @xmath69 . it would be interesting to find better general bounds on @xmath12 , especially not involving @xmath24 .
the independence number seems not to be useful .
for example , there is no function @xmath26 such that every graph @xmath6 of independence number @xmath70 and diameter @xmath45 has pebbling number @xmath71 .
indeed , we define a family of graphs @xmath72 which satisfy @xmath73 and @xmath74 , but which have pebbling number @xmath75 as @xmath76 .
define @xmath77 , where @xmath78 and the edge set @xmath79 . since @xmath52 is a cut vertex we know by the above comment that @xmath80 .
however , the domination number ( minimum size of a dominating set ) can be useful .
chan and godbole @xcite made the following improvements on the general upper bound .
[ betterupper ] let @xmath81 denote the domination number of @xmath6
. then 1 .
@xmath82 , 2 .
@xmath83 , and 3 . @xmath84 .
the inequalities in parts 1 and 2 are sharp , and the coefficient of @xmath85 in part 3 can be reduced to @xmath86 in the case of perfect domination .
for any two graphs @xmath56 and @xmath57 , we define the _ cartesian product _ @xmath87 to be the graph with vertex set @xmath88 and edge set @xmath89 and @xmath90 or @xmath91 and @xmath92}. the cube @xmath50 can be built recursively from the cartesian product , and chung s result @xcite that @xmath93 ( and more generally theorem [ pathprod ] below ) would follow easily from graham s conjecture , which has generated a great deal of interest .
( graham)[graham ] @xmath94 .
it is worth mentioning that there are some results which verify graham s conjecture . among these
, the conjecture holds for a tree by a tree @xcite , a cycle by a cycle @xcite , and a clique by a graph with the @xmath85-pebbling property @xcite ( see below ) .
recently , feng and kim verified the conjecture for complete bipartite graphs @xcite and for wheels or fans @xcite .
it is also proven in @xcite that the conjecture holds when each @xmath95 is a path .
let @xmath63 be a path with @xmath24 vertices and for @xmath96 let @xmath97 denote the graph @xmath98 .
[ pathprod ] for nonnegative integers @xmath99 , @xmath100 .
the conjecture was also verified recently @xcite for graphs of high minimum degree , using theorem [ conn ] .
[ mindeg ] if @xmath56 and @xmath57 are connected graphs on @xmath24 vertices that satisfy @xmath101 and @xmath102 , then @xmath94 . in particular
, there is a constant @xmath103 so that if @xmath104 then @xmath105 is class 0 .
we will present probabilistic versions of conjecture [ graham ] in sections [ graphthresh ] and [ pebblingthresh ] .
a graph @xmath6 is said to have the _
2-pebbling property _ if two pebbles can be moved to any specified vertex when the initial configuration @xmath0 has size @xmath106 , where @xmath107 is the number of vertices @xmath4 with @xmath108 .
this property is crucial in the proof of theorem [ pathprod ] . until recently , only one graph @xmath109 ( see figure [ lemke ] ) , due to lemke , was known not to have the 2-pebbling property , although a family of related graphs were conjectured in @xcite not to have the property either .
although that conjecture is still unresolved , wang @xcite proved that a slight modification of snevily s graphs yield an infinite family of graphs , none of which have the 2-pebbling property .
also found in @xcite is the conjecture that all bipartite graphs have the 2-pebbling property .
it is possible that the square of the lemke graph might be a counterexample to graham s conjecture .
find @xmath110 .
in order to describe the generalization employed by chung to prove the kleitman - lemke theorem , we need to introduce generalized pebbling . a
_ @xmath111-pebbling step _ in @xmath6 consists of removing @xmath111 pebbles from a vertex @xmath7 , and placing one pebble on a neighbor @xmath4 of @xmath7 .
we say that a pebbling step from @xmath7 to @xmath4 is _ greedy _ ( _ semigreedy _ ) if @xmath112 ( @xmath113 ) , where @xmath14 is the root vertex , and that a graph @xmath6 is _ ( semi ) greedy _ if for any configuration of @xmath12 pebbles on the vertices of @xmath6 we can move a pebble to any specified root vertex @xmath14 , in such a way that each pebbling step is ( semi ) greedy .
let @xmath114 be a product of paths , where @xmath115 . then
each vertex @xmath116 can be represented by a vector @xmath117 , with @xmath118 for each @xmath119 .
let @xmath120 , be the @xmath121 standard basis vector .
denote the vector @xmath122 by * 0*. then two vertices @xmath7 and @xmath4 are adjacent in @xmath97 if and only if @xmath123 for some integer @xmath124 . if @xmath125 , then we may define * q*-_pebbling _ in @xmath97 to be such that each pebbling step from * u * to * v * is a @xmath126-pebbling step whenever @xmath123 .
we denote the * q*-pebbling number of @xmath97 by @xmath127 .
chungs s proof of theorem [ pathprod ] uses the following theorem @xcite . for integers
@xmath128 , @xmath124 , we use @xmath129 as shorthand for the product @xmath130 .
[ genpathprod ] suppose that @xmath129 pebbles are assigned to the vertices of @xmath97 and that the root @xmath131 .
then it is possible to move one pebble to @xmath14 via greedy * q*-pebbling .
in addition , it was shown in @xcite that @xmath132 , and moreover , that @xmath97 is greedy . also in @xcite the following generalization of graham s conjecture to * q*-pebbling was made .
[ gengraham ] @xmath133 . can conjecture [ graham ] be proved in the case when both graphs share certain extra properties such as greediness and tree - solvability ?
we say that a graph @xmath6 is _ tree - solvable _ if , whenever @xmath0 is a configuration on @xmath6 of size @xmath12 , it is possible to solve @xmath0 in such a way that the edges traversed by pebbling steps form an acyclic graph .
there are graphs which are neither greedy nor tree - solvable .
for example , let @xmath134 be the vertices ( in order ) of the cycle @xmath135 , and form @xmath6 from @xmath135 by adjoining new vertices @xmath26 to @xmath136 and @xmath103 , and @xmath137 to @xmath136 and @xmath138 .
it is not difficult to show that @xmath139 . with root @xmath140 and @xmath141
, @xmath0 has no greedy solution ( nor does it have a semi - greedy solution ) . with @xmath142 and @xmath143
, @xmath0 has no tree - solution .
( it is worth noting that @xmath6 is bipartite . )
another illustrative graph is @xmath144 , which is also bipartite .
( @xmath145 is the _ star _ with @xmath24 vertices , also denoted @xmath146 ) .
easily , @xmath147 and @xmath148 ( these numbers also follow from our classification of diameter two graphs below ) .
although each pebbling number is only one more than the number of vertices , @xmath149 is far greater than @xmath150 ; it would be interesting to discover how much greater this gap can be for other graphs .
also , notice that @xmath151 , a strict inequality .
more importantly , as observed by moews @xcite , @xmath6 is not semi - greedy .
indeed , think of @xmath6 as three pages of a book , let @xmath14 be the corner vertex of one of the pages , @xmath52 the farthest corner vertex of another page , @xmath7 , @xmath4 and @xmath8 the three vertices of the third page , and let @xmath152 .
this shows that even semi - greediness is not preserved by the product of two greedy graphs .
probably , the following holds .
[ greedyconj ] almost every graph is greedy and tree - solvable .
because of the lower bound of @xmath153 on the pebbling number @xmath12 , it is natural to try to classify class 0 graphs , or at least give conditions which either guarantee or prohibit class 0 .
this is of course extremely difficult , although some preliminary results have proved quite interesting . as argued above , graphs of connectivity 1 are not class 0 . in @xcite
we find the following theorem .
[ diam2 ] if @xmath154 then @xmath48 or @xmath155 . class 0 graphs of diameter 2 are classified in @xcite .
a particularly crucial graph in the characterization is the graph @xmath6 , built from the bipartite graph @xmath135 by connecting all the vertices of one of the parts of the bipartition to each other .
@xmath6 has connectivity 2 and diameter 2 but @xmath156 , as witnessed by the configuration @xmath157 , where @xmath136 , @xmath158 and the root @xmath14 are independent .
the following corollary to the characterization appears in @xcite .
denote the connectivity of a graph @xmath6 by @xmath159 .
[ 3conn ] if @xmath154 , and @xmath160 then @xmath6 is of class 0 .
from this it follows that almost all graphs ( in the probabilistic sense ) are of class 0 , since almost every graph is 3-connected with diameter 2 .
the following result , conjectured in @xcite , was proved in @xcite .
this result was used to prove a number of other theorems , including theorems [ mindeg ] , [ girth ] and [ class ] .
[ conn ] there is a function @xmath161 such that if @xmath6 is a graph with @xmath73 and @xmath162 then @xmath6 is of class 0 .
moreover , @xmath163 .
an upper bound for diameter 3 graphs was recently obtained in @xcite .
[ diam3 ] if @xmath164 then @xmath165 , which is best possible .
another guarantee for class 0 membership may reside in the following , as yet unexplored , question . [ greedy ] is it true that every greedy graph is of class 0 ?
a nice family of graphs in relation to theorem [ conn ] is the following . for @xmath166 , the _ kneser graph _ ,
@xmath167 , is the graph with vertices
@xmath168 \choose t}$ ] and edges @xmath169 whenever @xmath170 .
the case @xmath171 yields the complete graph @xmath172 and the case @xmath173 and @xmath174 yields the petersen graph @xmath64 , both of which are class 0 .
when @xmath175 and @xmath176 we have @xmath177 . also , it is not difficult to show that @xmath178 in this range , implying that @xmath167 is class 0 by theorem [ 3conn ] .
furthermore , chen and lih @xcite have shown that @xmath167 is connected , edge transitive , and regular of degree @xmath179 . a theorem of lovsz @xcite states that such a graph has connectivity equal to its degree , and thus @xmath180 .
therefore , using theorem [ conn ] , it is not difficult to prove the following .
@xcite[kneser ] for any constant @xmath181 , there is an integer @xmath182 such that , for @xmath183 , @xmath184 and @xmath185 , we have that @xmath167 is class 0 . in the context of graph pebbling ,
the family of kneser graphs is interesting precisely because the graphs become more sparse as @xmath13 decreases toward @xmath186 , so the diameter ( as well as the girth ) increases and yet the connectivity decreases . [ classkneser ] for @xmath187 , is @xmath188 class 0 ? because pachter et al .
@xcite also proved that diameter two graphs have the 2-pebbling property , it is interesting as well to ask whether @xmath167 has the 2-pebbling property when @xmath189 ( i.e. , when its diameter is at least 3 ) .
in addition , one should ask whether graphs of large enough connectivity ( in terms of the diameter ) have the 2-pebbling property .
this would follow from theorem [ conn ] if all class 0 graphs have the 2-pebbling property . regarding conditions which prohibit class 0 membership
, one can easily show that if @xmath190 then @xmath53 .
the following question was asked in @xcite : is there a constant @xmath26 so that @xmath191 implies @xmath53 ?
this question was answered in the negative in @xcite by using theorem [ conn ] , along with a probabilistic ( deletion ) method analogous to erds s construction @xcite of graphs of arbitrarily high girth and chromatic number . [ girth ]
let @xmath192 denote the maximum number @xmath26 such that there exists a class 0 graph @xmath6 on at most @xmath24 vertices with finite @xmath193 .
the for all @xmath194 we have @xmath195 finally , denote by @xmath196 the family of all connected , bipartite graphs on @xmath197 vertices with minimum degree at least @xmath198 .
what can be said of the pebbling numbers of graphs in @xmath196 ?
in particular , the only graph in @xmath199 is @xmath200 , having pebbling number @xmath201 .
the only graph in @xmath202 is @xmath51 , having pebbling number @xmath197 , as first shown in @xcite . also shown in @xcite is the result that if @xmath203 then @xmath6 is of class 0 . in @xcite
we find that if @xmath204 ( and @xmath205 ) then every @xmath206 is of class 0 .
define @xmath207 to be the smallest @xmath198 such that @xmath6 is of class 0 for every @xmath206 . recently in @xcite
was proved the following result .
[ regularbipartite ] for all @xmath208 we have @xmath209 . for all @xmath210 we have @xmath211 . if @xmath6 has some pseudo - random properties then the bound for @xmath212 seems to be significantly smaller . as far as sparse graphs are concerned , a related problem is to find the transitions for @xmath198 at which the pebbling numbers of the graphs in @xmath196 change from exponential to subexponential in @xmath24 , and from superlinear to linear in @xmath24 . in connection with theorem [ regularbipartite ] ,
it is natural to consider the family @xmath213 of connected , minimum degree @xmath198 graphs instead of bipartite graphs . in @xcite
is proved an analogous result .
let @xmath214 be the minimum @xmath198 such that every graph in @xmath213 is class 0 .
[ normal ] for all @xmath215 we have @xmath216 .
for all @xmath217 we have @xmath218 .
in terms of possible applications it may be relevant to consider having to move a pebble to every vertex simultaneously , rather than to just one root . more generally , given a weight function @xmath8 on the vertices of a graph @xmath6 , we define the _ weighted cover pebbling number _ @xmath219 to be the minimum number of pebbles needed to place , after a sequence of pebbling steps , @xmath220 pebbles on vertex @xmath4 , for all @xmath4 , regardless of the initial configuration . the _ cover pebbling number _
@xmath221 regards the case @xmath222 for all @xmath4 .
we say that @xmath8 is _ positive _ if @xmath223 for every vertex @xmath4 . in @xcite , for positive weight functions the weighted cover pebbling number for cliques and trees is found , and it is shown that the ratio between this parameter and the pebbling number can be arbtrarily large , even within the class of trees . given @xmath8
define @xmath224 and @xmath225 .
the results are as follows .
[ wcovpebbk ] for every positive weight function @xmath8 we have @xmath226 . the _ fuse _ @xmath227 is the graph on @xmath228 vertices composed of a path ( or _ wick _ ) on @xmath229 vertices with @xmath230 independent vertices ( _ sparks _ ) incident to one of its endpoints .
[ wcovrat ] for every @xmath24 and @xmath229 we have @xmath231 .
thus , for @xmath232 , we have @xmath233 . given a graph @xmath6
define @xmath234 , where @xmath235 denotes the distance between @xmath7 and @xmath4 .
[ wcovpebbt ] for every positive weight function @xmath8 we have @xmath236 for every tree @xmath237 .
it would be interesting to find the weighted pebbling numbers of other graphs such as cycles , cubes , complete @xmath14-partite graphs and random graphs , among others .
the main technique in proving theorem [ wcovpebbt ] involves showing that the largest @xmath8-unsolvable configuration is _ simple _ ; that is , concentrated on a single vertex .
the same can not be said in general for weight functions that are not positive .
an important question to investigate along these lines is whether , for arbitrary graphs and positive weight functions , the largest @xmath8-unsolvable configuration is always simple . a positive answer verifies the so - called stacking conjecture ( now theorem ) , below . [ stack ] for every graph @xmath6 and positive weight function
@xmath8 we have @xmath238 .
this theorem originally was verified in @xcite for the case of covering @xmath45-dimensional cubes ( @xmath239 ) .
other graphs were considered in @xcite .
theorem [ stack ] was recently proved in general independently in @xcite .
one should also search for classes of graphs for which there is some constant @xmath103 so that every graph @xmath6 in the class satisfies @xmath240 .
for example , the classes of cliques and paths both have @xmath241 .
finally , we define the weighted pebbling number @xmath242 to be the minimum number @xmath212 so that , for every weight function @xmath8 with @xmath243 , every configuration of @xmath212 pebbles is @xmath8-solvable .
if one is concerned only with positive @xmath8 , then such an evaluation on any tree @xmath237 is trivial .
one would compute @xmath244 .
however , nonpositive @xmath8 must be considered , and thus finding weighted cover pebbling numbers for such @xmath8 is critical . instead of looking for the size of the largest unsolvable configuration in a graph , which is essentially the task in finding the pebbling number of the graph ( minus one )
, one could look for the size of the smallest solvable configuration , which is the task in finding what is called the _
optimal _ pebbling number , @xmath245 .
the first results in this direction showed that @xmath246 @xcite .
in @xcite it is proved that @xmath247 for all @xmath6 ( equality also holds for cycles ) , and the lower bound @xmath248 is found in @xcite .
caterpillars , cycles and other graphs have been considered in @xcite , and the following interesting analog of graham s conjecture was recently proven in @xcite . [ optgraham ] for all graphs @xmath6 and @xmath249 we have @xmath250 .
what is most surprising is that , while @xmath251 is known exactly , only @xmath252 is known at present @xcite .
clearly , much work needs to be done in this area .
some of the newest results consider graphs of high minimum degree . the following result of czygrinow appears in @xcite . [ optdeg ] if @xmath6 is a connected graph with @xmath24 vertices and @xmath253 , then @xmath254 .
the result is not known to be sharp , as given by the following result in @xcite .
[ optdeglow ] for all @xmath255 , @xmath256 and @xmath257 , there is a graph @xmath6 with @xmath253 and @xmath258 .
the graphs discovered for this theorem are clever modifications of a blow - up of the vertices of a cycle into cliques .
it is noted that the @xmath259 term disappears when @xmath24 is a multiple of @xmath260 .
it would be interesting to discover the correct asymptotic coefficient of @xmath261 , somewhere between 2.4 and 4 .
the authors also ask if the general upper bound of @xmath262 can be improved because of high minimum degree .
[ nover2 ] is it true that @xmath263 for all graphs @xmath6 with @xmath264 ?
such a bound would be sharp for @xmath265 .
if girth is also considered then one can say more .
let @xmath266 and @xmath267 .
the following theorem of @xcite displays an asymptotic bound of @xmath268 .
[ deggir ] let @xmath269 , @xmath175 and @xmath270 .
then every @xmath24-vertex graph @xmath6 with @xmath253 and @xmath271 satisfies @xmath272 .
the optimal pebbling numbers of linear ( @xmath273 ) , circular ( @xmath274 ) and m " obius ( circular with a twist ) are also determined in @xcite : @xmath275 unless @xmath276 , with @xmath13 as a lower bound always . in terms of a taxi service ,
army unit , or other operational outfit being able to strategically place their limited resources so as to reach any desired location , this line of research may be most fruitful and significant .
many combinatorial optimization parameters have fractional counterparts . graph parameters like matching number , chromatic number @xcite are well known examples ,
as is the dimension of posets @xcite , and the pebbling number is no exception .
let @xmath277 be the minimum @xmath212 such that , for every choice of root @xmath14 and for every configuration of @xmath212 pebbles on @xmath6 , it is possible to move @xmath198 pebbles to @xmath14 . then the fractional pebbling number @xmath278 . for example , @xmath279 .
also , using moews s tree theorem @xcite , it is easy to compute @xmath280 for every tree @xmath237 .
the following general result was recently proven in @xcite .
[ fracpebbthm ] for every graph @xmath6 we have that @xmath281 .
typically , the parameter in question can be formulated in terms of an integer linear program , and the relaxation of the program gives rise to the fractional version of the parameter . in the case of graph pebbling , while it is routine to verify in polynomial time that a particular configuration reaches a particular root , no simple matrix condition exists to capture whether every configuration of a fixed size @xmath212 is solvable for a particular root . thus minimizing @xmath212 as an integer program remains elusive , and
so it is difficult to say at present whether or not a parameter exists that is dual to @xmath282 .
recently , lourdusamy and somasundaram @xcite found a proof that verifies graham s conjecture for the case @xmath283 .
their proof uses linear programming and lends credence to the above ideas while suggesting that further explorations into the connections between linear programming duality and pebbling numbers may be fruitful and worthwhile indeed .
because it is easy to verify a pebbling solution , the question of deciding whether a given configuration @xmath0 reaches a given root @xmath14 is in np .
in fact it has been shown @xcite that this question is at least has hard as np - complete ( see also @xcite ) . [ hard ] for every @xmath284-regular hypergraph @xmath249 , there is a graph @xmath6 , a configuration @xmath0 , and a root @xmath14 , so that deciding if @xmath0 reaches @xmath14 in @xmath6 is at least as hard as deciding if @xmath249 has a perfect matching .
a complexity upper bound is found in @xcite , where it shown that deciding whether @xmath285 is a @xmath286-complete problem .
this means that it is complete for the class of languages computable in polynomial time by conp machines equipped with an oracle for an np - complete language .
they also proved that deciding whether @xmath287 is np - complete , while the same result is proved in @xcite for deciding whether @xmath288 . instead of computing @xmath12 exactly
, one could try to approximate it .
in other words the following question is of interest .
[ comp2 ] does there exist a ( constant / linear / polynomial ) function @xmath289 and an algorithm which , given a graph @xmath6 on @xmath24 vertices , can find a number in polynomial time @xmath290 such that @xmath291 just like in the case of other `` hard '' problems one can design a `` good '' approximation algorithm for special cases of graphs .
in particular , is it possible to design an approximation algorithm for the case when @xmath6 is a dense graph ? perhaps an easier and more natural question is to ask for an algorithm which , given a configuration and a root , will pebble to the root efficiently .
[ comp3 ] does there exist an algorithm which , given a graph @xmath6 , a root @xmath14 , and an @xmath14-solvable configuration @xmath0 , can pebble to @xmath14 in time polynomial in @xmath24 ?
of course , theorem [ hard ] shows that this is in general impossible unless @xmath292 , although the question for appropriately restricted classes of graphs , such as cubes , remains interesting .
note that for some graphs ( like trees ) the greedy algorithm will pebble to @xmath14 efficiently .
however , as noted in section [ pebnum ] , there are graphs which are neither tree - solvable , nor greedy ( nor semi - greedy ) , and so the question of how to pebble is far from obvious . even in the case of greedy graphs like @xmath50
, there are solvable configurations that require decision - making ; that is , some greedy approaches fail while others succeed .
the notion that graphs with very few edges tend to have large pebbling number and graphs with very many edges tend to have small pebbling number can be made precise as follows .
let @xmath293 be the random graph model in which each of the @xmath294 possible edges of a random graph having @xmath24 vertices appears independently with probability @xmath42 . for functions @xmath295 and @xmath26 on the natural numbers we write that @xmath296 ( or @xmath297 ) when @xmath298 as @xmath299 .
let @xmath300 and define @xmath301 ( resp . , @xmath302 ) to be the set of functions @xmath295 for which there are constants @xmath303 such that @xmath304 ( resp . ,
@xmath305 ) whenever @xmath306 .
finally , let @xmath307 .
let @xmath308 be a property of graphs and consider the probability @xmath309 that the random graph @xmath293 has @xmath308 . for large @xmath42 it may be that @xmath310 as @xmath299 , and for small @xmath42 it may be that @xmath311 as @xmath299 .
more precisely , define the _ threshold _ of @xmath308 , @xmath312 , to be the set of functions @xmath212 for which @xmath313 implies that @xmath310 as @xmath299 , and @xmath314 implies that @xmath311 as @xmath299 .
it is not clear that such thresholds exist for arbitrary @xmath308 .
however , we observe that class 0 is a monotone property ( adding edges to a class 0 graph maintains the property ) , and a theorem of bollobs and thomason @xcite states that @xmath312 is nonempty for every monotone @xmath308 . it is well known @xcite that @xmath315 , and since connectedness is required for class 0 , we see that @xmath316 . in @xcite it is noted that @xmath317 . in @xcite theorem [ conn ]
was used to prove the following result .
[ class ] for all @xmath318 , @xmath319 . as a probabilistic version of theorem [ regularbipartite ]
, we might replace the probability space of @xmath293 by @xmath320 , the random graph model in which each of the @xmath321 possible edges of a random bipartite graph having @xmath13 vertices in each bipartition appear independently with probability @xmath42 .
one might be interested in the following problem .
[ randbip ] find the class 0 threshold @xmath322 of the random bipartite graph @xmath320 .
we now reconsider graham s conjecture [ graham ] from a probabilistic point of view .
we would like to say at the very least that it holds almost always .
more precisely , we offer the following .
[ randgraham ] let @xmath323 for any @xmath324
. then @xmath325\rar 1 $ ] as @xmath299 .
if it turns out that the threshold for class 0 is bigger than that for connectivity , then it would be a very interesting to investigate the pebbling numbers of graphs in @xmath293 for @xmath42 in that range .
another important problem , especially in light of the previous conjecture , is the following .
[ 2ppthresh ] find the threshold @xmath326 for the 2-pebbling property of the random graph @xmath293 .
for this section we will fix notation as follows . the vertex set for any graph on @xmath17 vertices will be taken to be @xmath327\}$ ] , where @xmath328=\{0,\ldots , n-1\}$ ] .
that way , any configuration @xmath329 is independent of @xmath330 .
( here we make the distinction that @xmath24 is the index of the graph @xmath330 in a sequence @xmath331 , whereas @xmath332 denotes its number of vertices . )
let @xmath333 denote the sequence of complete graphs , @xmath334 the sequence of paths , and @xmath335 the sequence of @xmath24-dimensional cubes .
let @xmath336\rar \n$ ] denote a configuration on @xmath332 vertices .
let @xmath337 and for fixed @xmath17 consider the probability space @xmath338 of all configurations @xmath339 of size @xmath340 , we denote by @xmath341 the probability that @xmath342 is @xmath330-solvable and let @xmath343 .
we say that @xmath212 is a _ pebbling threshold _ for @xmath344 , and write @xmath345 , if @xmath346 whenever @xmath347 and @xmath348 whenever @xmath349 .
the existence of such thresholds was recently established in @xcite .
[ existthresh ] every graph sequence @xmath344 has nonempty @xmath350 .
the first threshold result is found in @xcite .
the result is merely an unlabeled version of the so - called `` birthday problem '' , in which one finds the probability that 2 of @xmath212 people share the same birthday , assuming @xmath17 days in a year .
[ cliquethresh ] @xmath351 .
the same threshold applies to the sequence of stars ( @xmath352 ) .
it was discovered in @xcite that every graph sequence @xmath344 satisfies @xmath353 , where @xmath354 is meant to signify that @xmath355 for every @xmath356 .
the authors also discovered that @xmath357 when @xmath344 is a sequence of graphs of bounded diameter , that @xmath358 , and that @xmath359 .
surprisingly , the threshold for the sequence of paths has not been determined .
the lower bound found in @xcite was improved in @xcite to @xmath360 for every @xmath361 , while the upper bound of @xmath362 found in @xcite was improved in @xcite to @xmath363 for every @xmath364 .
finally the lower bound was tightened recently in @xcite to nearly match the upper bound .
[ lower ] for any constant @xmath365 , we have @xmath360 . frustratingly , this still leaves room for a wide range of possible threshold functions .
it is interesting that even within the family of trees , the pebbling thresholds can vary so dramatically , as in the case for paths and stars .
diameter seems to be a critical parameter .
it is quite natural to guess that families of graphs with higher pebbling numbers have a higher threshold , but this kind of monotonicity result remains unproven .
[ monothresh ] if @xmath366 for all @xmath24 then @xmath367 .
in fact , this conjecture remains unproven even in the case that @xmath344 and @xmath368 are sequences of trees . moreover , there is some reason to believe the conjecture may be false , since the pebbling number is a worst - case scenario , while the threshold is an average - case scenario .
consider the following .
for a positive integer @xmath212 and a graph @xmath6 denote by @xmath369 the probability that a randomly chosen configuration @xmath0 of size @xmath212 on @xmath6 solvable . then conjecture [ monothresh ]
would follow from the statement that , if @xmath366 then for all @xmath212 we have @xmath370 . unfortunately , although seemingly intuitive , this implication is false . using the class 0 pebbling characterization theorem of @xcite , in @xcite
is found a family of pairs of graphs @xmath371 , one pair for each @xmath372 , for which the implication fails .
however , the implication may yet hold when @xmath330 and @xmath373 are trees .
[ trees ] let @xmath330 and @xmath373 both be trees on @xmath24 vertices so that @xmath366 .
is it true that , for all @xmath212 ( or those @xmath212 `` near '' one of the thresholds ) , we have @xmath370 ? at the heart of such investigations into monotonicity is the following , most natural conjecture . [ spec ]
let @xmath374 be any functions satisfying @xmath375 .
then there is some graph sequence @xmath344 such that @xmath376 .
this conjecture was proven in @xcite in the case that @xmath377 is replaced by @xmath378 .
in fact , the family of fuses ( defined in section [ coverpebbopt ] ) covers this whole range .
what behavior lives above @xmath378 remains unknown .
it is interesting to consider a pebbling threshold version of graham s conjecture . given graph sequences @xmath379 and @xmath344 , define the sequence @xmath380 .
suppose that @xmath381 , @xmath382 , and @xmath383 , where @xmath384 , @xmath385 , and @xmath386 .
[ prodthresh ] is it true that , for @xmath379 , @xmath344 , and @xmath368 as defined above , we have @xmath387 ?
in particular , one can define the sequence of graphs @xmath388 in the obvious way . in @xcite one
finds tight enough bounds on @xmath389 to show that the answer to this question is yes for @xmath390 and @xmath391 .
another important instance is @xmath392 .
boyle @xcite proved that @xmath393 .
this was improved in @xcite , answering question [ prodthresh ] affirmatively for squares of complete graphs .
[ squarecliques ] for @xmath394 we have @xmath395 .
this result is interesting because , by squaring , the graphs become fairly sparse , and yet their structure maintains the low pebbling threshold .
the proof of the result tied the behavior of pebbling in @xmath396 to the existence of large components in various models of random complete bipartite graphs . another interesting related sequence to consider is @xmath397 . when @xmath398 we have @xmath399 , and the best result to date is the following theorem of @xcite ( obtained independently in @xcite ) . [ cubes ] for the sequence of cubes we have @xmath400 for all @xmath401 .
let @xmath402 and @xmath403 and denote by @xmath404 the graph sequence @xmath405 , where @xmath406 .
most likely , fixed @xmath229 yields similar behavior to theorem [ cubes ]
. [ grids ] for fixed @xmath229 we have @xmath407 .
in contrast , the results of @xcite show that @xmath408 for fixed @xmath45 .
thus it is reasonable to believe that there should be some relationship between the two functions @xmath402 and @xmath403 , both of which tend to infinity , for which the sequence @xmath404 has threshold on the order of @xmath17 .
[ gridbalance ] find a function @xmath403 for which @xmath409 .
in particular , how does @xmath45 compare to @xmath24 ?
finally one might consider the behavior of graphs of high minimum degree .
define @xmath410 to be the set of all connected graphs on @xmath24 vertices having minimum degree at least @xmath411 .
let @xmath412 denote any sequence of graphs with each @xmath413 . in @xcite
is proven the following .
[ dense ] for every function @xmath414 , @xmath415 .
in particular , if in addition @xmath416 then @xmath417 .
a very recent incarnation of graph pebbling is introduced in @xcite , in which the _ critical pebbling number _ of a graph @xmath6 is defined . a configuration @xmath0 of pebbles on @xmath6 is _ minimally solvable _ if it is solvable but the removal of any pebble leaves it unsolvable .
while the optimal pebbling number measures the size of the smallest minimally solvable configuration on @xmath6 , the critical pebbling number measures the size of the largest minimally solvable configuration .
another variation is developed in @xcite , combining ideas from sections [ pebnum ] and [ coverpebbopt ] .
the _ support _ of a configuration @xmath0 is the set of vertices @xmath4 having @xmath418 . the _ domination cover pebbling number _
@xmath419 is the minimum number @xmath212 so that from every configuration of @xmath212 pebbles one can reach a configuration whose support is a dominating set of @xmath6 .
their motivation stems from transporting devices from initial positions to eventual positions that allow them to monitor the entire graph . because of the superficial similarity of graph pebbling to other positional games on graphs , like `` cops - and - robbers '' @xcite and `` chip - firing '' @xcite for instance , and the high degree of applicability of many of these games to structural graph theory @xcite and theoretical computer science @xcite
, one should nt neglect the possibility that graph pebbling will have similar impact .
for example , one can think of the loss of a pebble during a pebbling step as a toll or as a loss of information , fuel or electrical charge .
@xmath111-pebbling is one generalization of this rate of loss ; another is simply to choose any fixed rate @xmath70 of loss . in any case , instead of restricting the initial configuration to integer values , let @xmath0 range among all nonnegative reals .
a pebbling step removes weight @xmath52 from one vertex and places weight @xmath420 at an adjacent vertex , for some fixed @xmath421 .
still the objective is to place weight 1 at any prescribed root @xmath14 so that there is enough money , fuel , information , or energy at that location in the network . of course ,
all of the questions raised herein may be asked about this more general @xmath70-_pebbling_. it is conceivable that chip - firing may even come into play as a useful model . for a given graph @xmath6
, one might be able to build an auxiliary graph @xmath249 , so that chip - firing results on @xmath249 can be brought to bear on @xmath12 .
this opens up the theory to questions of the eigenvalues of the laplacian of @xmath6 , and so on .
999 w. r. alford , a. granville and c. pomerance , there are infinitely many carmichael numbers , _ annals math .
ser . 2 _ * 139 * , no . 3 ( 1994 ) , 703722 . n. alon , _ personal communication _ ( 2003 )
. n. alon , s. friedland and g. kalai , regular subgraphs of almost regular graphs , _
j. combin .
theory ser .
b _ * 37 * ( 1984 ) , 7991 .
r. anderson , l. lovsz , p. shor , j. spencer , e. tardos and s. winograd , disks , balls , and walls : analysis of a combinatorial game , _ amer .
monthly _ , * 96 * ( 1989 ) , 481493 . s. arnborg , d. g. corneil and a. proskurowski , complexity of finding embeddings in a @xmath198-tree , _ siam j. algebraic disc
. methods _ * 8 * ( 1987 ) , 277284 .
a. bekmetjev , g. brightwell , a. czygrinow and g. hurlbert , thresholds for families of multisets , with an application to graph pebbling , _ discrete math .
_ * 269 * ( 2003 ) , 2134 .
a. bekmetjev and g. hurlbert , the pebbling threshold of square cliques , _ preprint _
b. bollobs and a. thomason , threshold functions , _ combinatorica _ * 7 * ( 1987 ) , 3538
. j. boyle , thresholds for random distributions on graph sequences with applications to pebbling , _ discrete math .
_ * 259 * ( 2002 ) , 5969 .
g. brightwell and e. scheinerman , fractional dimension of partial orders , _ order _ * 9 * ( 1992 ) , 139158 .
b. bukh , maximum pebbling number of graphs of diameter three , _ preprint _
d. bunde , e. chambers , d. cranston , k. milans and d. west , pebbling and optimally pebbling in graphs , _ preprint _ ( 2005 ) .
y. caro , zero - sum problems a survey , _ discrete math . _
* 152 * ( 1996 ) , 93113 . m. chan and a. godbole , improved pebbling bounds , _ preprint _
s. chapman , on the davenport constant , the cross number , and their application in factorization theory , in : _ zero - dimensional commutative rings , lecture notes in pure appl .
_ * 171 * , marcel dekker , new york , 1995 , 167190 .
chen and k.w .
lih , hamiltonian uniform subset graphs , _ j. combin .
theory ser .
b _ * 42 * ( 1987 ) , 257263 .
f. r. k. chung , pebbling in hypercubes , _
siam j. disc .
_ * 2 * ( 1989 ) , 467472 .
f. r. k. chung and r. ellis , a chip - firing game and dirichlet eigenvalues , _ discrete math .
_ * 257 * ( 2002 ) , 341355 .
t. clarke , pebbling on graphs , _
master s thesis _ ,
arizona st . univ .
t. clarke , r. hochberg and g. hurlbert , pebbling in diameter two graphs and products of paths , _
j. graph th . _
* 25 * ( 1997 ) , 119128 .
b. crull , t. cundif , p. feltman , g. hurlbert , l. pudwell , z. szaniszlo and z. tuza , the cover pebbling number of graphs , _ discrete math . _ * 296 * ( 2005 ) , 1523 .
a. czygrinow , n. eaton , g. hurlbert and p. m. kayll , on pebbling threshold functions for graph sequences , _ discrete math .
_ * 247 * ( 2002 ) , 93105 .
a. czygrinow and g. hurlbert , pebbling in dense graphs , _ austral .
j. combin .
_ , * 29 * ( 2003 ) , 201208 .
a. czygrinow and g. hurlbert , girth , pebbling and grid thresholds ,
_ siam j. discrete math .
_ , to appear .
a. czygrinow and g. hurlbert , on the pebbling threshold of paths and the pebbling threshold spectrum , _ preprint _
a. czygrinow , g. hurlbert , h. kierstead and w. t. trotter , a note on graph pebbling , _ graphs and combin . _
* 18 * ( 2002 ) , 219225 .
a. czygrinow and m. wagner , on the pebbling threshold of the hypercube , _ preprint _
t. denley , on a result of lemke and kleitman , _ combin .
, prob . and comput .
_ * 6 * ( 1997 ) , 3943 . s. elledge and g. hurlbert , an application of graph pebbling to zero - sum sequences in abelian groups , _ integers : elec .
j. number theory _
* 5(1 ) * ( 2005 ) , # a17 .
p. van emde boas and d. kruyswijk , a combinatorial problem on finite abelian groups iii , _
report zw-1969 - 007 _ , math .
centre , amsterdam .
p. erds , on pseudoprimes and carmichael numbers , _ publ .
debrecen _ * 4 * ( 1956 ) , 201206 .
p. erds , applications of probabilistic methods to graph theory , in _ a seminar on graph theory _ ,
holt , rinehart and winston , new york ( 1967 ) , 6061 .
p. erds , a. ginzburg and a. ziv , a theorem in additive number theory , _ bull .
council israel _ * 10f * ( 1961 ) , 4143 .
p. erds and a. rnyi , on random graphs i , _ publ .
debrecen _ * 6 * ( 1959 ) , 290297 .
r. feng and j. y. kim , graham s pebbling conjecture on product of complete bipartite graphs , _ sci .
china ser .
a _ * 44 * ( 2001 ) , 817822 .
r. feng and j. y. kim , pebbling numbers of some graphs , _ sci .
china ser . a _ * 45 * ( 2002 ) , 470478 .
j. foster and h. snevily , the 2-pebbling property and a conjecture of graham s , _ graphs and combin . _ * 16 * ( 2000 ) , 231344 .
t. friedman and c. wyels , optimal pebbling of paths and cycles , _ preprint _
h. l. fu and c. l. shiue , the optimal pebbling number of the complete @xmath13-ary tree , _ discrete math . _
* 222 * ( 2000 ) , 89100 .
h. l. fu and c. l. shiue , the optimal pebbling number of the caterpillar , _ elec .
notes in disc .
_ * 11 * ( 2002 ) .
w. gao , on davenport s constant of finite abelian groups with rank three , _ discrete math . _
* 222 * ( 2000 ) , 111124 .
w. gao and a. geroldinger , zero - sum problems and coverings by proper cosets , _ euro .
j. combin . _
* 24 * ( 2003 ) , 531549 . w. gao and x. jin , weighted sums in finite cyclic groups , _ discrete math . _
* 283 * ( 2004 ) , 243247
. w. gao and r. thangadurai , on the structure of sequences with forbidden zero - sum subsequences , _ colloq .
_ * 98 * ( 2003 ) , 213222 .
j. gardner , a. godbole , a. teguia , a. vuong , n. watson and c. yerger , domination cover pebbling : graph families , _ preprint _
a. geroldinger , on a conjecture of kleitman and lemke , _
j. number theory _ * 44 * ( 1993 ) , 6065 .
a. geroldinger and r. schneider , on davenport s constant , _
j. combin .
theory ser .
a _ * 61 * ( 1992 ) , 147152 . c. gibbons , j. laison and e. paul , critical pebbling numbers of graphs ,
_ preprint _
a. godbole , m. jablonski , j. salzman and a. wierman , an improved upper bound for the pebbling threshold of the @xmath24-path , _ discrete math .
_ * 275 * ( 2004 ) , 367373 .
g. gunda and a. higgins , pebbling on directed graphs , _ elec .
day _ * 1 * ( 2003 ) , 113 .
d. herscovici , graham s pebbling conjecture on products of cycles , _ j. graph theory _ * 42 * ( 2003 ) , 141154 .
d. herscovici and a. higgins , the pebbling number of @xmath283 , _ discrete math .
_ * 187 * ( 1998 ) , 123135 .
g. hurlbert , two pebbling theorems , _ congr .
* 135 * ( 1998 ) , 5563 .
g. hurlbert , a survey of graph pebbling , _ congr .
_ * 139 * ( 1999 ) , 4164 .
g. hurlbert , the graph pebbling page , g. hurlbert and h. kierstead , on the complexity of graph pebbling , _ preprint _
g. hurlbert and b. munyan , cover pebbling hypercubes , _ bull .
_ , to appear .
p. lemke and d. j. kleitman , an addition theorem on the integers modulo @xmath24 , _ j. number th .
_ * 31 * ( 1989 ) , 335345 .
a. lourdusamy and s. somasundaram , pebbling @xmath422 using linear programming , _ j. discrete math .
sci . cryptography _ * 4 * ( 2001 ) , 115
. l. lovsz , _ combinatorial problems and exercises _ , north holland pub . ,
amsterdam , new york , oxford ( 1979 ) .
k. milans and b. clark , the complexity of graph pebbling , _ preprint _ ( 2005 ) .
d. moews , pebbling graphs , _ j. combin
( b ) _ * 55 * ( 1992 ) , 244252 .
d. moews , _ personal communication _ ( 1997 ) .
d. moews , optimally pebbling hypercubes and powers .
_ discrete math .
_ * 190 * ( 1998 ) , 271276 .
m. nathanson , _ additive number theory : inverse problems and the geometry of sumsets _
( graduate texts in mathematics ; 165 ) , springer - verlag , new york , 1996 , 4851 . r. nowakowski and p. winkler ,
vertex - to - vertex pursuit in a graph , _ discrete math . _ * 43 * ( 1983 ) , 235239 . j. olson , a combinatorial problem on finite abelian groups i , _ j. number theory _ * 1 * ( 1969 ) , 810
. l. pachter , h. s. snevily , and b. voxman , on pebbling graphs , _ congr .
_ * 107 * ( 1995 ) , 6580 .
n. robertson and p. d. seymour , graph minors _ iii _ : planar tree - width , _ j. combin .
theory ser .
b _ * 36 * ( 1984 ) , 4964 .
e. scheinerman and d. ullman , _ fractional graph theory : a rational approach to the theory of graphs _ , wiley , new york ( 1997 ) .
p. d. seymour and r. thomas , graph searching and a min - max theorem for tree - width ,
_ j. combin .
theory ser .
b _ * 58 * ( 1993 ) , 2233 .
j. sjostrand , the cover pebbling theorem , _ preprint _ ( 2004 ) . z. sun , unification of zero - sum problems , subset sums and covers of @xmath423 , _ elec .
soc . _ * 9 * ( 2003 ) , 5160 . m. tomova and c. wyels , cover pebbling cycles and certain graph products , _ preprint _ ( 2005 ) .
a. vuong and i. wyckoff , conditions for weighted cover pebbling of graphs , _ preprint _ ( 2004 ) .
m. wagner and m. wokasch , graph pebbling numbers of line graphs , _ preprint _
s. wang , pebbling and graham s conjecture , _ discrete math . _
* 226 * ( 2001 ) , 431438 .
n. watson , the complexity of pebbling and cover pebbling , _ preprint _
n. watson and c. yerger , cover pebbling numbers and bounds for certain families of graphs , _ preprint _ ( 2005 ) .
d. b. west , _ introduction to graph theory _ , prentice - hall , upper saddle river , nj ( 1996 ) . | the subject of graph pebbling has seen dramatic growth recently , both in the number of publications and in the breadth of variations and applications . here
we update the reader on the many developments that have occurred since the original _ survey of graph pebbling _ in 1999 .
c * v * p s # 1#2 |
experimental determination of fission barrier height in mass a @xmath2 200 continues to be a challenging problem .
accurate knowledge of fission barrier height is vital not only to understand the heavy ion induced fusion - fission dynamics and prediction of super heavy elements , but also other areas , such as stellar nucleosynthesis and nuclear energy applications as well . in the actinide region ,
the fission barrier heights are comparable to the neutron separation energies and could be determined accurately from the measured fission excitation functions , which exhibit a characteristic rise at the barrier energy followed by a flat plateau . in the a @xmath2 200 mass region ,
fission barrier heights are much higher than the neutron separation energies .
most of the measurement of fission cross sections in this mass region are performed at energies much higher than the fission barrier , where there are other open channels and a statistical description is essential .
although a number of studies have been made , there are still ambiguities in choosing various input parameters for the statistical model analysis . according to the statistical model of compound nucleus decay , the probabilities of decay to different channels
are governed by the transmission coefficient and relative density of states ( phase space ) .
the nuclear level density depends of the level density parameter ( @xmath3 ) related to the single particle density near the fermi surface and the available thermal energy ( u ) .
the ground state shell corrections in the nuclei around the doubly closed shell nucleus @xmath4pb ( z=82 , n=126 ) are large and its damping with excitation energy has to be incorporated properly in the statistical model analysis .
the nuclear level density of a shell closed nucleus shows same energy dependence at high excitation energy ( @xmath5 40 mev ) as that of nuclei away from shell closure , if the excitation energy is measured from the liquid drop surface , indicating the complete washing out of the shell corrections at those energies @xcite . at intermediate energy
the dependence is phenomenologically described in terms of energy dependent level density parameter approaching asymptotically to the liquid drop value @xcite .
the phenomenological description of gradual damping of the ground state shell correction with excitation energy was obtained by examining the density of neutron resonances situated near the neutron threshold ( @xmath2 8 mev ) .
recently , it has also been studied by measuring evaporation spectra in @xmath4pb region @xcite .
the knowledge about the shell corrections at the saddle point in a @xmath2 200 is obscure .
the heavy - ion induced fission excitation functions are not sensitive to the correlated variation of the fission barrier height and the ratio of the level density parameter at the saddle point to that at the equilibrium deformation ( @xmath6 ) @xcite . however , the pre - fission neutron multiplicity ( @xmath7 ) data is sensitive to this correlated variation and hence it could be used to constrain the statistical model parameters . however , the measured @xmath7 can have dynamical contribution , which should be taken care of
. analysis @xcite of the fission and evaporation residue cross - sections along with pre - fission neutron multiplicities data for @xmath8c+@xmath9pt system required large shell corrections at the saddle point and yielded fission barriers much smaller ( 13.4 mev ) than those ( @xmath2 22 mev ) obtained for same compound nuclei from the analysis of light ion induced reactions . in the analysis ,
the @xmath7 data was corrected for a dynamical emission corresponding to fission delay of 30@xmath10s . in the present article
we report the statistical model analysis for @xmath0+@xmath11bi , @xmath1+@xmath12w , @xmath13pb , @xmath11bi systems using the same prescription for level density and fission barrier as in ref .
the experimental fission excitation functions are taken from ref .
+ @xmath11bi system . ] + @xmath11bi system . ] + @xmath11bi and @xmath1+@xmath14pb systems are compared with statistical model calculations using parameters from ref .
@xcite ( b@xmath15(0 ) = 13.4 mev ) obtained from the fit to the experimental er and fission excitation functions along with the pre - fission neutron multiplicity data for @xmath8c+@xmath9pt system . ]
the lines are to guide the eye . ]
the statistical model analyses have been carried out using the code pace @xcite with a modified prescription for fission barrier and level density @xcite .
the fission barrier is expressed as @xmath16 where @xmath17 , @xmath18 and @xmath19 are scaling factor to the rotating finite range model ( rfrm ) fission barrier @xcite , shell correction at the equilibrium and shell correction at the saddle deformation , respectively .
the shell corrections at the equilibrium deformations are taken from ref .
fermi gas level density formula has been used to calculate level densities at the equilibrium and saddle point deformation .
the excitation energy of the compound nucleus is calculated as @xmath20 where e@xmath21 , q , e@xmath22(j ) and @xmath23 are the energy in the center of mass , q - value for fusion , rotational energy and pairing energy , respectively . the q - value for fusion as well as the particle separation energies for subsequent decays
are calculated using the experimental masses @xcite .
the excitation energy available at the saddle point deformation is taken as @xmath24 .
the damping of the shell correction with excitation energy is taken care by assuming energy dependent level density parameter as @xmath25\ ] ] with @xmath26 and @xmath27 corresponding to the equilibrium deformation and saddle point deformation , respectively .
the asymptotic liquid drop value of the level density parameter at the equilibrium deformation is taken as @xmath28 = a/9 mev@xmath29 .
the asymptotic liquid drop value of the level density parameter at the saddle point deformation ( @xmath30 ) may be different from that of @xmath28 due to difference in nuclear shapes at these two cases .
hence , the ratio @xmath31 have been kept as free parameter to be decided by the fit to the data .
the spin distribution of the compound nucleus is taken as @xmath32}\ ] ] the value of the @xmath33 and @xmath34 are determined from the experimental fusion cross - section ( @xmath35 ) and fission fragment angular anisotropy data .
sum of the @xmath36 cross - sections available in literature @xcite have been taken as fusion cross - section for @xmath0 + @xmath11bi system .
for @xmath1+@xmath14pb system @xmath36 cross - sections are not available .
however , @xmath36 cross - sections for @xmath1 + @xmath11bi system are available @xcite
. statistical model analysis of @xmath1 + @xmath11bi system taking fusion cross - section from the bass systematics @xcite reproduces the experimental @xmath36 cross - sections well .
hence , fusion cross - sections for @xmath1 induced reactions have been estimated using the bass systematics .
the value of @xmath34 = 2 and 3 reproduces the experimental fission fragment angular anisotropy data for @xmath0 and @xmath1 induced reactions , respectively . in fig . 1 and 2 ,
we have compared the statistical model predictions with the experimental @xmath36 cross - sections for @xmath0 and @xmath1 induced reaction on @xmath11bi target .
fusion process leads to bell shaped excitation functions for @xmath36 channels and the presence of pre - equilibrium particle emission gives rise to high energy tails to these distributions . a significant contribution from pre - equilibrium particle emission to the @xmath36 cross - sections
will lead to over estimation of the fusion cross - section .
as can be seen from the figures , the distribution of the @xmath36 cross - sections could be well reproduced , except the high energy tails in the experimental distributions .
we have estimated the contribution of the pre - equilibrium emission to to @xmath36 cross - sections from these high energy tails .
it was found that the pre - equilibrium contribution to the @xmath36 cross - sections is of the order of 10% for excitation energies below @xmath3740 mev and it becomes around 20% at 50 mev for @xmath0 induced reaction .
fission following pre - equilibrium particle emission will be negligible because of the population of the target - like nuclei with much lower excitation energies than that of the compound nucleus . for
the @xmath1 induced reaction , the pre - equilibrium contribution to @xmath36 cross - section is found not to be significant in the energy range considered in the present analysis . for @xmath0 + @xmath11bi .
the values of @xmath38 is varied to get the best fit . ] for @xmath0 + @xmath11bi .
the values of @xmath38 is varied to get the best fit . ]
experimental fission probabilities ( @xmath39 ) for @xmath0 + @xmath11bi and @xmath1 + @xmath14pb systems are compared with the predictions of the statistical model using the parameters ( b@xmath15 = 13.4 mev ) of ref .
@xcite , which fits the fission excitation functions and the @xmath7 data for @xmath8c+@xmath9pt system simultaneously .
the statistical model calculation using the parameters of ref .
@xcite fails to reproduce the shape of the excitation functions .
attempts to fit the excitation functions by varying the values of @xmath19 and @xmath31 results @xmath40 0.5 mev .
since the resulted shell correction at the saddle point ( @xmath19 ) is within the uncertainty of the rfrm prediction and the damping of shell corrections at the saddle point is also not well known , the value of @xmath19 have been assumed to be zero and the values of @xmath17 and @xmath31 have been varied to fit the excitation functions .
initially , we attempted to fit the excitation functions using an energy independent damping factor @xmath41 = 0.054 mev@xmath29 @xcite for damping of the shell correction at the equilibrium deformation .
it could not fit the excitation functions for the entire excitation energy range considered in the analysis .
the ratio of the experimental fission probabilities to the predictions of the statistical model using energy independent damping factor for @xmath0 + @xmath11bi , @xmath1 + @xmath12w , @xmath14pb systems are shown in fig .
while the ratio shows a systematic deviation from unity for the @xmath0 + @xmath11bi , @xmath1 + @xmath14pb ( @xmath18 = -10.6 mev ) systems , no such deviation was observed for the @xmath1 + @xmath12w ( @xmath18 = -1.9 mev ) system . from this observation
it was concluded that the deviation is due to the improper damping of the shell correction with excitation energy .
an energy dependent shell damping factor @xmath42 is found to give better agreement with the experimental data .
we have also studied the sensitivity of the parameters to the excitation functions .
predictions of the statistical model with different values of @xmath31 for b@xmath15 = 21.8 mev are compared with the experimental fission probabilities for @xmath0 + @xmath11bi systems in fig .
it was found that the lower part of the excitation function is not sensitive to the values of the @xmath31 .
predictions of the statistical model with different values of b@xmath15 for @xmath0 + @xmath11bi system are shown in fig .
6 . for each value of the fission barrier ,
the value of the @xmath31 has been varied to obtain the best fit .
the low energy part of the fission excitation functions are found to be very sensitive to the variation of b@xmath15 .
it should be mentioned here that pre - equilibrium particle emission will not have significant effect at the lower part of the excitation function and hence on the extracted fission barrier hight .
and @xmath1 induced reactions are compared with statistical model calculations using the parameters given in the table 1 . ] the results of the statistical model calculations are shown in fig .
best fit parameters are listed in table [ t1 ] .
.[t1 ] relevant statistical model parameters corresponding to best fits .
fission barriers extracted from the present analysis ( b@xmath43 ) are also compared with the macroscopic - microscopic finite - range liquid - drop model @xcite fission barrier ( b@xmath44 ) . [ cols="^,^,^,^,^,^ " , ]
we have carried out statistical model calculations for @xmath0 and @xmath1 induced reactions to determine the fission barrier heights in a@xmath2200 mass region .
sensitivity of relevant parameters have been studied .
while the low energy part of the excitation functions are found to be sensitive to the height of the fission barrier , high energy part of the excitation functions are found to be sensitive to the value of @xmath31 .
effect of pre - equilibrium particle emission on the extracted fission barrier is estimated to be not significant .
the shell correction at the saddle point is found to be not significant .
however , the statistical model calculation without shell correction at the saddle point substantially under predict the measured pre - fission neutron multiplicity data in heavy ion fusion - fission reactions @xcite . further investigation is required to study the statistical nature of these pre - fission neutron and contributions of other sources ( e.g. dynamical emission and near scission emission ) .
the author gratefully acknowledges the suggestions and advice of drs .
s. kailas and s. s. kapoor . | statistical model analysis has been carried out for @xmath0 and @xmath1 induced fission reactions using a consistent description for fission barrier and level density in a @xmath2 200 mass region .
a continuous damping of shell correction with excitation energy have been considered .
extracted fission barriers agree well with the recent microscopic - macroscopic model .
the shell corrections at the saddle point were found to be not significant . |
gravitational lensing is one of the prediction of einstein s general theory of relativity and many samples of images caused by gravitational lensing have been obtained observationally @xcite .
light rays obey null geodesics in curved spacetime and they are deflected by gravitational potential of lens objects . in weak gravitational field with thin lens approximation , a path of a light ray obeys so called lens equation for gravitational lensing and many analysis concerning the gravitational lensing effect are carried out based on this equation .
especially , we can obtain images of source objects by solving the lens equation using a ray tracing method . as a path of light ray
is derived as the high frequency limit of electromagnetic wave , wave effects of gravitational lensing become important when the wavelength is not so much smaller than the size of lens objects and in such a situation , we must take into account of wave effects .
for example , when we consider gravitational wave is scattered by gravitational lens objects , the wave effect gives significant impact on the amplification factor of intensity for waves @xcite .
another example is direct detection of black holes via imaging their shadows @xcite .
the apparent angular size of black hole shadows are so small that their detectability depends on angular resolution of telescopes which is determined by diffraction limit of image formation system .
thus it is important to investigate wave effects on images for successful detection of black hole shadows .
although interference and diffraction of waves by gravitational lensing has been discussed in connection with amplification of waves , a little was discussed about how images by gravitational lensing are obtained based on wave optics . for electromagnetic wave , e. herlt and h.
stephani @xcite discussed the position of images by a spherical gravitational lens evaluating the poynting flux of scattered wave at an observer .
they claimed that there is a disagreement between wave optics and geometrical optics concerning the position of double images of a point source .
but they have not presented complete understanding of image formation . in this paper
, we consider image formation in gravitational lensing using wave optics and aim to understand how images by gravitational lensing are obtained in terms of waves . for this purpose , we adopt the diffraction theory of image formation in wave optics @xcite , which explains image formation in optical systems in terms of diffraction of waves .
we review the basic formalism of gravitational lensing based on wave optics @xcite . in this paper
, we does not consider polarization of waves and treat scalar waves as a model for electromagnetic waves
. is the deflection angle .
@xmath0 is the distance from the observer to the lens object and @xmath1 is the distance from the observer to the source . @xmath2 . ]
let us consider waves propagating under the influence of the gravitational potential of a lens object .
the background metric is assumed to be @xmath3 where @xmath4 is the gravitational potential of the lens object with the condition @xmath5 .
the scalar wave propagation in this curved spacetime is described by the following wave equation : @xmath6 and for a monochromatic wave with the angular frequency @xmath7 , @xmath8 where @xmath9 is the flat space laplacian .
we show the configuration of the gravitational lensing system considering here ( figure [ fig : glens1 ] ) .
the wave is emitted by a point source , scattered by the gravitational potential of the lens object and reaches the observer .
we assume the wave scattering occurs in a small spatial region around the lens object and outside of this region , the wave propagates in a flat space . with the assumptions of the eikonal and the thin lens approximation
, the fresnel - kirchhoff diffraction formula provides the following amplitude of the wave at the observer @xcite @xmath10\ ] ] where @xmath11 is the effective path length ( eikonal ) along a path from the source position @xmath12 to the observer position @xmath13 via a point @xmath14 on the lens plane @xmath15^{1/2 } + \left[(\boldsymbol{\xi}-\boldsymbol{\delta})^2+r_l^2\right]^{1/2 } -\hat\psi(\boldsymbol{\xi } ) \nonumber \\ & \approx \frac{(\boldsymbol{\eta}-\boldsymbol{\delta})^2}{2r_s}+r_s+\frac{r_lr_s}{2r_{ls}}\left ( \frac{\boldsymbol{\xi}-\boldsymbol{\delta}}{r_l}-\frac{\boldsymbol{\eta } -\boldsymbol{\delta}}{r_s}\right)^2-\hat\psi ( \boldsymbol{\xi})\end{aligned}\ ] ] and we assume that @xmath16 and @xmath17 .
a constant @xmath18 represents the intensity of a point source .
the two dimensional gravitational potential is introduced by @xmath19 then the wave amplitude at the observer can be written as @xcite @xmath20 where @xmath21 is the wave amplitude at the observer in the absence of the gravitational potential @xmath22 : @xmath23 , \qquad s_0(\boldsymbol{\eta},\boldsymbol{\delta } ) = \frac{(\boldsymbol{\eta}-\boldsymbol{\delta})^2}{2r_s } + r_s.\ ] ] @xmath24 is the path length along a straight path from @xmath12 to @xmath13 .
the amplification factor @xmath25 is given by the following form of a diffraction integral @xmath26,\\ & \qquad s_1(\boldsymbol{\eta},\boldsymbol{\xi},\boldsymbol{\delta } ) = \frac{r_lr_s}{2r_{ls}}\left(\frac{\boldsymbol{\xi}-\boldsymbol{\delta}}{r_l } -\frac{\boldsymbol{\eta}-\boldsymbol{\delta}}{r_s}\right)^2 -\hat\psi(\boldsymbol{\xi } ) \end{aligned}\ ] ] where @xmath27 is the fermat potential along a path from the source position @xmath12 to the observer position @xmath13 via a point @xmath14 on the lens plane .
the first term in @xmath28 is the difference of the geometric time delay between a straight path from the source to the observer and a deflected path .
the second term is the time delay due to the gravitational potential of the lens object .
now we introduce the following dimensionless variables : @xmath29 where we choose @xmath30 as @xmath31 @xmath32 is the mass of the gravitational source , @xmath30 and @xmath33 represent the einstein radius and the einstein angle , respectively . using these dimensionless variables , @xmath34 , \label{eq : ampf}\\ & \phi_0(\boldsymbol{y},\boldsymbol{d})=\frac{a_0}{r_s } \exp\left[iw\left(\frac{r_{ls}}{2r_l}\left(\boldsymbol{y } -\frac{r_l}{r_{ls}}\boldsymbol{d}\right)^2+\frac{r_{ls}r_{l}}{\xi_0 ^ 2}\right)\right].\end{aligned}\ ] ] in the geometrical optics limit @xmath35 ,
the diffraction integral ( [ eq : ampf ] ) can be evaluated around the stationary points of the phase function in the integrand .
the stationary points are determined by the solution of the following equation : @xmath36 this is the lens equation for gravitational lensing and determines the location of the image @xmath37 for given source position @xmath38 .
as the specific model of gravitational lensing , we consider a point mass as a gravitational source . in this case , the two dimensional gravitational potential is @xmath39 and the deflection angle is given by ] @xmath40 for @xmath41 , the solution of the lens equation ( [ eq : lens ] ) is @xmath42 and represents the einstein ring with the apparent angular radius @xmath33 defined by ( [ eq : einstein ] ) .
we show an example of images obtained as solutions of the lens equation ( [ eq : lens ] ) ( figure [ fig : image1 ] ) . to produce these images ,
we have assumed an extended source with gaussian distribution of intensity .
.,title="fig : " ] .,title="fig : " ] .,title="fig : " ] .,title="fig : " ] the wave property is obtained by evaluating the diffraction integral ( [ eq : ampf ] ) . for a point
mass lens potential ( [ eq : pot ] ) , the integral can be obtained exactly @xmath43 on the observer plane , an interference pattern appears ( figure 3 ) . . ] near @xmath44 , the distance between adjacent fringes of the interference pattern is @xmath45 this fringe pattern is interpreted as interference between double images of a point source by the gravitational lensing .
the question we aim to raise in this paper is how the interference pattern on the observer plane is related to the images of gravitational lensing in the geometrical optics limit .
the wave amplitude on the observer plane does not make the image of the source and we have to transform the wave function to extract images . to answer this question ,
we introduce a `` telescope '' in the gravitational lensing system and simulate observation of a star ( a point source ) using the telescope . with this setup , it is possible to understand how images of a source are formed in the framework of wave optics .
to establish relation between the interference pattern of the wave and the images of the source in the gravitational lensing system , we first consider an image formation system composed of a single convex lens and review how images of source objects appear in the framework of wave optics @xcite .
let us @xmath46 is the incident wave from a point source in front of a thin convex lens and @xmath47 is the transmitted wave by the lens ( figure [ fig : convex1 ] ) .
they are connected by the following relation @xmath48 where @xmath49 is called a lens transformation function .
the action of a convex lens is to modify the phase of the incident wave . for a point source placed at @xmath50 ( front focal point ) ,
the incident wave and the transmitted wave are @xmath51 where we have used @xmath52 assuming @xmath53 .
thus , a convex lens converts a spherical wave front to a plane wave front . using this action of a convex lens for the incident wave and the transmitted wave
, we can demonstrate the image formation by a convex lens in the framework of wave optics .
let us consider the configuration of the lens system shown in figure [ fig : lens - config ] .
we assume the distribution of the source field on the object plane @xmath54 as @xmath55 . using the fresnel - kirchhoff diffraction formula ,
the amplitude of the wave in front of the lens is given by @xmath56 where @xmath57 is the path length from a point on the object plane to a point on the lens plane and we have assumed @xmath58 .
the amplitude of the wave just behind the lens is given by the relation ( [ eq : conv - lens ] ) @xmath59 where @xmath60 is the aperture function of the lens defined by @xmath61 for @xmath62 and @xmath63 for @xmath64 .
@xmath65 represents a radius of the lens . with the assumption @xmath66 ,
the amplitude of the wave on the @xmath67 plane behind the lens is @xmath68 \nonumber\\ \fl & \qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\times \exp\left[i\frac{\omega}{2 } \left(\frac{|\boldsymbol{x}_0|^2}{a}+\frac{|\boldsymbol{x}|^2}{b } \right)\right ] .
\label{eq : phi2}\end{aligned}\ ] ] for a value of @xmath69 satisfying the following relation ( the lens equation for a convex thin lens ) , @xmath70 the wave amplitude becomes @xmath71\exp\left[i\omega\frac{|\boldsymbol{x}_0|^2}{2a}\right ] \nonumber \\
\fl & \propto\int d^2x_0\phi_0(\boldsymbol{x}_0 ) \left(\frac{2j_1(\omega d|\boldsymbol{x}/b+\boldsymbol{x}_0/a| ) } { \omega d|\boldsymbol{x}/b+\boldsymbol{x}_0/a|}\right ) \exp\left[i\omega\frac{|\boldsymbol{x}_0|^2}{2a}\right ] .
\label{eq : phi2bess}\end{aligned}\ ] ] for @xmath72 limit , the bessel function in ( [ eq : phi2bess ] ) becomes the delta function and we obtains the following wave amplitude on @xmath67 : @xmath73 = \phi_0\left(-\frac{a}{b}\boldsymbol{x}\right).\ ] ] thus , a magnified image of the source field appears on the @xmath67 plane .
this reproduces the result of image formation in geometric optics ; we have shown that an inverted images with magnification @xmath74 of a source object appears on @xmath67 satisfying the lens equation ( [ eq : lens - eq1 ] ) .
if we do not take @xmath72 limit , due to the diffraction effect , an image of a point source has finite size on the image plane called the airy disk @xcite .
its size is given by @xmath75 this value determines the resolving power of image formation system . for two point sources at @xmath76 , their separation on the image plane
is @xmath77 . to resolve them
, their separation must be larger than the size of the airy disk : @xmath78 the lefthand side of this inequality is the angular separation of the sources and @xmath79 determines the resolving power of the image formation system . as we have observed that a convex lens can be a device for image formation in wave optics , we combine it with a gravitational lensing system and obtain images by gravitational lensing .
we consider a configuration of the gravitational lens system shown in figure [ fig : setting2 ] and examine how the images of the source object appear using wave optics . as the source object
, we assume a point source of wave .
the amplitude of the wave just in front of a convex lens is @xmath80 this equation is the same as ( [ eq : phi ] ) . after passing through the convex lens , the wave amplitude on the image plane @xmath81
is given by @xmath82\exp\left[\frac{i\omega}{2z_2 } \left(\boldsymbol{\delta}-\boldsymbol{\delta}'\right)^2\right]\ ] ] where @xmath83 denotes the aperture of the convex lens .
using dimensionless variables , the wave amplitude on the image plane is @xmath84 .
\nonumber\end{aligned}\ ] ] if we choose the location of the image plane @xmath81 to satisfy the following `` lens equation '' for a convex lens , @xmath85 then the wave amplitude on the image plane becomes @xmath86 \times e^{i(w/2)g(\boldsymbol{y},\boldsymbol{d})}\ ] ] where @xmath87 .
thus the wave amplitude on the image plane is the fourier transform of the amplification factor @xmath25 which gives the interference fringe pattern . under the geometrical optics
limit @xmath35 , @xmath37 integral in the amplification factor ( [ eq : ampf ] ) can be approximated by the wkb form @xmath88},\ ] ] where @xmath89 is the solution of the lens equation @xmath90 we have assumed that the aperture of the convex lens is sufficiently smaller than the size of the gravitational lensing system and @xmath91 holds .
then , the wave amplitude on the image plane is @xmath92}e^{iwg(\boldsymbol{y},\boldsymbol{d})/2 } \int_{|\boldsymbol{d}'|\leq d_0 } d^2d ' \exp\left[-iw\left(\boldsymbol{x}_*+\frac{r_lr_s}{r_{ls}f}\,\boldsymbol{d}\right ) \cdot\boldsymbol{d}'\right ] \nonumber \\
\fl & = a\,e^{iw\left[\frac{1}{2}(\boldsymbol{x}_*-\boldsymbol{y})^2-\psi(\boldsymbol{x}_*)\right ] } e^{iwg(\boldsymbol{y},\boldsymbol{d})/2}\times 2\pi d_0 ^ 2\ , \frac{j_1(w|\boldsymbol{x}_*+\beta\boldsymbol{d}|d_0)}{w|\boldsymbol{x}_*+\beta \boldsymbol{d}|d_0},\quad\beta\equiv\frac{r_lr_s}{r_{ls}f}. \ ] ] for @xmath93 limit ( large lens aperture limit or high frequency limit ) , we obtains @xmath94\ ] ] and the image of the point source appears at the following location on the image plane determined by the lens equation ( [ eq : lens - eq22 ] ) : @xmath95 ( [ eq : gl1 ] ) and ( [ eq : gl2 ] ) reproduce the same result of image formation in the geometrical optics ( ray tracing ) in terms of the wave optics .
this is what we aim to clarify in this paper .
if the lens equation ( [ eq : lens - eq22 ] ) has multiple solutions @xmath96 , the wave amplitude on the image plane becomes @xmath97 where @xmath98 are constants .
as an example of image formation in a gravitational lensing system using wave optics , we present the wave optical images of a point source by the gravitational lensing of a point mass ( figure 7 ) .
they are obtained by fourier transformation of the amplification factor @xmath25 ( equation ( [ eq : waveimage ] ) ) and teh lens equation for gravitational lensing ( [ eq : lens ] ) has not been used .
this procedure corresponds to image formation by a convex lens .
these images correspond to images obtained by geometric optics ( figure 2 ) .
we can observe wave effect in these images .
( aperture of a convex lens ) , @xmath99.,title="fig : " ] ( aperture of a convex lens ) , @xmath99.,title="fig : " ] ( aperture of a convex lens ) , @xmath99.,title="fig : " ] ( aperture of a convex lens ) , @xmath99.,title="fig : " ] in each images , we can observe concentric interference pattern which is caused by finite size of the lens aperture and this is not intrinsic feature of the gravitational lensing system .
we can also observe radial non - concentric patterns .
they are caused by interference between double images and represent the intrinsic feature of the gravitational lensing system . for @xmath44 case which corresponds to the einstein ring in the geometrical optics
limit , we can observe a bright spot at the center of the ring , which is the result of constructive interference and does not appear in geometric optics .
for sufficiently large values of @xmath100 , the wave amplitude at the observer coincides with the result obtained by geometric optics .
we investigated image formation in gravitational lensing system based on wave optics . instead of using a ray tracing method , we obtained images directly from wave functions at the observer without using a lens equation of gravitational lensing .
for this purpose , we introduced a `` telescope '' with a single convex thin lens , which acts as a fourier transformer for the interference pattern formed at an observer .
the analysis in this paper relates the wave amplitude and images of the gravitational lensing directly . in the geometric optics limit of waves ,
images by lens systems are obtained by a lens equation which determines paths of each light rays . as light rays are trajectories of massless test particles ( photon ) , expressing image in terms of wave
is to express particle motion in terms of waves . as an application and extension of analysis presented in this paper
, we plan to investigate gravitational lensing by a black hole and obtain wave optical images of black holes .
this subject is related to observation of black hole shadows @xcite
. as the apparent angular sizes of black hole shadows are so small , the diffraction effect on images are crucial to resolve black hole shadows in observation using radio interferometer . for sgra@xmath101 , which is the black hole candidate at galactic center
, the apparent angular size of its shadow is estimated to be @xmath102 arc seconds and this value is the largest among black hole candidates . for a sub - mm vlbi with a baseline length @xmath65 , using ( [ eq : resolution ] ) , the condition to resolve the shadow becomes @xmath103 and this requirement shows the possibility to detect the black hole shadow of sgra@xmath101 using the present day technology of vlbi telescope .
thus , analysis of black hole shadows based on wave optics is an important task to evaluate detectability of shadows and determination of black hole parameters via imaging of black holes .
the topic of wave optical image formation in black hole spacetimes belongs to a classical problem of wave scattering in black hole spacetimes @xcite .
as is well known , waves incident to a rotating black hole are amplified by the superradiance @xcite due to dragging of spacetimes .
this effect enables waves to extract the rotation energy of black holes . on the other hand
, it is known that particles can also extract the rotation energy of black holes via so called penrose process . by investigating images of scattered waves by a rotating black hole
, we expect to find out new aspect or interpretation of phenomena associated with superradiance in connection with the penrose process .
this work was supported in part by the jsps grant - in - aid for scientific research ( c ) ( 23540297 ) .
the author thanks all member of `` black hole horizon project meeting '' in which the preliminary version of this paper was presented . | we discuss image formation in gravitational lensing systems using wave optics . applying the fresnel - kirchhoff diffraction formula to waves scattered by a gravitational potential of a lens object ,
we demonstrate how images of source objects are obtained directly from wave functions without using a lens equation for gravitational lensing . _ wave optics ; image formation ; gravitational lens _ |
estimates on mixed norm spaces @xmath0 ( see definition [ definition : mixed norm spaces ] ) already appeared in the works of gagliardo @xcite and nirenberg @xcite , to prove an endpoint case of the classical sobolev embeddings .
however , a more systematic approach to these spaces was first shown explicitly by fournier @xcite .
recall that the sobolev space @xmath1 @xmath2 consists of all functions in @xmath3 whose first - order distributional derivatives also belong to @xmath4 we write @xmath5 for the closure of smooth function , with compact support , in @xmath6 the classical sobolev embedding theorem claims that @xmath7 sobolev @xcite proved this embedding for @xmath8 but his method , based on integral representations , did not work when @xmath9 that case was settled affirmatively by gagliardo @xcite and nirenberg @xcite , who first observed that @xmath10 and then , using an iterated form of hlder s inequality , completed the proof ; i.e. , @xmath11 later , a new approach based on properties of mixed norm spaces was introduced by fournier @xcite and was subsequently developed , via different methods , by various authors , including blei @xcite , milman @xcite , algervik and kolyada @xcite and kolyada @xcite . to be more precise ,
the central part of fournier s work was to prove @xmath12 and then taking into account , he obtained the following improvement of : @xmath13 the embedding is due to poornima @xcite , and it can be also traced in the work of peetre @xcite ( the case @xmath14 , @xmath15 ) , and kerman and pick @xcite , where a characterization of sobolev embeddings in rearrangement invariant ( r.i . ) spaces was obtained .
since the embedding was first proved , many other proofs and different extensions have appeared in the literature . in particular , relations between mixed norm spaces of lorentz spaces were studied in @xcite , where it was shown , for instance , @xmath16 all these works provide us a strong motivation to better understand the embeddings between mixed norm spaces , as well as to provide a characterization of for more general r.i .
spaces .
the paper is organized as follows . a precise definition of r.i . spaces , and also other definitions and properties concerning function spaces to be used throughout the paper can be found in section [ sec : preliminaries ] .
in section [ mixed norm spaces ] we introduce the benedek - panzone spaces and the mixed norm spaces . here , among other things , we find an explicit formula for the peetre @xmath17-functional for the couple of mixed norm spaces @xmath18 in section [ section : embeddings between mixed norm spaces ] , we explore connections between mixed norm spaces .
it is important to observe that , for any @xmath19 @xmath20 ( see lemma [ lema : inclusion the benedek - panzone spaces ] ) . that is , looking at each specific component , the embeddings are trivial .
however , there are examples , for instance , showing that if in we replace benedek - panzone spaces by the global mixed norm spaces , then the corresponding equivalence is not longer true .
this fact illustrates that mixed norm spaces may have a much more complicated structure than benedek - panzone spaces .
therefore , it is natural to analyze embeddings between mixed norm spaces .
motivated by this problem we find necessary and sufficient conditions for the existence of the following types of embeddings : @xmath21 after this discussion our analysis focuses on a particular embedding : @xmath22 to be more specific , theorem [ teo : rango r(x1,loo)->r(x2,l1 ) ] provides a characterization of the smallest mixed norm space of the form @xmath23 in , once the mixed norm space @xmath24 is given .
to finish , we make some comments for the case when @xmath25 is replaced by @xmath26 ( see also @xcite ) . section [ section : fournier embeddings ] is devoted to the study of the embedding for more general r.i . spaces .
in particular , theorem [ theorem : mixta inclusion r(x , loo)->z ] gives necessary and sufficient conditions for the following embedding to hold : @xmath27 a general consequence of theorem [ theorem : mixta inclusion r(x , loo)->z ] is contained in theorem [ theorem : dominio r(x , loo)->z ] , which provides a characterization of the largest space of the form @xmath28 in , once the r.i .
space @xmath29 is given . finally , for a fixed mixed norm space @xmath28 , theorem [ teo : rango r(x , loo)->z ] describes the smallest r.i .
space for which holds .
some remarks about the notation : the measure of the unit ball in @xmath30 will be represented by @xmath31 as usual , we use the symbol @xmath32 to indicate that there exists a universal positive constant @xmath33 , independent of all important parameters , such that @xmath34 .
the equivalence @xmath35 means that @xmath32 and @xmath36 .
finally , the arrow @xmath37 stands for a continuous embedding .
we collect in this section some basic notations and concepts that will be useful in what follows .
let @xmath38 with @xmath39 and let @xmath40 be a finite interval .
we write @xmath41 for the set of all real - valued measurable functions on @xmath42 given @xmath43 its distribution function @xmath44 is defined by @xmath45 ( where @xmath46 denotes the lebesgue measure ) and the decreasing rearrangement @xmath47 of @xmath48 is given by @xmath49 it is easily seen that if @xmath50 is any radial function on @xmath30 of the form @xmath51 for some @xmath43 then @xmath52 as usual , we shall use the notation @xmath53 a basic property of rearrangements is the hardy - littlewood inequality ( cf .
e.g. ( * ? ? ?
* theorem ii.2.2 ) ) , which says : @xmath54 a rearrangement invariant banach function space @xmath55 ( briefly an r.i .
space ) is the collection of all @xmath56 for which @xmath57 where @xmath58 satisfies the following properties : 1 .
@xmath58 is a norm ; 2 . if @xmath59 then @xmath60 3 . if @xmath61 then @xmath62 4 .
@xmath63 5 .
@xmath64 6 . if @xmath65 then @xmath66 given an r.i .
space @xmath67 the set @xmath68 equipped with the norm @xmath69 is called the associate space of @xmath55 .
it turns out that @xmath70 is again an r.i .
space ( * ? ? ?
* theorem i.2.2 ) . the fundamental function of an r.i .
space @xmath55 is given by @xmath71 where @xmath72 and @xmath73 denotes the characteristic function of the set @xmath74 .
it is known ( * ? ? ?
* theorem 5.2 ) that if @xmath55 is an r.i .
space , then @xmath75 the lorentz space @xmath76 consists of all @xmath56 for which the expression @xmath77 is finite .
it is well - known ( * ? ? ?
* theorem ii.5.13 ) that if @xmath55 is an r.i .
space then , @xmath78 a basic tool for working with r.i .
spaces is the hardy - littlewood - plya principle ( cf .
* proposition ii.3.6 ) ) which asserts that if @xmath79 and @xmath80 then @xmath81 and @xmath82 for later purposes , let us recall the luxemburg _ representation theorem _
* theorem ii.4.10 ) .
it says that given an r.i .
space @xmath83 , there exists another r.i .
space @xmath84 such that @xmath85 and in this case @xmath86 next , we define the boyd indices of an r.i . space .
first we introduce the dilation operator : @xmath87 let us recall that the operator @xmath88 is bounded on @xmath89 for every r.i .
space @xmath55 and for every @xmath90 by means of the norm of @xmath88 on @xmath84 , denoted by @xmath91 we define the lower and upper boyd indices of @xmath55 as @xmath92 it is easy to see that @xmath93 basic examples of r.i .
spaces are the lebesgue spaces @xmath94 @xmath95 another important class of r.i .
spaces is provided by the lorentz spaces .
we recall that the lorentz space @xmath96 , with @xmath97 and @xmath98 or @xmath99 is the r.i .
space consisting of all @xmath56 for which the quantity @xmath100^{q}\frac{ds}{s}\bigg)^{1/q } } , & \textnormal{if $ q<\infty,$}\\ \displaystyle{\sup_{0<t<|i|^{n}}t^{1/p}f^{*}(t)},&\textnormal{if $ q=\infty$ } \end{cases}\end{aligned}\ ] ] is finite .
observe that @xmath101 for a comprehensive treatment of r.i .
spaces we refer the reader to @xcite . finally , let us recall some special results from interpolation theory , which we shall need in what follows .
for further information on this topic see @xcite . given a pair of compatible banach spaces @xmath102 ( compatible in the sense that they are continuously embedded into a common hausdorff topological vector space ) , their @xmath17-functional
is defined , for each @xmath103 by @xmath104 the fundamental result concerning the @xmath17-functional is : [ interpolkf ] let @xmath105 and @xmath106 be two compatible pairs of banach spaces and let @xmath107 be a sublinear operator satisfying @xmath108 then , there exists a constant @xmath109 ( depending only on the norms of @xmath107 between @xmath110 and @xmath111 and between @xmath112 and @xmath113 ) such that @xmath114
in what follows and throughout the paper we shall assume @xmath115 our goal in this section is to present some basic properties of mixed norm spaces .
let @xmath116 we write @xmath117 for the point in @xmath118 obtained from a given vector @xmath119 by removing its @xmath120 coordinate .
that is , @xmath121 moreover , for any @xmath43 we use the notation @xmath122 for the function obtained from @xmath123 with @xmath117 fixed .
let us recall that , since @xmath48 is measurable , @xmath124 is also measurable for a.e . @xmath117 . for later purposes ,
let us first enumerate some geometric properties of the projections .
we refer to the book @xcite for basic facts on this topic .
let @xmath74 be any measurable set and let @xmath125 be fixed .
the @xmath126 of @xmath127 is defined as @xmath128 let us emphasize that since @xmath127 is measurable , its @xmath126 is also measurable , for a.e .
the essential projection of @xmath127 onto the hyperplane @xmath129 is defined as @xmath130 an important result is the so - called loomis - whitney inequality @xcite which says that @xmath131 we now recall the benedek - panzone spaces , which were introduced in @xcite . for further information on this topic see @xcite .
[ definition : benedek - panzone ] let @xmath132 given two r.i
. spaces @xmath133 and @xmath134 the benedek - panzone space @xmath135 is defined as the collection of all @xmath56 satisfying @xmath136 where @xmath137 buhvalov @xcite and blozinski @xcite proved that @xmath135 is a banach function space .
moreover boccuto , bukhvalov , and sambucini @xcite proved that @xmath135 is an r.i .
space , if and only if @xmath138 .
now , we shall give the definition of the mixed norm spaces , sometimes also called symmetric mixed norm spaces .
[ definition : mixed norm spaces]given two r.i .
spaces @xmath133 and @xmath134 the mixed norm space @xmath0 is defined as @xmath139 for each @xmath140 we set @xmath141 it is not difficult to verify that @xmath0 is a banach function space . since the pioneering works of gagliardo @xcite , nirenberg @xcite , and fournier @xcite ,
many useful properties and generalizations of these spaces have been studied , via different methods , by various authors , including blei @xcite , milman @xcite , algervik and kolyada @xcite , and kolyada @xcite .
all these works , together with the embedding , provide us a strong motivation to better understand the mixed norm spaces of the form @xmath142 for this , we start with a useful lemma : [ lemma : essential projection property ] let @xmath56 and let @xmath143 with @xmath144 then , @xmath145 to prove this lemma , it is enough to consider @xmath146 let us see that @xmath147 in fact , if @xmath148 then , by definition of @xmath149 we have @xmath150 but @xmath151 and hence , we get @xmath152 as a consequence , we have @xmath153 this proves that holds . to complete the proof
, it only remains to see that @xmath154 in fact , if @xmath155 then @xmath156 so , by definition , @xmath157 therefore , @xmath158 thus , the proof is complete . as an immediate consequence of inequality and lemma [ lemma : essential projection property ]
, we have the following inequality , which was previously proved in @xcite : [ corollary : distribution and essential projection ] let @xmath159 then , for any @xmath160 @xmath161 our next goal is to describe the @xmath17-functional for pairs of the form @xmath18 then , we shall apply it to the characterization of the r.i .
hull of a mixed norm space of the form @xmath28 ( see theorem [ teo : rango r(x , loo)->z ] ) .
we refer to the paper @xcite , for other results concerning interpolation of mixed norm spaces .
we first prove a lower bound for the @xmath17-functional for the couple of mixed norm spaces @xmath162 [ lemma : lower bound for k - functional ] let @xmath133 and @xmath163 be r.i . spaces .
then , @xmath164 we fix @xmath165 and @xmath132 if @xmath166 with @xmath167 and @xmath168 then @xmath169 so , it holds that @xmath170 therefore , we have @xmath171 hence , taking the infimum over all decompositions of @xmath48 of the form @xmath166 with @xmath167 and @xmath168 we obtain @xmath172 for any @xmath173 and @xmath174 consequently , @xmath175 from which the result follows .
[ theorem : k - functional of r(x , loo ) and loo ] let @xmath133 be an r.i . space and let @xmath176 then , @xmath177 where @xmath178 is the fundamental function of @xmath179 in view of lemma [ lemma : lower bound for k - functional ] , we only need to prove @xmath180 for this , we fix any @xmath174 then , we define @xmath181 @xmath182 and @xmath183 let @xmath184 be fixed . then
, for any @xmath185 @xmath186 where @xmath187 thus , for any @xmath188 and @xmath185 @xmath189 now , let us suppose that @xmath190 then , by definition , @xmath191 as a consequence , we have that if @xmath192 then @xmath193 for any @xmath194 therefore , for any @xmath195 it holds that @xmath196 so , lemma [ lemma : essential projection property ] implies that , for any @xmath195 @xmath197 hence , for any @xmath198 we get @xmath199 therefore , we obtain @xmath200 but , by hypothesis , @xmath201 and hence we get @xmath202 therefore , using the above inequality , we obtain @xmath203 but , it holds that @xmath204 and hence , we have @xmath205 thus , the proof is complete . as a consequence
we obtain the following result .
recall that @xmath206 stands for the real interpolation space of the couple @xmath207 ( * ? ? ?
* definition v.1.7 ) : [ corollary : mixtas e interpolacion ] let @xmath133 be an r.i . space .
let @xmath208 and @xmath209 then , @xmath210 with equivalent norms . to prove it , we first need to recall a result concerning the @xmath17-functional of pairs of r.i .
spaces . for further information
see @xcite .
[ theorem : k - functional of r.i .
spaces ] let @xmath55 be an r.i . space .
then , @xmath211 for the sake of simplicity , we prove this result only when @xmath212 let @xmath213 then , by a change of variables , we get @xmath214^{q}dt\\ & = \int^{\infty}_{0}\bigl(\varphi_{x}(s)\bigr)^{-\theta q-1}\bigl[k(f,\varphi_{x}(s ) , \mathcal{r}(x , l^{\infty}),l^{\infty})\bigr]^{q}d\varphi_{x}(s).\end{aligned}\ ] ] hence , using theorem [ theorem : k - functional of r(x , loo ) and loo ] , we obtain @xmath215^{q}ds\\ & \geq \int^{\infty}_{0}\bigl(\varphi_{x}(s)\bigr)^{-\theta q-1}\bigl\|\psi^{*}_{k}(f , l^{\infty})\chi_{(0,s)}\bigr\|^{q}_{\overline{x}(0,|i|^{n-1})}ds,\end{aligned}\ ] ] for any @xmath184 .
so , theorem [ theorem : k - functional of r.i . spaces ] implies that @xmath216^{q}d\varphi_{x}(s)\\ & = \bigl\|\psi_{k}(f , l^{\infty})\bigr\|^{q}_{(x , l^{\infty})_{\theta , q}}=\bigl\|f\bigr\|_{\mathcal{r}_{k}((x , l^{\infty})_{q,\theta},l^{\infty})}.\end{aligned}\ ] ] as a consequence , we get @xmath217 thus , we have seen that the embedding @xmath218 holds .
hence , to complete the proof , it only remains to see that @xmath219 also holds .
to do it , we fix any @xmath220 then , using theorem [ theorem : k - functional of r(x , loo ) and loo ] and the subadditive property of @xmath221 we get @xmath222 so , using theorem [ theorem : k - functional of r.i .
spaces ] , we obtain @xmath223 that is , @xmath224 thus , the proof is complete .
our aim in this section is to characterize certain embeddings between mixed norm spaces . before that , let us emphasize that relations between mixed norm spaces of lorentz spaces were studied in @xcite , where it was shown , for instance , @xmath225 let us start with some preliminary lemmas : [ lema : inclusion the benedek - panzone spaces ] let @xmath226 let @xmath227 @xmath228 @xmath229 and @xmath230 be r.i . spaces .
then , @xmath231 to prove the implication @xmath232 " , we just have to apply the hypothesis to the functions @xmath233 with @xmath234 and @xmath235 .
the converse follows from definition [ definition : benedek - panzone ] .
it is important to observe that there are examples , for instance , showing that if in lemma [ lema : inclusion the benedek - panzone spaces ] we replace the benedek - panzone spaces by mixed norm spaces , then the corresponding equivalence is not longer true .
however , we always have this result : [ lema : inclusion norma mixta ] let @xmath227 @xmath228 @xmath229 and @xmath230 be r.i . spaces
. then , @xmath236 it is immediate from definition [ definition : mixed norm spaces ] and lemma [ lema : inclusion the benedek - panzone spaces ] .
lemma [ lema : inclusion the benedek - panzone spaces ] and lemma [ lema : inclusion norma mixta ] show that it is natural to study when embeddings between mixed norm spaces are true .
motivated by this problem , we shall find necessary and sufficient conditions in the following cases : @xmath21 [ teo : mixta inclusion r(x1,y1)->r(x2,y2)]for any r.i
. spaces @xmath227 @xmath228 @xmath237 and @xmath238 if the following embedding @xmath239 holds , then @xmath240 by lemma [ lema : inclusion norma mixta ] , we may assume , without loss of generality , that the following embedding @xmath241 holds .
also , we shall suppose that @xmath242 with @xmath243 let @xmath244 such that @xmath245 given any function @xmath246 with @xmath247 we define @xmath248 for any @xmath249 , we denote @xmath250 now , if @xmath195 we have @xmath251 thus , @xmath252 and so @xmath253 hence , we get @xmath254 therefore , @xmath255 hence , our assumption on @xmath50 ensures that @xmath256 and @xmath257 thus , using @xmath258 and , we get @xmath259 now , let us compute @xmath260 in order to do it , we fix any @xmath184 and @xmath261 and set @xmath262 as before , if @xmath195 we have @xmath263 but , @xmath264 so we obtain @xmath265 for any @xmath194 as a consequence , if @xmath261 then @xmath266 thus , @xmath267 and so @xmath268 therefore , inequality gives us that @xmath269 for any @xmath270 with @xmath271 now , let us consider any @xmath272 we define @xmath273 and @xmath274 since @xmath275 the inequality , with @xmath50 replaced by @xmath276 implies that @xmath277 thus , combining @xmath278 a.e . with , we get @xmath279 on the other hand , by hlder s inequality , we obtain @xmath280 finally , using and , we get @xmath281 and the proof is complete . as a consequence we have the following corollary : let @xmath228 @xmath237 and @xmath230 be r.i
then , @xmath282 it follows from lemma [ lema : inclusion norma mixta ] and theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] .
another consequence of theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] is the following result : let @xmath283 @xmath284 and either @xmath285 @xmath286 or @xmath287 and @xmath288 then , @xmath289 it follows from theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] and the classical embeddings for lorentz spaces ( see @xcite ) .
let us now study embeddings between mixed norm spaces of the form @xmath142 [ teo : mixta inclusion r(x1,loo)->r(x2,loo ) ] let @xmath290 and @xmath291 be r.i . spaces .
then , @xmath292 in view of lemma [ lema : inclusion norma mixta ] , we only need to prove that if the embedding @xmath293 holds , then @xmath294 as before , we assume that @xmath242 with @xmath295 and @xmath245 given any @xmath296 with @xmath297 we define @xmath298 we fix any @xmath132 then , it holds that @xmath299 and @xmath300 otherwise .
so , for any @xmath19 we have @xmath301 hence , since we are assuming that @xmath296 we obtain @xmath302 and @xmath303 so , using @xmath304 and the previous inequality , we get @xmath305 but , as before , @xmath306 hence , we have @xmath307 this proves that if @xmath296 with @xmath297 then @xmath308 the rest of the proof is essentially the same as in theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] . for a general r.i .
space @xmath134 we have a similar result assuming some conditions on @xmath309 [ theorem : r(x1,y)->r(x2,y ) ] let @xmath290 be an r.i .
space , with @xmath310 and let @xmath291 and @xmath163 be r.i . spaces .
then , @xmath311 as before , according to lemma [ lema : inclusion norma mixta ] , it suffices to prove the necessary part of this result .
also , by theorem [ teo : mixta inclusion r(x1,loo)->r(x2,loo ) ] , we assume that @xmath312 let us suppose that the embedding @xmath313 holds and suppose that @xmath242 with @xmath295 and @xmath245 given any function @xmath296 with @xmath297 we define @xmath314 we fix @xmath184 and @xmath315 using now , we obtain @xmath316 then , fubini s theorem gives @xmath317 hence , using that @xmath318 is an increasing function , we deduce that @xmath319 since @xmath310 ( * ? ? ?
* theorem v.5.15 ) ensures that the integral operator @xmath320 is bounded on @xmath321 and , as a consequence , we obtain @xmath322 for any @xmath132 hence , our assumption on @xmath48 gives that @xmath323 and @xmath324 so , using @xmath325 and , we get @xmath326 we next find a lower estimate for @xmath327 . in fact , we fix @xmath184 and @xmath328 then , by hlder s inequality , we get @xmath329 on the other hand , by a change of variables , it holds that @xmath330 and so fubini s theorem gives @xmath331 hence , using , we obtain @xmath332 therefore , @xmath333 thus , using , we get @xmath334 this proves that if @xmath296 with @xmath297 then @xmath308 the general case can be treated as at the end of the proof of theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] . as a consequence of theorem [ teo : mixta inclusion r(x1,loo)->r(x2,loo ) ] and theorem [ theorem : r(x1,y)->r(x2,y ) ] , we obtain the following result : let @xmath283 @xmath284 and either @xmath285 @xmath286 or @xmath287 and @xmath288 then @xmath335 it follows from theorem [ teo : mixta inclusion r(x1,loo)->r(x2,loo ) ] , theorem [ theorem : r(x1,y)->r(x2,y ) ] and the classical embeddings for lorentz spaces ( see @xcite ) .
finally , let us study the embedding @xmath336 let us start by analyzing the case @xmath337 the following result will be useful for our purposes .
[ lemma : fubini mixed norm ] let @xmath338 be an r.i . space .
then , @xmath339 let @xmath340 and @xmath341 then , using fubini s theorem and hlder s inequality , we get @xmath342 that is , @xmath343 and the proof is complete .
for any couple of r.i .
spaces @xmath344 and @xmath345 , we have @xmath346 using lemma [ lema : inclusion norma mixta ] and lemma [ lemma : fubini mixed norm ] , we get @xmath347 as we wanted to see .
now , let us consider the embedding , for the case @xmath348 . in particular , we shall provide a characterization of the smallest mixed norm space of the form @xmath23 in once the mixed norm space @xmath349 is given .
in order to do it , we begin with a preliminary lemma . for simplicity
, we will assume that @xmath350 .
[ lemma : optimal range norm r(x , loo)->r(y , l1 ) ]
let @xmath133 be an r.i .
space , with @xmath351 .
then , the functional defined by @xmath352 is an r.i . norm .
the positivity and homogeneity of @xmath353 are clear .
next , let @xmath48 and @xmath50 be measurable functions on @xmath42 then , @xmath354 since @xmath355 ( cf .
* theorem ii.3.4 ) ) , hardy - littlewood - plya principle implies that @xmath356 the proof of ( a2)-(a4 ) and ( a6 ) for @xmath353 requires only the corresponding axioms for @xmath357 hence we shall omit them .
finally , to prove property ( a5 ) , we fix any @xmath358 then , @xmath359 [ teo : rango r(x1,loo)->r(x2,l1 ) ] let @xmath360 let @xmath133 be an r.i .
space and let @xmath361 be as in . then , the embedding @xmath362 holds .
moreover , @xmath363 is the smallest space of the form @xmath364 that verifies .
lemma [ lemma : optimal range norm r(x , loo)->r(y , l1 ) ] gives us that @xmath361 is an r.i .
space equipped with the norm @xmath365
now , let us see that the embedding holds .
in fact , if @xmath366 then , combining @xmath367 with the embedding and using theorem [ interpolkf ] , we deduce that @xmath368 hence , using lemma [ lemma : lower bound for k - functional ] and theorem [ theorem : k - functional of r(x , loo ) and loo ] , we get @xmath369 so , using the hardy - littlewood inequality and the subadditivity of @xmath357 we get @xmath370 for any @xmath371 therefore , the embedding @xmath372 holds , as we wanted to see .
now , let us prove that @xmath363 is the smallest r.i .
space for which holds .
that is , let us see that if a mixed norm space @xmath364 satisfies @xmath373 then @xmath374 we assume that @xmath242 with @xmath295 and @xmath245 given any function @xmath375 with @xmath376 we define @xmath377 we fix @xmath378 then , @xmath379 and @xmath300 , otherwise .
thus , for any @xmath380 @xmath381 on the other hand , @xmath382 and @xmath383
so , @xmath384 therefore , using and , we get @xmath385 so , using @xmath386 and , we get @xmath387 but , as before , @xmath388 and hence , using , we get @xmath389 from this , we obtain that any @xmath390 , with @xmath391 , belongs to @xmath392 .
the general case can be proved as in the proof of theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] .
thus , we have @xmath393 , and so , lemma [ lema : inclusion norma mixta ] implies that the embedding holds , as we wanted to see .
theorem [ teo : rango r(x1,loo)->r(x2,l1 ) ] can be extended to the case @xmath15 as follows : if we define @xmath394 , then with a similar proof one can get @xmath395 although this embedding is not optimal in general . for example , with @xmath396 then @xmath397 , but in this case @xmath398 and clearly @xmath399 . however , if @xmath400 , then gives @xmath401 which is indeed optimal ( as proved in @xcite ) .
our main goal , in this section , is to study the following embedding @xmath402 in particular , we are interested in the following problems : a. given a mixed norm space @xmath403 we would like to find the largest r.i .
range space @xmath29 satisfying .
b. now , let us suppose that the range space is given @xmath404 we would like to find the largest of the form @xmath28 for which holds .
let us first suppose that the embedding @xmath407 holds . as before , we assume that @xmath242 with @xmath409 and @xmath245 given any @xmath410 with @xmath411 we define @xmath412 now , we fix any @xmath132 then , @xmath413 and @xmath300 otherwise .
thus , using the boundedness of the dilation operator in r.i .
spaces , we get @xmath414 so , our assumption on @xmath48 shows that @xmath415 and @xmath416 thus , using @xmath417 and , we obtain @xmath418 but , @xmath419 and hence gives @xmath420 this proves , for any function @xmath410 with @xmath421 the general case can be proved as in the proof of theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] .
now , let us suppose that holds .
we fix any @xmath366 and @xmath422 then , for any @xmath19 it holds that @xmath423 so , we get @xmath424 hence , corollary [ corollary : distribution and essential projection ] , with @xmath425 replaced by @xmath426 implies that @xmath427 for any @xmath428 as a consequence , @xmath429 therefore , we obtain @xmath430 thus , we have @xmath431 hence , using , we get @xmath432 that is , the embedding @xmath433 holds and the proof is complete .
let @xmath29 be an r.i .
now , we want to find the largest space of the form @xmath28 satisfying @xmath434 in order to do this , let us introduce a new space , denoted by @xmath435 , consisting of those functions @xmath436 for which the quantity @xmath437 is finite .
it is not difficult to verify that @xmath435 is an r.i .
space equipped with the norm @xmath438 [ theorem : dominio r(x , loo)->z ] let @xmath29 be an r.i .
space , with @xmath441 and let @xmath442 be the r.i . space defined in then , the embedding @xmath443 holds .
moreover , @xmath444 is the largest space of the form @xmath28 for which the embedding holds .
the embedding follows from theorem [ theorem : mixta inclusion r(x , loo)->z ] .
thus , to complete the proof , it only remains to see that @xmath444 is the largest domain space of the form @xmath28 corresponding to @xmath29 .
in fact , we shall see that that if @xmath445 is another mixed norm space such that holds with @xmath444 replaced by @xmath445 , then @xmath446 we fix any @xmath447 then , theorem [ theorem : mixta inclusion r(x , loo)->z ] ensures us that @xmath448 and so , using lemma [ lemma : funcional equivalente dominio r(x , loo)->z ] , we get @xmath449 that is , @xmath450 hence , using theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] , we deduce that @xmath451 as we wanted to see .
lemma [ lemma : optimal range norm ] gives us that @xmath459 is an r.i .
space equipped with the norm @xmath461 now , let us see that the embedding @xmath460 holds .
in fact , let @xmath48 be any function from @xmath142 then , as a consequence of , we obtain now , let us see that @xmath459 is the smallest r.i .
space for which this embedding holds , i.e. , let us see that if an r.i .
space @xmath29 satisfies @xmath466
then @xmath467 as before , assume that @xmath242 with @xmath409 and @xmath245 given any function @xmath468 with @xmath469 we define @xmath470 then , applying the same technique as in the proof of theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] and using the boundedness of the dilation operator in r.i .
spaces , we get @xmath471 by hypothesis @xmath468 and hence @xmath472 so , using the embedding @xmath433 and we get @xmath473 but , @xmath50 and @xmath48 are equimeasurable functions , and hence we obtain @xmath474 from this , we obtain that any @xmath475 with @xmath469 belongs to @xmath404 the general case can be proved as in the proof of theorem [ teo : mixta inclusion r(x1,y1)->r(x2,y2 ) ] .
thus , the proof is complete .
we shall give now a corollary of theorems [ teo : rango r(x , loo)->z ] .
in particular , we shall see that the fournier s embedding can not be improved within the class of r.i .
spaces . this should be understood as follows : if we replace the range space in @xmath476 by a smaller r.i .
space , say @xmath477 , then the resulting embedding @xmath478 can no longer be true . | our main goal in this work is to further improve the mixed norm estimates due to fournier @xcite , and also algervik and kolyada @xcite , to more general rearrangement invariant ( r.i . )
spaces . in particular we find the optimal domains and the optimal ranges for these embeddings between mixed norm spaces and r.i . spaces . |
neptune mass extrasolar planets around main sequence stars were first detected five years ago . since then , 27 planets with a projected mass lower than 25 earth masses have been discovered , the lightest one having a projected mass of 1.94 m@xmath2 . among these objects ,
17 have a semi major axis smaller than 0.1 astronomical unit ( au ) .
they are called hot neptunes or hot super earths .
so far , two multiple planet systems with at least two such objects have been observed .
one of them is a four planet system around the m dwarf gj 581 ( bonfils et al .
2005 , udry et al .
2007 , mayor et al .
the projected masses of the planets are 1.9 , 15.6 , 5.4 and 7.1 m@xmath2 and the periods are 3.15 , 5.37 , 12.93 and 66.8 days , respectively .
note that although the eccentricities of the two innermost planets are compatible with zero , those of the next outermost and outermost planets are 0.16 and 0.18 respectively .
the second multiple system which we shall focus upon in this paper is that around the k dwarf hd 40307 ( mayor et al .
it comprises three planets with projected masses of 4.2 , 6.9 and 9.1 m@xmath2 and periods of 4.31 , 9.62 and 20.46 days , respectively .
the eccentricities are all compatible with zero .
if we label the planets in a system with successive integers starting from the innermost labelled , 1 , and moving outwards , then , in the system gl 581 , the ratio of the periods of planets 2 and 1 is 2.41 and that of the periods of planets 3 and 2 is 1.70 .
we note that these numbers depart from 5/2 and 5/3 only by 4% and 2% , respectively.a in the hd 40307 system , the ratio of the periods of planets 2 and 1 is 2.23 and that of planets 3 and 2 is 2.13 , which depart from 2 by 11.5% and 6.5% , respectively .
because departure from exact mean motion resonances in these systems is significant , the importance of resonances for their dynamical evolution has been ruled out ( mayor et al .
2009a , 2009b , barnes et al .
we shall address this aspect for the hd 40307 system in this paper .
migration due to tidal interaction with the disc is probably the mechanism by which planets end up on short period orbits , as _ in situ _ formation requires very massive discs ( e.g. , raymond et al .
planet scatterings may also lead to short period orbits , but only when tidal circularization is efficient . according to kennedy & kenyon ( 2008 ) ,
scatterings are not a likely way of producing hot super
earths , because of the long circularization timescales involved . in a previous paper ( terquem & papaloizou 2007 , see also brunini & cionco 2005 ) , we proposed a scenario for forming hot super
earths on near commensurable orbits , in which a population of cores that assembled at some distance from the central star migrated inwards due to tidal interaction with the disc while merging on their way in .
we found that evolution of an ensemble of cores in a disc almost always led to a system of a few planets , with masses that depended on the total initial mass of the system , on short period orbits with mean motions that frequently exhibited near commensurabilities and , for long enough tidal circularization times , apsidal lines that were locked together . starting with a population of 10 to 25 planets of 0.1 or 1 m@xmath3
, we ended up with typically between two and five planets with masses of a few tenths of an earth mass or a few earth masses inside the disc inner edge .
interaction with the central star led to the initiation of tidal circularization of the orbits which , together with possible close scatterings and final mergers , tended to disrupt mean motion resonances that were established during the migration phase .
the system , however , often remained in a configuration in which the orbital periods were close to commensurability .
apsidal locking of the orbits , if established during migration , was often maintained through the action of these processes .
this scenario has been questioned ( mayor et al .
2009a , 2009b ) because the multiple planet systems of hot super earths detected so far do not exhibit close enough mean motion commensurabilities . in this paper
, we argue that _ the system around hd 40307 does actually exhibit the effects of resonances _ through secular effects produced by the action of the resonant angles coupled with the action of tides from the central star , and that such a configuration could be produced if the system formed as described in terquem & papaloizou ( 2007 ) , but with the addition of relatively small perturbations to the orbits arising from close encounters or collisions after the system enters the disc inner cavity .
the reason that what are regarded as resonant effects can arise even when departures from strict commensurability are apparently large is that tidal circularization produces small eccentricities which , for first order resonances , can be consistent with resonant angle libration under those conditions ( see murray & dermott 1999 ) .
we consider in detail the evolution of this system and other putative similar systems with scaled masses and orbital periods after they form a strict three planet laplace resonance as a result of convergent inward orbital migration .
we go on to consider the onset of tidal circularization and how it leads to a separation of the semi - major axes and consequent increasing departure from commensurability .
we find that if the strict laplace resonance is broken by a fairly small relative perturbation corresponding to a few parts in a thousand , continuing evolution in a modified form of the three planet resonance could in principle in the case of hd 40307 lead to a system like the observed one .
however , this would require masses significantly larger than the minimum estimates and possibly unrealistically efficient tidal dissipation with the tidal dissipation parameter @xmath4 however , weaker tidal effects could be responsible for some of the deviation from strict commensurability , with the rest being produced by a larger perturbation resulting from the processes mentioned above inducing eccentricities of the order @xmath5 note however that tidal evolution could play a more significant role in similar systems with shorter orbital periods . in section [ sec :
model ] , we describe the numerical model we use to simulate the evolution of a system of three planets migrating in a disc and evolving under the tidal interaction with the star after they enter the disc inner cavity . in section [ sec : hd40307 ] , we simulate the system around hd 40307 using either the minimum or twice the minimum masses for the planets under varying assumptions about initial eccentricities and the strength of the tidal interaction .
we show that three of the four possible resonant angles associated with 2:1 commensurabilities either librate or have long term time averages , indicating evolution of the system towards a resonant state driven by the tidal influence of the star . in a related appendix
, we give a semi analytic model for a system in a three planet resonance undergoing circularization with departures from strict mean motion commensurabilities .
we note in passing that the discussion may also be applied to a two planet system near a 2:1 resonance .
we show that the departure from commensurability increases with time being @xmath6 and derive an expression for the timescale required to attain a given departure that can be used to interpret extrapolate from and generalize the numerical simulations .
in section [ sec : simulations ] , we present results of numerical simulations of a three planet system migrating in a disc , establishing a 4:2:1 laplace like resonance , and evolving under tidal effects induced by the star after entering the disc inner cavity . as expected ,
as the orbits are being tidally circularized , the planets move away from exact commensurabilites , and the period ratios of the outer and inner pairs of planets increases .
the system is still resonant in the sense that some of the resonant angles continue to librate . to get a larger period ratio for the inner pair of planets , as in hd 40307 ,
some disruption of this laplace resonant state is needed . in our simulations , this is achieved by applying an impulse ( that could result from an encounter or collision ) to the system .
this leads to departures from exact commensurabilites of the form seen in hd 40307 while still retaining libration of three of the four resonant angles .
finally , in section [ sec : discussion ] we discuss and summarize our results .
the model we use has been described in terquem & papaloizou ( 2007 ) .
we recall here the main features .
we consider a system consisting of a primary star and n planets embedded in a gaseous disc surrounding it ( for hd 40307 , n = 3 ) .
the planets undergo gravitational interaction with each other and the star and are acted on by tidal forces from the disc and the star .
the evolution of the system is studied by means of the solution of @xmath7body problems , in which the tidal interactions are included as dissipative forces .
the equations of motion are : @xmath8 where @xmath9 , @xmath10 and @xmath11 denote the mass of the central star , that of planet @xmath12 and the position vector of planet @xmath12 , respectively .
the acceleration of the coordinate system based on the central star ( indirect term ) is : @xmath13 and that due to tidal interaction with the disc and/or the star is dealt with through the addition of extra forces as in papaloizou & larwood ( 2000 ) : @xmath14 where @xmath15 , @xmath16 and @xmath17 are the timescales over which , respectively , the angular momentum , the eccentricity and the inclination with respect to the unit normal @xmath18 to the assumed fixed gas disc midplane change .
evolution of the angular momentum and inclination is due to tidal interaction with the disc , whereas evolution of the eccentricity occurs due to both tidal interaction with the disc and the star .
we have : @xmath19 where @xmath20 and @xmath21 are the contribution from the disc and tides raised by the star , respectively .
relativistic effects are included through @xmath22 ( see papaloizou & terquem 2001 ) .
the way it is implemented , interaction with the star does not modify the angular momentum of a single orbit ( the orbital decay timescale due to tidal interaction with the star , which as indicated below is estimated to be much longer than the circularization timescale , has been ignored here ) .
however , in the formulation above , eccentricity damping causes radial velocity damping , which results in energy loss at constant angular momentum . as a consequence
, both the semi major axis and eccentricity are reduced together ( at least for a single planet ) .
the circularization timescale due to tidal interaction with the star can be estimated , from expressions given by goldreich & soter ( 1966 ) , to be : @xmath23 where @xmath24 is the semi
major axis of planet @xmath25 here we have adopted a mass density of 1 g @xmath26 for the planets ( uncertainties in this quantity could be incorporated into a redefinition of @xmath27 ) .
the parameter @xmath28 where @xmath29 is the tidal dissipation function and @xmath30 is the love number .
equation ( [ teccs ] ) applies to the situation where tides raised by the central star are dissipated in the planet in the limit of small eccentricity which is appropriate here .
tides raised by the planet in the star have been estimated to be unimportant for the low mass planets of interest ( eg .
barnes et al 2009 ) . for solar system planets in the terrestrial mass range ,
goldreich & soter ( 1966 ) give estimates for @xmath29 in the range 10500 and @xmath31 , which correspond to @xmath27 in the range 502500 .
of course this parameter must be regarded as being very uncertain for extrasolar planets , and accordingly we have considered values of @xmath27 in the range @xmath32 in this paper . for an earth mass planet at 0.05 au and @xmath33 , we get @xmath34 years .
this indicates that circularization processes will be important for close in extrasolar planets in this mass range over the entire range of @xmath27 considered .
but note that although circularization operates on each orbit according to equation ( [ teccs ] ) , because the planets interact they may enter into a decay mode with a decay timescale that is not straightforwardly estimated as above ( see section [ initecc ] ) .
nonetheless the above estimate is found to be indicative .
we have pointed out above that , in our formulation , tidal interaction with the star does not change the angular momentum of a single orbit but modifies the semi major axis .
the physical basis for this is that the planets rapidly attain pseudo - synchronization ( e.g. , ivanov & papaloizou 2007 ) , after which they can not store significant angular momentum changes through modifying their intrinsic angular momenta .
we restrict consideration for the moment to a single planet .
since the orbital angular momentum is conserved , @xmath35 for small eccentricities , where @xmath36 is the eccentricity of the planet we consider . therefore @xmath24 changes only by about 5 percent in @xmath37 years for @xmath38 @xmath39 m@xmath2 and @xmath40 however , larger changes may occur for larger planet masses .
we investigate the dependence of this form of evolution on the system configuration in detail below .
finally we remark that all the simulation results reported here were obtained assuming that @xmath27 is the same for all the planets .
however , tests were carried out with this assumption relaxed and these are alluded to in section [ sec : simulations ] below . in the local linear treatment of type i migration ( e.g. , tanaka et al .
2002 ) , if the planet is not in contact with the disc , there is no interaction between them so that @xmath15 , @xmath20 and @xmath17 are taken to be infinite .
when the planet is in contact with the disc , disc planet interactions occur leading to orbital migration as well as eccentricity and inclination damping ( e.g. , ward 1997 ) . in that case , away from the disc edge , for some simulations we have adopted : @xmath41 \left [ 1- \left ( \frac { e_i } { 1.1 h / r}\right)^4 \right]^{-1 } \ ; \left ( \frac{h / r}{0.05 } \right)^2 \ ; \frac{{\rm m}_\odot}{m_d } \ ; \frac{{\rm m}_\oplus}{m_i } \ ; \frac{a_i}{{\rm 1~au } } \ ; \ ; \ ; { \rm years } , \label{tm}\ ] ] @xmath42 \ ; \left ( \frac{h / r}{0.05 } \right)^4 \ ; \frac{{\rm m}_\odot}{m_d } \ ;
\frac{{\rm m}_\oplus}{m_i } \ ; \frac{a_i}{{\rm 1~au } } \ ; \ ; \ ; { \rm years } , \label{te}\ ] ] and @xmath43 ( eq . [ 31 ] and [ 32 ] of papaloizou & larwood 2000 with @xmath44 ) .
here @xmath36 is the eccentricity of planet @xmath45 , @xmath46 is the disc aspect ratio and @xmath47 is the disc mass contained within 5 au .
we have assumed here that the disc surface mass density varies like @xmath48 .
for a 1 earth mass planet on a quasi
circular orbit at 1 au , we get @xmath49 yr and @xmath50 yr for @xmath51 m@xmath52 and @xmath53 .
the timescales given by equations ( [ tm ] ) and ( [ te ] ) can be used not only for small values of @xmath36 , but also for eccentricities larger than @xmath46 .
we have also carried out simulations with @xmath15 and @xmath20 taken to be constants and thus independent of eccentricity .
the commensurabilities that are formed in the system depend on the ratio of the migration rate to the orbital frequency , with commensurabilities of low order and low degree forming when this ratio is small .
therefore , using equations ( [ tm ] ) and ( [ te ] ) or constant timescales give very similar results provided the eccentricity damping is ultimately effective and the magnitudes of the migration rates are comparable in both cases at zero eccentricity .
the type i migration rates to be used are uncertain even when the surface density is known , largely because of uncertainties regarding the effectiveness of coorbital torques ( e.g. , paardekooper & melema 2006 , pardekooper & papaloizou 2008 , 2009 ) . in this context there are indications from modelling the observational data that the adopted type i migration rate should be significantly below that predicted by ( [ tm ] ) and ( [ te ] ) or tanaka et al ( 2002 ) ( see schlaufman et al 2009 ) . hence we explore a range of possible values .
three super earths have been detected orbiting the star hd 40307 through the radial velocity variations they produce ( mayor et al .
the planets have projected masses of 4.2 , 6.9 and 9.2 m@xmath54 and the reported semi
major axes are 0.047 , 0.081 and 0.134 au , respectively , which correspond to period ratios of 2.26 and 2.13 for the inner and the outer pairs of planets , respectively .
the periods reported by mayor et al .
( 2009a ) are 4.31 , 9.62 and 20.46 days , which correspond to period ratios of 2.23 and 2.13 .
the derived eccentricities are compatible with zero
. we will refer to the planets as 1 , 2 and 3 , planet 1 being the innermost object .
because the period ratios differ significantly from 2 , resonances have been ruled out ( mayor et al .
2009a , barnes et al . 2009 )
. however , the period ratios are not far from 4:2:1 , being characteristic of a three planet resonance .
this indicates that such a resonance might play a role at some evolutionary phase of the system . to investigate this hypothesis , we have simulated the system by means of @xmath7body calculations using the method described in the previous section .
the equations of motion are integrated using the bulirsch stoer method ( e.g. , press et al .
1993 ) .
the planets were set up on circular orbits around a 0.77 m@xmath55 star ( the estimated mass of hd 40307 ) with both the minimum masses indicated above and also masses a factor of two larger . the initial semi
major axes were taken to be those indicated above , except in two simulations described below , where the innermost semi
major axis was adjusted to make the period ratio of the innermost pair @xmath56 instead of 2.26 .
the results of a calculation with the tidal parameter @xmath57 and minimum masses are displayed in fig .
[ figjp1 ] , which shows the evolution of the eccentricities and runnimg time averages of the cosines of the resonant angles @xmath58 @xmath59 @xmath60 and @xmath61 over a timescale of 3.6 million years . here
@xmath62 and @xmath63 are the mean longitude and the argument of pericentre of planet @xmath45 , respectively .
the semi major axes do not vary significantly with time while the eccentricities remain small , below @xmath64 . however , tidal circularization processes are important for this run as the eccentricity of the outermost planet and the amplitude of the oscillations of the eccentricity of the inner two planets tends to decrease .
interestingly , the running time averages of the cosines of the resonant angles are not zero , as would be expected for a non resonant system .
they are such that the values for @xmath65 and @xmath66 approach unity , while the value for @xmath67 moves towards @xmath68 , corresponding to the angle approaching @xmath69 the angle @xmath70 exhibits non resonant behaviour .
( full line ) , @xmath71 ( dotted line ) , @xmath72(dashed line ) , @xmath73 ( dot dashed line ) .
for all plots time is measured in yr.,width=604 ] we have also investigated the dependence of the evolution on @xmath74 small variations of the initial semi - major axes and the magnitudes of the planet masses .
figure [ figjp2 ] shows the evolution for @xmath75 , minimum masses and the initial innermost semimajor axis adjusted to make the innermost period ratio 2.23 . in this case , the mean values of the eccentricities show evidence only of limited evolution . nonetheless ,
running time averages of the cosines of @xmath76 @xmath77 and @xmath78 quickly attain steady non zero values .
this is consistent with the evolution of these runs being driven by tidal effects and towards a resonant state in which the resonant angles are associated with secular evolution of the orbital elements .
this is able to occur without precise commensurability only because the eccentricities are very small ( see appendix a ) .
for comparison in fig .
[ figjp2a ] we plot the running means of @xmath79 @xmath80 and @xmath81 for the same run .
these also attain non zero values also indicating long term variations associated with the corresponding resonant angles .
note that the cosines of the resonant angles are not special in this respect ( see also appendix a ) .
but for @xmath75 , minimum masses and initial innermost period ratio 2.23.,width=604 ] the dependence on the planet mass scale is illustrated in fig .
[ figjp3 ] which shows the evolution for @xmath57 , twice the minimum masses and without adjustment of the innermost semi - major axis .
figure [ figjp4 ] gives the evolution for the same system but with the initial innermost semimajor axis adjusted to make the innermost period ratio 2.23 .
both these runs show similar evolution to the minimum mass case illustrated in fig .
[ figjp1 ] but scaled to larger eccentricities .
thus the results are not sensitive to the choice of minimum masses or small changes in the semi - major axes .
( solid line ) @xmath82 ( dashed line ) and @xmath81 ( dot - dashed line ) for q=1000 and minimum masses .
time is measured in yr.,width=604 ] but with q=10 and with masses a factor of two larger than the minimum.,width=604 ] but with initial inner period ratio 2.23.,width=604 ] the first second and third rows of panels correspond to the cases where the inner , middle and outer planets were started with an initial eccentricity ( see text ) respectively and minimum masses were adopted .
the fourth row is the same as the third except that the planet masses were twice the minimum values . from left to right for each row ,
the first panel shows the eccentricity of the innermost planet , the second the eccentricity of the middle planet and the third panel shows the eccentricity of the outermost planet .
the fourth panel shows running time averages of @xmath83 ( solid line ) , @xmath84 ( dotted line ) @xmath85 ( dashed line ) and @xmath86 ( dot - dashed line).,width=604 ] and the innermost planet was started with a non zero eccentricity ( see text ) . in the upper right panel
the masses were twice the minimum values , @xmath57 and the outermost planet was started with a non zero eccentricity .
the lower left panel was as for the upper left panel but @xmath87 the lower right panel was as for the upper right panel but @xmath87 time is measured in yr .
note that the shifts occurring in the period ratios are independent of @xmath88,width=604 ] and @xmath75 respectively , minimum masses were adopted and the innermost planet was started with a non zero eccentricity ( see text ) . for the fourth @xmath89 and the masses were twice the minimum values . in this case
the outermost planet was started with a non zero eccentricity . from left
to right for each row , the first panel shows the eccentricity of the innermost planet , the second the eccentricity of the middle planet and the third panel shows the eccentricity of the outermost planet .
the fourth panel shows running time averages of @xmath83 and @xmath85 ( ultimately uppermost).,width=604 ] the calculations described above started with the planets on circular orbits .
the motivation for this is that it is expected that tidal circularization will lead to a state closely approaching this provided that initial eccentricities are not so large as to disrupt the system . in order to explore this aspect in more detail ,
we have carried out a number of simulations which started with non circular orbits .
we have explored initial eccentricities @xmath90 and found that the system enters into an eccentricity decay mode that causes the system to evolve towards the state we found starting from circular orbits in which three of the resonant angles move towards libration with non zero values for the running time averages of their cosines . in order to illustrate this we describe here results obtained when
just one of the planets was started with a non zero eccentricity for various values of @xmath91 to initiate the system it was started as for the simulation illustrated in figure [ figjp1 ] but with one of the planets set at pericentre in an orbit with the same angular momentum but with an eccentricity @xmath92 this is equivalent to transfer to a less tightly bound orbit with the original specific angular momentum , but with the specific energy multiplied by @xmath93 in all cases the eccentricities of the planets become coupled and start to decay .
after a while the system attains a slow decay mode for which , in a time average sense , the eccentricities are in an approximately constant ratio such that @xmath94 and @xmath95 fig .
[ figjp6 ] illustrates the evolution of the eccentricities and running time averages of the cosines of the resonant angles for minimum mass simulations with @xmath96 cases for which the innermost , middle and outermost planet respectively had an initial eccentricity are shown . for comparison a case with twice the minimum masses and for which the outermost planet was started with a non zero eccentricity
is also shown .
the evolution of all of these systems was as described above .
however , the introduction of the initial eccentricity while conserving the angular momentum of the system causes the semi - major axes and hence the period ratios in the system to change .
this evolution involves energy dissipation at fixed angular momentum and thus the system must spread radially as for an accretion disc ( eg .
lynden - bell & pringle 1974 ) .
this causes the period ratios to increase ( see fig . [ figjp7 ] ) .
when this happens the system attains period ratios that differ from the original ones .
however , this is easily compensated for by starting with appropriately shifted initial conditions .
the important point is that we get the same type of evolution independently of such shifts .
[ figjp7 ] shows that the largest shift occurs when the outermost planet starts with an initial eccentricity .
the period ratio of the innermost pair shifts from @xmath56 to @xmath97 a consequence of this is that the forced eccentricity of the innermost planet is @xmath98 smaller in this case than in the case where the inner planet started with a non zero eccentricity ( see fig .
[ figjp6 ] ) .
we remark that the same evolution is obtained when the masses are increased by a factor of two except that the ultimate eccentricities are a factor between @xmath99 and @xmath100 larger .
the above trends are in qualitative agreement with a simplified discussion of the resonant interaction between the planets that we give in appendix a even though only two of the resonant angles @xmath76 and @xmath77 are found to be librating at this point with @xmath86 having a non zero time average of relatively small magnitude .
finally we have checked that changing @xmath27 gives the same form of solution but with an evolutionary time that scales with @xmath88 results for @xmath89 and @xmath75 that confirm this view when compared with results already described for @xmath57 are illustrated in fig .
[ figjp8 ] . however , practical considerations mean that runs with large @xmath27 can not be followed to libration of the resonant angles .
nonetheless when @xmath101 @xmath85 and @xmath83 begin to attain non zero running means at the end of simulations with minimum masses after @xmath102 this indicates that resonance could be effectively entered for @xmath103 within the lifetime of such systems .
these results taken together suggest that the planets may be undergoing circularization in a three planet resonance .
as the circularization process involves energy dissipation at fixed angular momentum , the system must spread increasing the period ratios of neighbouring planets .
although this means an increasing departure from strict commensurability , resonant angles remain librating or circulating in such a way as to enable the required redistribution of angular momentum to take place . a simple model illustrating how this works is given in appendix a. the amount of spreading and increase in the period ratios that can take place in a given time depends on @xmath88 we investigate this in the context of a migration scenario for producing the configuration below .
simulations incorporating migration indicate that the planets were likely to have been in a strict laplace resonance with all four angles librating . however , if this persists the period ratios would increase maintaining a larger departure from commensurability for the outer pair than the inner pair in disagreement with observations .
thus some perturbation event must have disrupted this resonance enabling the planets to continue circularizing with the correct period ratio relationships . as we have seen above ,
giving one of the planets a relatively small eccentricity can produce significant changes to the period ratios . to investigate this scenario ,
we now describe simulations of systems of three planets migrating inwards in a disc that set up a laplace resonance .
we go on to study the evolution following from its disruption as the result of a perturbation applied once the planets have decoupled from the disc .
in all the runs we have performed , the planets converge on each other forming a three planet 4:2:1 resonance , and migrate in locked in that configuration . whether the resonance can be maintained all the way down to the disc inner cavity depends on how fast the migration is .
there is a tendency for the 4:2:1 resonance to be stable for slow migration rates but to be disrupted , with the appearance of higher order resonances once the migration rate becomes sufficiently fast .
however , the evolution is difficult to predict precisely , as the transition from the 4:2:1 resonance to higher order resonances happens through an instability which appears to be extended in time and chaotic .
small changes to the model parameters in this regime may affect whether this instability occurs or not within a given time interval .
similar phenomena have been seen in two planet resonant migration where a sequence of resonances can be formed , maintained and subsequently disrupted in turn ( see papaloizou & szuszkiewicz 2009 ) . figure [ fig3 ] shows the evolution of the semi major axes of the planets and of the period ratios for 3 planets with minimum masses initially located at 0.2 , 0.34 and 0.6 au , respectively .
the disc aspect ratio is @xmath104 equations ( [ tm ] ) and ( [ te ] ) for the migration and circularization times are used with @xmath105 m@xmath106 this was equal to the disc mass within 5 au in the original formalism of papaloizou & larwood ( 2000 ) .
however , as there is much uncertainty in the migration rates to be used for low mass planets we investigate different possibilities using this parameter to scale the migration and circularization rates ( see section [ typeimigration ] ) .
used in this way it no longer corresponds to a disc mass as in the original formulation .
the initial migration times for this case are then @xmath107 and @xmath108 years for the innermost and outermost planets respectively .
the initial period ratios for the inner and outer pairs are @xmath109 and @xmath110 , respectively .
this calculation led to the setting up of a laplace resonance that is maintained during the subsequent evolution of the system , although for a finite period of a time around @xmath111 years a temporary weak and inconsequential instability was present ( see below ) .
( in au , _ plots a _ and _ b _ ) , eccentricity @xmath112 ( _ plot c _ ) , period ratios ( _ plot d _ ) , resonant angles @xmath113 ( in rad . , _ plot e _ ) and @xmath114 ( in rad . , _ plot f _ ) versus time ( in million years ) for 3 planets initially located at 0.2 , 0.34 and 0.6 au . on plots a and b , each solid line represents a planet , whereas the dotted line indicates the location of the disc inner edge .
note the different time interval for plot b. on plot c , the solid and dotted lines represent the period ratio of the inner and outer pair of planets , respectively . here
@xmath115 m@xmath55 and @xmath116 ( note that the evolution during the disc migration phase is not sensitive to the value of this parameter ) .
even though strict commensurability is lost as a result of tidal interaction with the star , the planets stay locked in resonances , as indicated by the behaviour of the resonant angles . , width=604 ] in these calculations , the tidal @xmath27 parameter was set to @xmath117 .
this is at least 3 orders or magnitude smaller than the values supposed to apply in earth like planets , but this value was chosen to give evolutionary timescales that could be reasonably handled numerically ( i.e. runs that could be done within weeks on a single processor ) .
tests carried out below with more realistic values of @xmath27 indicate that the evolution during the disc migration phase is not affected by this specification .
furthermore , when the planets are inside the cavity , the scaling given by equation ( [ hscal ] ) holds , implying that increasing @xmath27 corresponds to scaling to a longer evolutionary timescale . with @xmath116 ,
the eccentricity damping timescale given by equation ( [ teccs ] ) is @xmath118 years for a 1 m@xmath54 planet at 0.05 au , inducing a timescale on which the semi
major axis evolves of about @xmath119 years for an eccentricity of @xmath64 ( see section [ sec : startide ] ) .
figure [ fig3 ] shows the evolution of the semi major axes , the eccentricities , the period ratios and the resonant angles @xmath120 and @xmath121 . as pointed out above , the laplace resonance for which all four resonant angles librate is established while the planets migrate in the disc , after about @xmath122 years .
the resonant angles @xmath123 and @xmath77 stay close to zero . then , while some of the planets are still in the disc , a weak instability appears for a finite period of time ( after about @xmath111 years ) and the resonance is weakened with the angles driven to large amplitude libration with an occasional circulation .
but the instability eventually dies out and the resonance is recovered before the planets decouple from the disc . note
that this instability is not always seen in the runs we have performed .
as indicated above it seems to be associated with an approach to a transition to a resonance of higher degree .
this is indicated by the fact that when the migration rate was speeded up by a factor of @xmath124 the outer pair of planets underwent a transition to a @xmath125 resonance while the inner pair maintained a @xmath126 commensurability at the point at which the semi - major axes had attained the same configuration . during the migration phase ,
the eccentricities are pumped up to several hundredths . because of the low value of @xmath27 , tidal interaction with the star begins to affect the evolution of the system slightly before the outermost planet penetrates inside the cavity .
once the outermost planet reaches the disc inner edge , the disc is removed altogether from the calculation , so that if the planet is pushed out again it will not interact with the disc .
this procedure is introduced to take into account the disappearance of the disc on a timescale of a few million years . in this particular run
, the outermost planet enters the cavity after @xmath127 years . at that point , changes in semi major axes are due only to tidal interaction with the star .
as the eccentricities of the two inner planets are damped from a few @xmath64 to @xmath64 , their semi major axes decrease by about 1 percent ( see plot b of fig . [ fig3 ] ) in such a way that their period ratio increases by somewhat more than a percent , going from 2 to 2.037 ( see plot d of fig . [ fig3 ] ) .
in the meantime , the eccentricity of the outermost planet is hardly affected , as it was already very small ( a few @xmath128 ) when the planet entered the cavity . because of the conservation of the total angular momentum though , its semi major axis increases by about 1 percent , which results in the period ratio of the outer pair of planets increasing by a few percent , from 2 to 2.077 .
therefore , strict commensurability is lost .
however , we note that the resonant angles still librate , so that the laplace resonance is not disrupted .
the fact that the deviation of the outer period ratio from 2 is twice the deviation of the inner period ratio from 2 is a consequence of the laplace relation ( equation [ [ lapr ] ] of appendix a ) and this feature is preserved in the subsequent evolution of this and similar cases .
as we shall see , these deviations increase approximately @xmath6 , as predicted from equation ( [ hscal ] ) .
as noted above , the laplace relation is not satisfied in hd 40307 . as we shall see below ,
it is possible to obtain the correct relationship between the period ratios provided the laplace resonance is disrupted such that only @xmath129 and @xmath78 librate and contribute to the secular evolution ( see also section [ initecc ] ) . in order to investigate the dependence on the specification of the circularization rates induced by the disc
, we have performed simulations for which these depend only on the planet mass and are therefore constant .
we here consider examples for which : @xmath130 and : @xmath131 results for cases both with the minimum masses for hd 40307 and twice the mimimum masses are plotted in figure [ figjpmig ] . during the disc induced migration ,
a three planet resonance is formed in both cases .
as expected , the evolution for the higher mass case is twice as fast but otherwise similar .
the resonance shows only small signs of instability while the planets migrate inwards .
this is probably helped by the fact that the migration rate is here constant rather than increasing inwards as in the previous simulation .
as the migration continues , the resonant angles @xmath132 ( @xmath133 ) all enter libration , but after the planets enter the cavity there is an adjustment so that only @xmath123 and @xmath77 librate .
the system then evolves back towards a laplace resonance where all angles librate or contribute to long term time averages .
this latter aspect is similar to what was seen in the simulations with larger @xmath27 illustrated in figs .
[ figjp1][figjp4 ] .
( although evolution during the disc migration phase is not sensitive to this parameter),width=604 ] . the _
upper panels _ give the evolution of the semi major axes and the _ lower panels _ the period ratios .
[ figjpmig ] this final evolution is one where the eccentricities all reduce and the planets physically separate with increasing period ratios for the inner and outer pairs of planets .
this form of evolution is to be expected in a keplerian system in which energy is dissipated while the total angular momentum is conserved , as is the case for an accretion disc ( e.g. , lynden bell and pringle 1974 ) .
it is this form of evolution we investigate below and how far the tidal effects of circularization can bring the system towards the hd 40307 system .
as the continuation of the simulations described above maintains a three planet laplace resonance , the period ratios will not have the same form as in hd 40307 .
as indicated above the system has a larger period ratio for the inner pair as compared to the outer pair whereas the laplace relation leads to the converse .
however , we have found that there are resonant interactions for which only @xmath134 and @xmath78 librate .
hereafter we refer to this as the standard state of libration .
this can lead to the required form for the period ratios but some disruption of the form of the resonance initially set up within the cavity needs to take place .
mechanisms could involve close encounters or collisions occurring shortly after the disc migration phase .
we comment that we have found it impossible to disrupt the 2:1 resonance of the inner pair of planets with a slow process .
for instance , we have added a fourth planet on an inclined orbit , but found that even a massive perturber ( 3 jupiter masses ) at 5 au could not take the inner pair away from the 2:1 resonance . to do this
, the parameters of one or more of the orbits have to be varied suddenly preventing the resonance from responding adiabatically .
an encounter with a fourth planet on a parabolic orbit is a possible means to achieve this and there are other possibilities involving one or more direct collisions . to affect the resonance , the semi
major axis of one of the planets should change by more than an effective libration width associated with one of the resonant angles .
we have performed simulations for which the velocity components of one of the planets were all suddenly increased or decreased by a constant factor .
this has the effect of preserving circular orbits which is reasonable under conditions of effective circularization .
we found that fractional changes exceeding @xmath135 were sufficient and , provided the changes were not too excessive , simulations approach the same form of evolution .
below we describe a simulation where an impulsive change that decreased the velocity components of the innermost planet by @xmath136 was applied just after the outermost planet entered the cavity . the end of the simulation illustrated in fig .
[ figjpmig ] ( or its larger planet mass counterpart where appropriate ) was used as a starting point for the simulations presented below , as these are reasonable outcomes for a system having undergone inward migration induced by a disc and then entered an inner cavity . in some cases ,
the tidal parameter @xmath27 was also adjusted .
we remark that we have performed simulations with other types of applied impulse .
for example we have augmented the above form by giving the innermost planet an initial eccentricity in a similar manner to that undertaken in section [ initecc ] .
when this is small and typically of order @xmath64 this makes no difference on account of rapid effective circularization . we have also considered impulsive changes to the outermost planet giving it an eccentricity @xmath137 as in section [ initecc ] . in this case period ratios can be shifted to be close to the observed ones just after the outermost planet has entered the cavity ( see below ) . in fig .
[ figjpmig1 ] , we illustrate the evolution of a model system of three planets with masses twice the minimum masses of hd 40307 with @xmath96 two runs are displayed .
the first one continues the evolution of the corresponding case plotted in fig .
[ figjpmig ] . in the second run ,
an impulse was first applied to the innermost planet that reduced all velocities by a factor of @xmath138 .
right panels _ correspond to the evolution after an impulse was applied to the inner planet ( see text ) . for both models @xmath139 the period ratios @xmath140 and @xmath141
are plotted as a function of time in the _
upper panels_. the upper curve corresponds to the outer pair ( _ left panel _ ) and the inner pair ( _ right panel _ ) . in this and similar figures below
, the triangles correspond to analytic fits of the form given by equation [ anfit ] ) ( and see text ) .
the _ lower panels _ show the evolution of the eccentricities of the three planets . in the _ lower left panel _ the curves from top down
correspond to the inner , middle and outer planets respectively . in the _ lower right panel _ the curves from top down
correspond to the middle , lower and outer planet.,width=604 ] as shown in fig .
[ figjpmig1 ] , during the subsequent evolution the system spreads radially with increasing period ratios .
the way this happens is illustrated by the simple model described in appendix a. in this discussion the period ratios are replaced by the quantities @xmath142 and @xmath143 that are easily obtained from fig . [ figjpmig1 ] . we have been able to find good fits to the time evolution of the form found in appendix a as given by equation ( [ hscal ] ) @xmath144 where @xmath145 and @xmath146 are constants and @xmath147 is the time in units of @xmath148 yr measured from the start of the relevant simulation shown in fig .
[ figjpmig ] . in the case with no initial impulse , @xmath149 and @xmath150 for the outer pair , and @xmath151 and @xmath1521.5 for the inner pair . for the case with an initial impulse , @xmath153 and @xmath154 for the outer pair , and @xmath155 and @xmath154 for the inner pair . for the latter case ,
a standard libration state occurred with the deviation of the inner period ratio from two being about a factor of two larger than the corresponding deviation for the outer period ratio , similar to the situation in hd 40307 .
this is in contrast to the case with no impulse that persisted in a laplace resonance with all angles librating . in that case
, the laplace relation persists which leads to the relative deviation of the period ratios from two behaving differently from hd 40307 ( see above ) . despite this , we can estimate the time it would take to obtain the outer period ratio of 2.13 appropriate to hd 40307 .
this is @xmath156 yr . the scaling of the circularization time , @xmath157 leading to the scaling of the evolution time , @xmath158 to reach a given state , with mass of the form @xmath159 implies that for a minimum mass system this would be @xmath160 yr .
thus , it might just be possible to achieve the outer period ratio in this case but then the inner period ratio would have to be determined subsequently by alternative means . for the case with the applied impulse , we may estimate the time to bring the inner period ratio to the appropriate value @xmath161 this is somewhat uncertain because of the small variation seen in the simulation , but it is about @xmath162 yr . note that the relative period ratios are not precisely identical to those of hd 40307 . in particular the ratio of the deviation of the inner period ratio from two and , the deviation of the outer period ratio from two is about @xmath163 too large
. however , this may be adjusted by changing simulation input parameters .
for example we recall that the simulations reported here were performed with the same value of @xmath27 for all of the planets . to indicate one possibility we note that tests we have performed indicate that the above ratio may be reduced by taking a smaller value of @xmath27 for the outermost planet than the others .
we remark that , while it might be barely possible to obtain an inner period ratio of 2.23 for a larger mass system , it is almost certainly not possible for the minimum mass system as the time required would be @xmath164 yr . note that , as the evolution progresses , the eccentricities of all the planets tend to decrease , as does their variation .
this is a general consequence of the spreading that is driven by the energy dissipation in these simulations . to investigate further the form of the evolution and its scaling with run parameters , we plot in fig .
[ figjpmig2 ] the results of two simulations of models with minimum masses and for which an impulse of the above form was applied to the innermost planet .
we compare cases with @xmath165 and @xmath166 as stated above , we had to consider rather small values in order to perform simulations that produced a reasonable variation in a reasonable time .
these resulted in a standard state of libration and relative period ratios very similar to the higher mass simulation with an initial impulse illustrated in fig .
[ figjpmig1 ] . ,
but for models with minimum masses and for which an impulse was applied to the innermost planet ( see text ) . for
the _ left panels _ , @xmath165 and for the _ right panels _
, @xmath166 in the _ upper panels _ , the upper curves correspond to the inner period ratio . in the _ lower panels _ , the curves from top down correspond to the middle , inner and outer planets respectively.,width=604 ] in the case with @xmath165 , the fit parameters are @xmath167 and @xmath168 for the inner pair , and @xmath169 and @xmath1521.7 for the outer pair . for the case with @xmath170 , they are @xmath171 and @xmath172 for the inner pair , and @xmath173 and @xmath1521.5 for the outer pair .
the values of @xmath145 , being about an order of magnitude larger in the second case , are consistent with the evolution time being an order of magnitude larger , thus scaling with @xmath27 as expected .
the time to attain an inner period ratio of @xmath56 in the @xmath165 case is @xmath174 yr , leading to an estimate of @xmath175 yr when @xmath96 this is less than the estimate based on the analytic scaling with planet mass implied by equation ( [ hscal ] ) applied to the higher mass simulations illustrated in fig .
[ figjpmig1 ] by a factor @xmath176 which we believe is a reasonable level of consistency in view of the extent of the extrapolation involved .
the eccentricities in these simulations decrease with time in a manner analogous to the simulations shown in fig .
[ figjpmig1 ] .
the long times required for the above cases to reach a configuration similar to that now observed in hd 40307 arise because only a very small perturbation was applied just after the systems decoupled from the disc , requiring a large amount of subsequent period ratio evolution .
such a scenario is only feasible for masses exceeding twice the minimum and small @xmath4 however , if a larger perturbation is allowed , it can provide a larger shift in the period ratios requiring less evolution later on ( see fig .
[ figjp7 ] ) .
this is equivalent to starting on the evolutionary curves shown in figs .
[ figjpmig1]-[figjpmig2 ] at a later time .
for example we found that if a perturbation of the type adopted in section [ initecc ] was applied to the outer planet a shift of the right type could be obtained .
we found that when the imposed eccentricity of the outer planet was taken to be @xmath177 the period ratios were shifted to very similar values to those in hd 40307 .
thus if large enough perturbations are allowed , only a small amount of subsequent evolution may be needed .
in this paper , we have shown that the planets around the star hd 40307 could be undergoing a resonant interaction , despite departure of the period ratios from very precise commensurability .
this is indicated by the fact that three of the four resonant angles librate or are associated with long term changes to the orbital elements .
note that such a resonant state can occur without exact commensurability only because the eccentricities are very small , which can be brought about as a result of tidal circularization from a state with initially higher eccentricities ( see section [ initecc ] ) .
when this occurs in a system that starts from a near 2:1 commensurability , the difference of the period ratios from two then increases @xmath178 we propose that the planets in this system were in a strict laplace resonance while they migrated through the disc , with all 4 resonant angles librating .
exact commensurability was then departed from as indicated above as a result of the tidal interaction with the star , which preserved libration of at least some of the resonant angles , after the planets entered the disc inner cavity . because of the laplace relation , the period ratios evolve in such a way that the ratio for the inner pair of planets is smaller than that of the outer pair of planets .
the opposite is observed in hd 40307 . to get the appropriate form for the period ratios , some disruption of the resonance set up in the cavity needs to take place ( this is in contrast to claims by zhou 2009 ) .
a close encounter for instance might be a possible cause .
an impulsive interaction would be sudden enough that the resonance would not respond adiabatically . if the velocity of the innermost planet is reduced by a factor 0.996 for instance , deviation of the inner period ratio from two
gets twice as large as the corresponding deviation for the outer period ratio , while 3 of the 4 resonant angles still librate , similar to the situation in hd 40307 , within the lifetime of the system
. however , we found that we could get a period ratio of 2.23 for the inner pair only if the masses are significantly larger than the minimum masses and tidal dissipation is possibly unrealistically effective . from the results obtained in section [ sec : simulations ] ,
when the laplace resonance is disrupted at an early stage , we estimate that the deviation of the inner period ratio from two is approximately given as a function of time by @xmath179 where @xmath180 @xmath181 and @xmath182 ( @xmath183 and @xmath184 are assumed fixed for the purposes of this discussion ) .
accordingly , to attain @xmath185 we would require masses exceeding the minimum estimate by about a factor of two for @xmath186 if the laplace resonance is preserved , from the results obtained in section [ sec : simulations ] , we may also estimate that the deviation of the outer period ratio from two is approximately given as a function of time by @xmath187 .
thus it may be possible to get a period ratio of 2.13 for the outer pair of planets with near to minimum masses in a reasonable time if @xmath188 for all the planets , but then the inner period ratio would be only @xmath189 it would thus seem likely that additional stronger dynamical interactions are required to bring the system to its final state . for example we found that if the outer planet was given an eccentricity @xmath190 shortly after decoupling from the disc , period ratios similar to those observed in hd 40307 could be produced on a time scale of a few @xmath191 ( see sections [ initecc ] and [ sec : simulations ] ) . then , the work presented in this paper indicates that tidal circularization would continue through resonant interaction with the period ratios separating after these interactions are over .
note that a similar type of scenario may apply to the system of planets around gj 581 , which is currently closer to exact commensurablilty than hd 40307 .
indeed , different mean motion resonances can be established while the planets migrate within the disc , depending on the initial relative location of the planets and on the migration timescale . however , as it appears that this system may have been in higher order resonances than hd 40307 , it is possible that these would have been broken completely as a result of circularization .
we intend to present a study of gj 581 in a forthcoming paper . in a study of the tidal evolution of the system around hd 40307 , barnes et al .
( 2009 ) have concluded that the planets can not be earth like . according to them ,
if it were the case , extrapolating back in time the tidal evolution of the eccentricities from current values ( estimated from direct simulations of the observed system ) , the system would have been unstable .
the work in this paper does not support such a conclusion .
we note that barnes et al .
( 2009 ) neglected the dynamical interactions between the planets , which clearly will have played a role if the current eccentricities are small because resonant interactions coupled with tidal effects have been associated with long term changes to the orbital elements .
the results reported in this paper indicate that resonances in multiple low mass planet systems may play more of a role than currently claimed in the literature .
significant departure from exact mean motion resonance does not necessarily preclude a resonant state when , for example , tidal effects are important causing eccentricities to be small , as libration of some of the resonant angles in that case may still exist , and be associated with long term evolution of the orbital elements .
although the confirmation of resonance effects is problematic for small eccentricities , we expect that further detections and analysis of systems similar to hd 40307 will test the relevance of the scenarios explored in this paper . finally , we comment that , because of the rapid increase of the effectiveness of tidal interactions with decreasing orbital period , the dynamical interactions discussed here would be of much greater importance for systems in which the period of the innermost planet is significantly shorter than in the hd 40307 system .
for example , we note that the planet corot7 b , which has a projected mass of 11.12 m@xmath2 , has a period of only 0.85 day ( lger et al .
if this short period orbit is a result of migration , it is likely this planet was pushed inwards by a companion which would have been on a commensurable orbit ( terquem & papaloizou 2007 ) .
we note in support of this idea that pre - main sequence stars typically have rotation periods of several days ( bouvier et al .
1993 ) , which would imply that the disc inner cavity should extend well beyond the orbit of corot7 b if that were magnetically maintained . in that case
, this planet could not have migrated so far inwards without being shepherded by a companion .
in this appendix we develop a semi analytic model that shows how a two or three planet system undergoes radial spreading as a result of evolution driven by tidal circularization .
the coupling between the planets occurs because resonant angles may librate or be associated with long term changes _ even though there may be significant deviations from exact commensurability_. we begin with the system without tides which is hamiltonian .
subsequently , we add the effects of tides . we adopt jacobi coordinates ( sinclair 1975 , papaloizou & szuszkiewicz 2005 ) for which the radius vector of planet @xmath192 @xmath193 is measured relative to the centre of mass of the system comprised of @xmath194 and all other planets interior to @xmath192 for @xmath195 with @xmath196 here and @xmath197 corresponding to the innermost planet .
the hamiltonian , correct to second order in the planetary masses , can be written in the form : @xmath198 here @xmath199 and @xmath200 here the orbital angular momentum of planet @xmath45 which has reduced mass @xmath202 with @xmath203 being the actual mass , is @xmath204 and the orbital energy is @xmath205 for motion around a central point mass @xmath194 we have : @xmath206 where @xmath207 @xmath24 denotes the semi - major axis and @xmath36 the eccentricity of planet @xmath25 the mean longitude of planet @xmath45 is @xmath208 where @xmath209 is its mean motion , with @xmath210 denoting its time of periastron passage and @xmath211 the longitude of periastron .
we remark that results valid for a two planet system , @xmath212 , may be obtained by setting @xmath213 here we are interested in the effects of two first order @xmath126 commensurabilities associated with the outer and inner pairs of planets . in this situation
, we expect that any of the four angles @xmath217 @xmath59 @xmath218 and @xmath61 may be slowly varying .
following standard practice ( see , e.g. , papaloizou 2003 , papaloizou & szuszkiewicz 2005 ) , only terms in the fourier expansion involving linear combinations of @xmath219 @xmath220 @xmath77 and @xmath78 as argument are retained
. working in the limit of small eccentricities applicable here , only terms that are first order in the eccentricities need to be retained .
expanding to first order in the eccentricities , the hamiltonian may be written in the form : @xmath221 where : @xmath222 , \label{hamil}\ ] ] with : @xmath223 here @xmath224 denotes the usual laplace coefficient ( e.g. , brouwer & clemence 1961 ) with the argument @xmath225 using equations ( [ eqnmo1])([eqnmo4 ] ) together with equation ( [ hamil0 ] ) we may first obtain the equations of motion without the effect of circularization tides and then add this in .
we obtain correct to lowest order in the eccentricities : @xmath226 the terms on the right hand sides involving the circularization times @xmath21 describe the effects arising from tides raised by the central mass on planet @xmath25 the terms @xmath227 in equations ( [ eqntid0 ] ) , ( [ eqntid1 ] ) and ( [ eqntid2 ] ) account for the orbital energy dissipation occurring as a result of circularization at the lowest order in @xmath36 for which it appears . in the absence of tides , the total energy @xmath228 and angular momentum @xmath229 are conserved . when circularizing tides act , the total angular momentum is conserved but energy is lost according to : @xmath230 when both pairs of planets @xmath231 and @xmath232 are in resonance , a three planet resonance is said to exist .
it is convenient to perform a time average of equations ( [ eqne1])-([eqntid5 ] ) taken over a time long compared to both the characteristic orbital period and the libration period , but short enough compared to the tidal time scale ( see eg .
sinclair 1975 for a discussion of this aspect ) .
the angles @xmath132 may be undergoing libration or circulation . equations ( [ eqne1])([eqntid1 ] ) may be averaged directly , while equations ( [ eqntid2])([eqntid5 ] ) may be averaged directly for small amplitude librations or alternatively after multiplying them by quantities such as either , e.g. , @xmath233
@xmath234 @xmath235 and @xmath236 or @xmath79 @xmath237 @xmath82 and @xmath81 , respectively .
the time averages of these quantities as well as @xmath238 and @xmath239 @xmath240 are assumed to vary on the tidal time scale or longer . when these are non zero , long term changes to the orbital elements @xmath24 and @xmath36 may occur through planet - planet interactions even without tides . as illustrated in figures [ figjp2 ] and [ figjp2a ] , six of these quantities are found to exhibit steady long term time averages of the required type that are quickly established for a model of hd 40307 with assumed @xmath241 as this averaging procedure results in equations of essentially the same form as the case where the participating resonant angles ( defined to be those that are associated with non zero time averages ) are assumed to have averages close to but not equal to zero or @xmath242 while undergoing small amplitude librations , leading to the same scaling with physical parameters , but with an adjustment of numerical coefficients . for simplicity we restrict consideration to that case . assuming small amplitude librations of all four angles , equations ( [ eqntid2])([eqntid5 ] ) give relations of the form : @xmath243 where : @xmath244 and the @xmath132 in the above can be set to either zero or @xmath69 we note that , when all four angles participate , equations ( [ lap2 ] ) and ( [ lap3 ] ) imply the strict laplace resonance relation @xmath245 however , simulations of hd 40307 indicate that only @xmath246 and @xmath78 participate .
thus equation ( [ lap2 ] ) is absent and the strict laplace relation does not hold . indeed , this relation implies that @xmath247 should be unity .
being about @xmath248 , the laplace relation is not satisfied .
when @xmath70 does not participate and the laplace relation does not hold , the quantity @xmath249 is not constrained from the outset but is determined as a consequence of the initial conditions and following time dependent evolution .
below , we consider the case when either all four angles contribute or only @xmath250 and @xmath78 contribute .
we note that a similar discussion applies when only @xmath251 and @xmath77 participate that leads to the same conclusions ( @xmath252 decays independently due to tides in that case and may be taken to be zero ) . in all cases
, one can readily verify that the above system provides a complete set of equations for the time averages of the resonant angles and the orbital elements when the form of the tidal forces is specified .
however , we note that we can obtain the characteristic form and time scale of the evolution by considering the conservation energy given by ( [ econ ] ) together with the constancy of the total angular momentum .
to do this we begin , by noting that equations ( [ lap1 ] ) and ( [ lap4 ] ) provide equations for the mean motion ratios of the form : @xmath253 and : @xmath254 and @xmath255 and @xmath256 are defined implicitly through equations ( [ lap1 ] ) and ( [ lap4 ] ) . in addition , equation ( [ lap3 ] ) gives @xmath257 in terms of @xmath252 , @xmath258 and @xmath259 in the form : @xmath260 when @xmath261 has zero time average and so is replaced by zero , equation ( [ lap2 ] ) does not hold and accordingly the laplace relation also does not hold .
we comment that @xmath258 and @xmath259 measure the deviation from strict commensurability of the corresponding mean motions and _ do not have to be small .
_ for a system such as hd 40307 these are large , particularly in comparison comparison to the characteristic size of the eccentricities .
we remark that according to equations ( [ lap1 ] ) - ( [ lap4 ] ) , other things remaining fixed , the eccentricities are proportional to the planet mass scale and , inversely proportional to the deviation from commensurability .
the results presented in section [ initecc ] confirm these trends but with some deviation that is to be expected because all the relevant resonant angles , although contributing to the evolution , are not in a state of small amplitude libration in those simulations .
replacing @xmath262 ( @xmath263 ) by @xmath194 to simplify expressions without loss of content ( as @xmath264 ) , the total energy and angular momentum may be written , respectively , as : @xmath265,\ ] ] and : @xmath266.\ ] ] here , we have neglected in @xmath267 terms proportional to the squares of the eccentricities as these can be shown to lead to small effects compared to those arising from the eccentricity dependence of @xmath255 and @xmath268 we now use the above together with equation ( [ econ ] ) to form the quantity @xmath269 and find that this is given by : we may write equation ( [ elcon ] ) in terms of the deviations from commensurabiity : @xmath271 and @xmath272 , obtaining : @xmath273 in addition , the right hand side can be expressed in terms of @xmath274 and @xmath275 using equations ( [ lap1])([lap4 ] ) which give : @xmath276 \frac{gm_3m_2 ^ 2n_3 ^ 2c_{2,3}^2\cos^2\phi_4 } { w_{2,3}^2ma_3n_3 ^ 2 t_{e,3}^s}. \label{elwcon}\ ] ] we may now obtain characteristic form of the time evolution of the departure from strict commensurability . to do this
, we may simply assume that @xmath277 @xmath275 are comparable , significantly less than unity and @xmath278 say , as has been found numerically in general and must be at least approximately the case for a strict laplace resonance for which @xmath279 ( see equation [ [ lapr ] ] ) .
in addition , we take the planet masses to be comparable and @xmath280 then the characteristic evolution expected from equations ( [ elcon1 ] ) and ( [ elwcon ] ) can be obtained from an equation of the form @xmath281 where we take @xmath282 to be the shortest circularization time @xmath283 thus : @xmath284 where @xmath285 specifies an arbitrary origin of time .
we recall that ( [ hscal ] ) applies to a two planet system by appropriately setting @xmath286 as well as a three planet system .
the time required to attain a departure @xmath287 is @xmath288 . in general
, this is very much longer than the characteristic circularization time . | in this paper , we consider the dynamics of a system of hot super earths or neptunes such as hd 40307 .
we show that , as tidal interaction with the central star leads to small eccentricities , the planets in this system could be undergoing resonant coupling even though the period ratios depart significantly from very precise commensurability . in a three planet system , this is indicated by the fact that resonant angles librate or are associated with long term changes to the orbital elements . in hd 40307
, we expect that three resonant angles could be involved in this way .
we propose that the planets in this system were in a strict laplace resonance while they migrated through the disc . after entering the disc inner cavity ,
tidal interaction would cause the period ratios to increase from two but with the inner pair deviating less than the outer pair , counter to what occurs in hd 40307 .
however , the relationship between these pairs that occurs in hd 40307 might be produced if the resonance is impulsively modified by an event like a close encounter shortly after the planetary system decouples from the disc .
we find this to be in principle possible for a small relative perturbation on the order of a few @xmath0 but then we find that evolution to the present system in a reasonable time is possible only if the masses are significantly larger than the minimum masses and tidal dissipation is very effective . on the other hand we found that a system like hd 40307 with minimum masses and more realistic tidal dissipation could be produced if the eccentricity of the outermost planet was impulsively increased to @xmath1 we remark that the form of resonantly coupled tidal evolution we consider here is quite general and could be of greater significance for systems with inner planets on significantly shorter orbital periods characteristic of for example corot 7 b. planetary systems : formation planetary systems : protoplanetary discs |
let @xmath5 be the farey sequence of order @xmath6 , which is defined to be the set of all subunitary irreducible fractions with denominators @xmath7 , arranged in ascending order .
for any interval @xmath8 $ ] , we write @xmath9 .
the cardinality of @xmath10 is well known to be @xmath11 when @xmath12 $ ] we write shortly @xmath13 instead of @xmath14}(q)$ ] .
since @xmath15 contains a large number of fractions obtained by a combined process of division , sieving and sorting of integers from @xmath16 $ ] , one would apriori expect little or even no special structure in the set of all differences between consecutive fractions ( which we also call _ intervals of a second _ ) .
though , this expectation is not fulfilled .
this is sustained from many points of view by a series of authors , such as franel @xcite , kanemitsu , sita rama chandra rao and siva rama sarma @xcite , hall and tenenbaum @xcite , @xcite , hall @xcite , augustin , boca and the authors @xcite , who have studied the set of gaps between consecutive farey fractions .
a regularity is expected also in the set of larger gaps @xmath17 , where @xmath18 runs over @xmath19 and @xmath2 .
( we use up - scripts , such as @xmath20 to write consecutive elements of @xmath15 . )
it is our object to treat here the case @xmath4 , that is , the case of _ intervals of a third_. geometrical representations of the set of pairs of neighbor intervals of fractions from @xmath15 created for different values of @xmath6 reveals sets of points whose density concentrates on different parts of the plane .
the aesthetical qualities of the pictures catches attention immediately . for for any @xmath21 they look like a swallow and the main topological distinctions are in the number of folds of the tail .
thus , when @xmath22 ( neighbor pairs of intervals of a second ) the swallow has a one - fold tail ( see @xcite ) . when @xmath4 , the case treated in the present paper , the swallow has a two - fold tail ( see figure [ figure1 ] ) and in section [ secsupport ] we have calculated explicitely the equations of the frontier . in the cases @xmath23 the tail appears always to have a three - folded tail , but this is more complex and its characterization will appear in a separate paper .
given @xmath24 real numbers @xmath25 with mean spacing @xmath26 , we consider _ the @xmath27-th level of intervals of a third probability _
@xmath28 on @xmath29 , defined , for @xmath30 , by @xmath31 in our case , we normalize @xmath10 to get the sequence @xmath32 , @xmath33 with mean spacing equal to one .
accordingly , we get the sequence @xmath34 of the @xmath27-th level of intervals of a third probabilities on @xmath35 .
we show that this sequence converges , as @xmath1 , to a probability measure @xmath28 , which is independent of @xmath36 , and can be expressed explicitly . for any @xmath37 and @xmath38 in @xmath15 , we set @xmath39 this is the numerator of the difference @xmath40 .
it is well known that @xmath41 for any consecutive elements of @xmath15 , and it turns out that this equality is responsible for the existence of the @xmath27-spacing distribution of the farey sequence .
though , this relation is no longer true for larger intervals , but there is a convenient replacement . to see this , let us note that a farey fraction can be uniquely determined by its two predecessors . indeed ,
if @xmath42 are consecutive fractions of @xmath15 , we have @xmath43 and @xmath44 , where @xmath45 $ ] .
the basic idea of our procedure is to parametrize the set of @xmath27-tuples of intervals of a third in terms of just two variables that run over a completely described domain .
the set of pairs of consecutive denominators of fractions in @xmath15 are exactly the elements of @xmath46 since we are mainly interested in what happens when @xmath47 , we reduce the scale @xmath6 times , and consider the background triangle @xmath48 , called _ the farey triangle_. we split it into a series of polygons as follows .
firstly , for each @xmath49 , we set @xmath50 , @xmath51 , and then , for @xmath52 , we define recursively : @xmath53 l_{i-1}(x , y)-l_{i-2}(x , y)\,.\ ] ] then , as in @xcite , we consider the map @xmath54 where @xmath55 $ ] .
the functions @xmath56 are locally constant , and the subsets of @xmath57 on which they are constant plays a special role .
thus , for any @xmath58 , we get the convex polygon @xmath59 notice that @xmath60 and @xmath61 whenever @xmath62 .
next we consider the application @xmath63 defined by @xmath64 our main result shows that , indeed , for @xmath1 , the sequence @xmath34 converges to a measure and @xmath65 is the needed tool to describe its support .
[ theorem1 ] the sequence @xmath34 converges weakly to a probability measure @xmath28 , which is independent of @xmath36 .
the support @xmath66 of @xmath28 is the closure of the range of @xmath67 , and @xmath68 for any parallelepiped @xmath69 . in table [ table1 ] from section [ secsupport ]
we provide explicit formulae for all the pieces that form @xmath70 .
it is plain that in order to prove theorem [ theorem1 ] , it suffices to see the effect of @xmath28 on bounded parallelepipeds . for any @xmath69
, we define @xmath71 we have to show that the sequence @xmath72 is convergent when @xmath1 and the limit is independent of @xmath36 . in the beginning
we treat the case of the complete interval @xmath12 $ ] . in the following we write shortly @xmath73 instead of @xmath74}$ ] . with the notations from the introduction
, we see that @xmath75 . then @xmath76 can be written as @xmath77 knowing that @xmath78 , we consider the set @xmath79 since neighbor denominators in @xmath80 are always coprime , relation turns into @xmath81 next , we select the points with coprime coordinates using mbius summation ( cf .
* lemma 2 ) ) , and we find that @xmath82 splitting @xmath57 into the series of polygons @xmath83 , we see that the error term in is @xmath84 . in the main term , we replace @xmath85 by the bounded set @xmath86 .
these yield @xmath87 it remains to replace in the set @xmath88 by a set as in , but with bounds independent of @xmath6 in the corresponding inequalities .
this set is @xmath89 notice that @xmath90 is exactly @xmath91 .
the replacement does not change the error term because , via @xmath92 , we have : @xmath93 which implies @xmath94 therefore , by and , we get @xmath95 in particular , this gives @xmath96 , concluding the proof of the theorem when @xmath12 $ ] .
suppose now that @xmath8 $ ] is fixed . in order to impose the condition that only the fractions from @xmath36 are involved in the calculations , we employ the fundamental property of neighbor fractions in @xmath80 .
this says that if @xmath97 and @xmath98 are consecutive then @xmath99 .
consequently , @xmath100 ( mod @xmath101 ) , and this allows us to write the fraction @xmath102 in terms of @xmath103 and @xmath101 .
thus @xmath104 this time we have to estimate @xmath105 we may write as @xmath106 where @xmath107 for any @xmath108 $ ] and @xmath109\colon\ \begin{array}{l } \frac{n_{\mathcal{i}}(q)}{|{\mathcal{i}}|q^2\beta_i } < \frac{l_{i-1}\big(\frac { q'}q,\frac { q''}q\big ) l_{i+1}\big(\frac { q'}q,\frac { q''}q\big ) } { k_{i}\big(\frac { q'}q,\frac { q''}q\big ) } < \frac{n_{\mathcal{i}}(q)}{|{\mathcal{i}}|q^2\alpha_i},\ i=1,\dots , h \end{array } \right\}.\ ] ] for the best available technique to estimate the size of @xmath110 one requires bounds for kloosterman sums ( cf . @xcite ) .
this is done when @xmath111 and @xmath112 are intervals , but it may be easily extended for finite unions of subintervals of @xmath113 $ ] ( as the set @xmath114 is ) , even with the same same formula . for our needs here , it suffices a version with a slightly weaker term : @xmath115 inserting into , we get @xmath116 to calculate the sum in , we employ the euler - maclaurin formula , noticing the fact that @xmath117 , as a function of @xmath118 , is piecewise continuous differentiable on @xmath119 $ ] .
we obtain @xmath120 the size of the error term is estimated observing , firstly , that @xmath121 . secondly , by the definition of @xmath114 it follows that there exists a partition of @xmath16 $ ] in finitely many intervals with the property that the cardinality of @xmath114 is monotonic on each of them .
therefore @xmath122 then , forgathering , , in and using again the estimate @xmath123 , we obtain @xmath124 this concludes the proof of the theorem .
for @xmath125 , we have @xmath126 . for @xmath127 , by theorem [ theorem1 ] , it follows that @xmath66 is a countable union of hyper - surfaces in @xmath128 . the support @xmath66 has some striking features .
let us see them in the case @xmath129 .
we write @xmath130 and observe that @xmath131 roughly speaking , by definition we find that @xmath132 corresponds to the set of 3-tuples @xmath133 of consecutive elements of @xmath15 with the property that @xmath134 .
similarly , @xmath135 corresponds to the set of 4-tuples @xmath136 of consecutive elements of @xmath15 with the property that @xmath134 and @xmath137 .
we remark that @xmath138 , and also @xmath139 whenever both @xmath140 and @xmath141 are @xmath142 , except when @xmath143 .
notice that the symmetry of the farey sequence of order @xmath6 with respect to @xmath144 produces a sort of balance between the polygons @xmath135 and @xmath145 .
then the support @xmath66 is the closure of the image of the function @xmath146 , which can be written as @xmath147 in which @xmath148 , @xmath149 , for @xmath150 .
a tedious , but elementary , computation allows us to find precisely the boundaries of @xmath151 .
the image obtained is shown in figure [ figure1 ] and the equations are listed in table [ table1 ] .
all the functions that produce the equations of the boundaries of @xmath151 are either of the form @xmath152 , with @xmath153 in a certain interval that might be unbounded , or the symmetric with respect to @xmath154 of such a curve . here
@xmath155 are integers .
we conclude by making a few remarks .
firstly , we mention that @xmath146 has a symmetrization effect , namely , it makes @xmath156 and to @xmath157 to be symmetric with respect to the first diagonal @xmath158 , for any @xmath159 .
the diamond are close to being , but are not exactly straight lines .
the same applies for the edges of the diamonds in the tail .
] @xmath160 is the single nonempty domain @xmath151 that has @xmath158 as axis of symmetry .
the top of the beak of the swallow @xmath66 has coordinates @xmath161 .
the asymptotes of the wings are @xmath162 and @xmath163 .
the highest density is on a region situated in the neck , where many components of the swallow overlap partially or completely .
table [ table1 ] below lists all the equations of the boundaries of @xmath151 . in the head of the table @xmath164 represents an edge of @xmath135 ( listed in counterclockwise order , starting either from the east or from the north side ) and @xmath165 is a parametrization of @xmath166 .
+ k , l & mn & g__mn(t ) & t + + k , l & mn & g__mn(t ) & t + + 1,2 & ( 13,1 ) ; ( 0,1 ) & & 92 t + 1,2 & ( 0,1 ) ; ( 15,45 ) & & 3 t + 1,2 & ( 15,45 ) ; ( 13,1 ) & & 92 t + 1,3 & ( 12,1 ) ; ( 13,1 ) & & 4 t92 + 1,3 & ( 13,1 ) ; ( 15,45 ) & & 92 t + 1,3 & ( 15,45 ) ; ( 14,34 ) & & 8 t + 1,3 & ( 14,34 ) ; ( 27,57 ) & & 8 t + 1,3 & ( 27,57 ) ; ( 12,1 ) & & 4 t + 1,4 & ( 35,1 ) ; ( 12,1 ) & & t4 + 1,4 & ( 12,1 ) ; ( 27,57 ) & & 4 t + 1,4 & ( 27,57 ) ; ( 13,23 ) & & t9 + 1,4 & ( 13,23 ) ; ( 35,1 ) & & t9 + 1 , l5 & ( , 1 ) ; ( , 1 ) & & t + 1 , l5 & ( , 1 ) ; ( , ) & & t + 1 , l7 & ( , ) ; ( , ) & & t + 1 , l=5,6 & ( , ) ; ( , ) & & t + 1 , l5 & ( , ) ; ( , 1 ) & & t + 2,1 & ( 1,1 ) ; ( 13,23 ) & & 2 t6 + 2,1 & ( 13,23 ) ; ( 25,35 ) & & 6 t + 2,1 & ( 25,35 ) ; ( 1,1 ) & & 2 t + 2,2 & ( 1,45 ) ; ( 1,1 ) & & 2 t + 2,2 & ( 1,1 ) ; ( 25,35 ) & & 2 t + 2,2 & ( 25,35 ) ; ( 12,12 ) & & t8 + 2,2 & ( 12,12 ) ; ( 1,45 ) & & t8 + 2,3 & ( 1,57 ) ; ( 1,45 ) & & t + 2,3 & ( 1,45 ) ; ( 12,12 ) & & t8 + 2,3 & ( 12,12 ) ; ( 45,35 ) & & 8 t + 2,3 & ( 45,35 ) ; ( 1,57 ) & & t + 2,4 & ( 1,23 ) ; ( 1,57 ) & & t6 + 2,4 & ( 1,57 ) ; ( 45,35 ) & & t + 2,4 & ( 45,35 ) ; ( 1,23 ) & & 6 t + 3,1 & ( 1,35 ) ; ( 1,23 ) & & 3 t + 3,1 & ( 1,23 ) ; ( 12,12 ) & & 3 t6 + 3,1 & ( 12,12 ) ; ( 47,37 ) & & 6 t + 3,1 & ( 47,37 ) ; ( 1,35 ) & & t + 3,2 & ( 1,12 ) ; ( 1,35 ) & & t6 + 3,2 & ( 1,35 ) ; ( 47,37 ) & & t + 3,2 & ( 47,37 ) ; ( 35,25 ) & & t + 3,2 & ( 35,25 ) ; ( 1,12 ) & & 6 t + 4,1 & ( 1,37 ) ; ( 1,12 ) & & 4 t + 4,1 & ( 1,12 ) ; ( 35,25 ) & & 4 t + 4,1 & ( 35,25 ) ; ( 23,13 ) & & 9 t + 4,1 & ( 23,13 ) ; ( 1,37 ) & & t9 + 4,2 & ( 1,25 ) ; ( 1,37 ) & & t + 4,2 & ( 1,37 ) ; ( 23,13 ) & & t9 + 4,2 & ( 23,13 ) ; ( 1,25 ) & & t9 + k5 , 1 & ( 1 , ) ; ( 1,2k ) & & k t + k5 , 1 & ( 1,2k ) ; ( , 2k+1 ) & & k t + k5 , 1 & ( , ) ; ( , ) & & t + k5 , 1 & ( , ) ; ( 1 , ) & & t + r. r. hall , g. tenenbaum _ the set of multiples of a short interval _ , in : number theory , new york seminar 19891990 , d. v. chudnovsky , g. v. chudnovsky , h. cohn , m. b. nathanson ( eds . ) , springer , ( 1991 ) , 119128 . | the spacing distribution between farey points has drawn attention in recent years .
it was found that the gaps @xmath0 between consecutive elements of the farey sequence produce , as @xmath1 , a limiting measure .
numerical computations suggest that for any @xmath2 , the gaps @xmath3 also produce a limiting measure whose support is distinguished by remarkable topological features . here
we prove the existence of the spacing distribution for @xmath4 and characterize completely the corresponding support of the measure .
[ multiblock footnote omitted ] |
80 k. j. eskola and h. honkanen , nucl .
phys . a * 713 * , 167 ( 2003 ) ; i. vitev , hep - ph/0212109 ; b. jager _
et al . _ ,
d * 67 * , 054005 ( 2003 ) ; d. de florian , phys . rev .
d * 67 * , 054004 ( 2003 ) ; p. levai _ et al .
_ , nucl - th/0306019 .
g. t. bodwin , phys .
d * 31 * , 2616 ( 1985 ) ; j. c. collins , d. e. soper and g. sterman , nucl .
b * 261 * , 104 ( 1985 ) ; adv .
high energy phys . * 5 * ( 1988 ) 1 ; j. w. qiu and g. sterman , nucl .
b * 353 * , 137 ( 1991 ) ; r. brock _ et al .
_ , rev .
* 67 * , 157 ( 1995 ) . c. adler _ et al .
_ , phys .
lett . * 90 * , 032301 ( 2003 ) ; x. n. wang , phys .
c * 63 * , 054902 ( 2001 ) ; m. gyulassy , i. vitev and x. n. wang , phys . rev .
* 86 * , 2537 ( 2001 ) ; m. gyulassy _ et al .
_ , phys .
b * 526 * , 301 ( 2002 ) ; d. molnar and m. gyulassy , nucl .
a * 697 * , 495 ( 2002 ) ; e. v. shuryak , phys .
c * 66 * , 027902 ( 2002 ) ; i. p. lokhtin _ et al .
_ , phys . atom
. nucl . * 65 * , 943 ( 2002 ) ; a. drees , h. feng and j. jia , nucl - th/0310044 .
i. arsene _ et al .
_ , phys .
lett . * 91 * , 072305 ( 2003 ) ; s. s. adler _ et al .
_ , phys . rev
. lett . *
91 * , 072303 ( 2003 ) ; b. b. back , phys .
lett . * 91 * , 072302 ( 2003 ) ; j. adams _ et al .
_ , phys .
lett . * 91 * , 072304 ( 2003 ) .
y. zhang , _ et al .
_ , phys .
c * 65 * , 034903 ( 2002 ) ; i. vitev , phys .
b * 562 * , 36 ( 2003 ) ; x. n. wang , phys .
b * 565 * , 116 ( 2003 ) ; a. accardi and m. gyulassy , phys .
b * 586 * , 244 ( 2004 ) ; x. f. zhang and g. fai , hep - ph/0306227 .
j. w. qiu and i. vitev , phys .
b * 587 * , 52 ( 2004 ) ; hep - ph/0309094 ; k. j. eskola , v. j. kolhinen and c. a. salgado , eur . phys .
j. c * 9 * , 61 ( 1999 ) ; h. j. pirner and j. p. vary , phys . rev .
lett . * 46 * , 1376 ( 1981 ) .
g. g. barnafoldi _ et al .
_ , hep - ph/0311343 ; g. g. barnafoldi , hep - ph/0312384 ; a. adil , m. gyulassy and i. vitev , in preparation ; a. polleri and f. yuan , nucl - th/0108056 ; t. hirano and y. nara , phys .
c * 68 * , 064902 ( 2003 ) . | i summarize the recent advances in jet tomographic studies of cold and hot nuclear matter based on perturbative qcd calculations of medium - induced gluon bremsstrahlung .
quantitative applications to ultrarelativistic heavy ion reactions at rhic indicate the creation of a deconfined state of qcd with initial energy density on the order of 100 times cold nuclear matter density .
* introduction .
* tomography is the study of the properties of matter through the attenuation pattern of a calibrated flux of fast particles that lose energy via multiple elastic and inelastic scatterings . for ultrarelativistic heavy ion reactions this technique
was first discussed for vector mesons but has only recently been generalized to hard partonic processes which dominate the experimentally accessible moderate- and high-@xmath0 region of hadron production @xcite .
jet tomography extracts the nuclear matter density from the modification to the differential hadron multiplicities @xmath1 .
the overlap function @xmath2 reflects the reaction geometry and the elementary nucleon - nucleon cross sections are calculable to leading and next - to - leading order @xcite based on the successful perturbative qcd factorization approach @xcite .
the potentially dramatic observable effects of jet quenching @xcite , @xmath3 , in @xmath4 versus @xmath5 collisions even for dynamically expanding systems come from the nuclear size enhancement @xmath6 and the large initial energy and parton number densities of the quark - gluon plasma ( qgp ) created in the interaction region .
in addition to the suppression of the single inclusive spectra @xcite and the disappearance of the back - to - back di - hadron correlations @xcite in central and semi - central @xmath7 reactions , parton energy loss plays a major role in the generation of the high-@xmath0 azimuthal asymmetry @xcite and the enhancement of the moderate-@xmath0 baryon to meson ratios @xcite .
the final state origin of many of these effects is now confirmed via the control @xmath8 measurement and its comparison to theoretical predictions @xcite .
the first calculations of the landau - pomeranchuk - migdal ( lpm ) destructive interference effect in qcd @xcite already indicated that medium induced non - abelian bremsstrahlung dominates over the elastic energy loss .
the current techniques for the computation of the gluon radiative spectrum @xmath9 and the mean energy loss @xmath10 have been reviewed in @xcite and include a path integral approach @xcite , an effective 2d schredinger equation @xcite , exact reaction operator formalism @xcite and a twist expansion series @xcite .
for static non - abelian media the mean energy loss depends quadratically on the nuclear size , @xmath11 , and measures the gluon transport coefficient @xmath12 . in dynamically expanding plasmas the radiative spectrum and the mean energy loss
can be related to the soft gluon rapidity density , @xmath13 .
the outlined path length dependence holds for both small and large numbers of scatterings but @xmath14 is more sensitive to the few semihard kicks that hard partons undergo as they traverse the medium @xcite .
major corrections to the above analytic approximations arise from the finite kinematics which is particularly important at sps and rhic and even at the lhc for @xmath15 gev .
these can be numerically evaluated as in @xcite . .
( b ) from @xcite : jet energy dependence of the mean fractional energy loss @xmath16 .
the sensitivity on the plasmon mass @xmath17 and the heavy quark mass @xmath18 is illustrated.,title="fig:",height=144 ] .
( b ) from @xcite : jet energy dependence of the mean fractional energy loss @xmath16 .
the sensitivity on the plasmon mass @xmath17 and the heavy quark mass @xmath18 is illustrated.,title="fig:",width=220,height=153 ] dependence of the partonic quenching factor for lhc energies .
( b ) from @xcite : effective kinematic modification of the fragmentation functions versus @xmath19.,title="fig:",height=144 ] dependence of the partonic quenching factor for lhc energies .
( b ) from @xcite : effective kinematic modification of the fragmentation functions versus @xmath19.,title="fig:",width=201,height=182 ] for qgp in a local thermal equilibrium all dimensional scales can be related to the temperature : @xmath20 , @xmath21 and @xmath22 .
moreover , gluon absorption @xcite can significantly modify the radiative spectrum and the mean energy loss for small jet energies as shown in fig .
[ fig1 - 2]a .
the average energy gain from gluon absorption to first order in opacity , @xmath23 , is power suppressed for large @xmath24 but will manifest itself for @xmath25 via weaker hadron quenching .
the reduction of available phase space for bremsstrahlung off heavy quarks , the dead cone effect @xcite , and the plasmon mass @xmath17 ter - mikayelian effect have been systematically incorporated in the medium induced energy loss formalism @xcite and shown to shorten the gluon formation times , @xmath26 , and regulate the color current propagators , @xmath27 .
the dependence of @xmath28 on the heavy quark mass is illustrated in fig .
[ fig1 - 2]b .
in addition to a reduction of the mean energy loss the path length dependence of @xmath10 in static plasmas was found to be closer to the linear incoherent bethe - heitler regime @xcite .
approximate ways of estimating the observable jet quenching effect on final state hadrons include a suppression of the hard partonic cross section @xcite or an effective attenuation of the fragmentation functions @xcite , @xmath29 .
here @xmath30 is the probability of fractional energy loss @xmath31 due to multiple gluon emission @xcite .
these incorporate the same kinematic modifications to the leading hard parton which originate from the medium induced gluon bremsstrahlung at a timescale @xmath32 .
the former approach is illustrated in fig .
[ fig3 - 4]a for a static qgp with @xmath33 and the latter is exemplified in fig .
[ fig3 - 4]b for @xmath34 .
quantitative applications , however , require the incorporation of energy loss calculations within the full perturbative qcd hadron production formalism @xcite . on the @xmath35 system size in semi - inclusive dis .
hadron quenching is shown versus @xmath36 .
( b ) from @xcite : calculated @xmath37 dependence of nuclear suppression with @xmath38 fixed from drell - yan data.,title="fig:",height=144 ] on the @xmath35 system size in semi - inclusive dis .
hadron quenching is shown versus @xmath36 .
( b ) from @xcite : calculated @xmath37 dependence of nuclear suppression with @xmath38 fixed from drell - yan data.,title="fig:",height=201 ] * applications of jet tomography . *
semi - inclusive deeply inelastic scattering ( dis ) is an ideal probe of final state quark energy loss in cold nuclear matter .
the known lepton @xmath39 and hadron @xmath40 energies allow for an approximate reconstruction of the effective modification to the measured cross section @xmath41 , where @xmath42 .
induced gluon bremsstrahlung was shown to result in an effective shift , @xmath43 , of the fragmentation momentum fraction @xcite .
comparison of hermes data on the nuclear modification for @xmath44 and @xmath45 targets to theoretical calculations in fig .
[ fig5 - 6]b provides strong evidence in support of the @xmath46 dependence of @xmath10 for static systems .
the absorptive power of the medium for quark jets was found to be @xmath47 .
constraints on the cold nuclear matter transport coefficient can also be obtained from the analysis of drell - yan data on initial state broadening and energy loss @xcite .
the extracted @xmath48 can be used to predict the final state hadronic suppression .
figure [ fig5 - 6]b shows the consistency of this approach on the example of the @xmath49-dependent quenching . and @xmath0 dependence of jet quenching for finite kinematics and finite system size .
( b ) from @xcite : comparison of the calculated approximately constant @xmath50 for @xmath51 and @xmath52 to finalized experimental data.,title="fig:",width=172,height=211 ] and @xmath0 dependence of jet quenching for finite kinematics and finite system size .
( b ) from @xcite : comparison of the calculated approximately constant @xmath50 for @xmath51 and @xmath52 to finalized experimental data.,title="fig:",width=172,height=211 ] the initial discovery of a qualitatively large high-@xmath0 particle deficit at rhic @xmath53 gev @xcite has revived the interest in hard physics in relativistic heavy ion reactions .
however , it is the recent advances in the perturbative qcd description of multi - parton dynamics that have taken jet tomography to a new and quantitative level @xcite .
figure [ fig7 - 8]a illustrates the predicted @xmath0 dependence of @xmath50 versus @xmath54 for central @xmath7 collisions @xcite .
the relative importance of the initial state multiple elastic scattering @xcite relative to the final state radiative energy loss was found to decrease from sps to the lhc . at high @xmath55 the @xmath56 shadowing vanishes @xcite but the small effects from the @xmath57 emc region
are expected to persist .
figure [ fig7 - 8]b shows a good agreement of the calculated quenching for @xmath51 and @xmath52 with @xmath58 , comparable to the @xmath59 net hadron rapidity density , with the precision high-@xmath0 data at rhic @xcite . ongoing re - analysis of the sps data @xcite is expected to establish a small hadronic suppression at @xmath60 gev and thus improve the agreement with the perturbative calculation in a medium of finite , @xmath61 , soft parton density .
the centrality and rapidity dependence of jet quenching in @xmath4 reactions has also been addressed both analytically and numerically @xcite . for 1 + 1d expansion hadronic
attenuation scales with the number of participants as @xmath62^{n-2 } $ ] and is insensitive to the effective partonic slope @xmath63 . as a function of rapidity @xmath64 , in agreement with the brahms results @xcite .
corrections to these estimates arise from the interplay of soft and hard physics and the steepening of particle spectra with rapidity @xmath65 .
di - hadron tomography @xcite is one of the new and promising direction of moderate- and high-@xmath0 studies of in - medium jet modification .
the differential cross section @xmath66 is more sensitive to the non - abelian dynamics when compared to the single inclusive measurements . triggering on a large transverse momentum particle biases the hard scatter toward the surface of the interaction region and the away - side partner carries most of the observable effect from the interactions in dense nuclear matter .
for the example of the correlation function , @xmath67 , energy loss reduces the area @xmath68 under the away - side peak .
figure [ fig9 - 10]a show a simulation of the disappearance of the back - to back correlations in a hydro+jet model @xcite which is qualitatively consistent with the data .
important progress in di - jet correlation analysis has been made through recent studies of the attenuation effect relative to the reaction plane @xcite .
it confirms the @xmath69 path length dependence of energy loss in an expanding quark - gluon plasma @xcite .
di - jet acoplanarity reflects the accumulated mean transverse momentum kick @xmath70 of a hard scattered parton pair @xcite and is measurable @xcite via the enhanced away - side width @xmath71 .
preliminary results for @xmath8 and central @xmath7 collisions suggest a significant difference in the observed effect consistent with final state interactions in media of very different density and scattering power @xmath72 .
a large @xmath71 enhancement in central @xmath7 reactions may also indicate a power - law non - gaussian semi - hard multiple scattering @xcite .
modification to the intra - jet correlations for approximately collinear gluon bremsstrahlung .
, title="fig:",height=172 ] modification to the intra - jet correlations for approximately collinear gluon bremsstrahlung .
, title="fig:",height=172 ] * future directions in jet quenching studies . *
improved experimental techniques for jet cone reconstruction in the high - multiplicity environment of relativistic heavy ion reactions @xcite open new possibilities for differential studies of the jet properties .
the transverse momentum broadening of the interacting parent parton+gluon system is expected to modify the intra - jet correlations and widen the jet cone @xcite .
a quantitative measure of this effect will be the deposition of smaller fractional transverse energy @xmath73 inside a cone of opening angle @xmath74 relative to the @xmath5 case .
simulations of this effect in fig .
[ fig9 - 10]b at lhc bulk matter densities show that this effect is very small @xmath75 even for narrow cones @xmath76 @xcite .
perturbative calculations of di - hadron fragmentation also find insignificant alteration of the near - side correlation function @xcite .
while the approximate collinearity of the medium induced gluon bremsstrahlung suggests that modifications of the jet shape may be difficult to measure experimentally , a more promising way to verify the predictions of the non - abelian energy loss theory @xcite may be to search for the `` remnants of lost jets '' @xcite .
the energy radiated off the leading parton is redistributed in the system , possibly within the jet cone itself .
if it is subsequently fully thermalized , the increase in the soft gluon multiplicity can be estimated as @xmath77 .
figure [ fig11 - 12]a shows the momentum density of hadrons @xmath78 associated with back - to - back jets embedded in a parton cascade @xcite and the reappearance of the lost energy at @xmath79 mev . for energetic quark jets.,title="fig:",width=172,height=163 ] for energetic quark jets.,title="fig:",height=153 ] the long formation times @xmath80 , which are the basis of the non - abelian lpm effect , may limit the possibility for secondary interactions of the radiative quanta . in this case one
expects fewer harder gluons with @xmath81 @xcite . as a function of the experimental @xmath82 the accessible @xmath83 , with @xmath84 and the probability distribution
@xmath85 computed numerically as in @xcite .
figure [ fig11 - 12]b shows that already at @xmath86 gev a large fraction of the induced multiplicity is recovered and for @xmath87 and @xmath88 .
i propose that upcoming comparison of the low-@xmath0 experimental data on hadron multiplicities associated with energetic di - jets to theoretical calculations may provide an experimental handle on the degree of thermalization of the medium induced semi - hard bremsstrahlung .
* summary and conclusions . * in this talk i emphasized the quantitative applications of the theory and phenomenology of medium - induced non - abelian energy loss in reactions involving nuclei . in qcd processes
cold nuclear matter can be systematically characterized by its scattering and absorptive power .
table [ table - cold ] , adapted from @xcite , collects results from recent studies of semi - inclusive dis hadron attenuation , drell - yan broadening and the cronin enhancement . within the uncertainties
inherent in the theory of strong interactions different approaches consistently find @xmath89 and final state @xmath90 for large nuclei . in ultrarelativistic heavy ion reactions , results of the jet tomographic analysis @xcite based on the assumptions of local thermal equilibrium and longitudinal bjorken expansion are presented in table [ table - hot ] . for their interpretation , we recall that lattice calculations with improved wilson and staggered fermions suggest @xmath91 mev for 2-flavor qcd , @xmath92 mev for 3-flavor qcd and @xmath93 @xcite .
although the energy density at sps may have reached or even exceeded the critical value for a phase transition , the lifetime of the qgp @xmath94 seems insufficient to clearly develop many of the signatures of deconfinement and collectivity .
in contrast , at rhic @xmath95 , @xmath96 and @xmath97 is comparable to the transverse extent of the system itself .
these initial conditions are in accord with the hydrodynamic estimates @xcite based on the bulk properties of the plasma .
even a sizable theoretical error in the determination of @xmath98 is unlikely to qualitatively change rhic s status on the qcd phase diagram .
one should also note the much higher @xmath99 , @xmath100 and the significantly longer lifetime of the qgp phase expected at the lhc . in conclusion ,
jet tomographic analysis at rhic presents strong evidence for the creation of the deconfined state of qcd .
future jet cone , di - jet correlation and energy balance studies are expected to map out the properties of the plasma and reveal details of the qcd multiparton dynamics in ultradense nuclear matter .
* acknowledgments .
* this work is supported by the united states department of energy under grant no .
de - fg02 - 87er40371 .
@lll & @xmath101 $ ] & @xmath102 $ ] + semi - inclusive dis quenching & 0.12 & 0.5 + drell - yan analysis & 0.14 @xmath103 0.1 & 0.6 @xmath103 0.45 + cronin effect analysis & 0.10 0.14 & 0.4 0.6 + theoretical estimates & 0.05 & 0.2 + @lllllll & @xmath104 $ ] & @xmath105 $ ] & @xmath106 $ ] & @xmath107 $ ] & @xmath98 & @xmath108 $ ] + sps & 0.8 & 1.3 2.3 & 205 245 & 1.2 2.6 & 200 350 & 0.75 1.3 + rhic & 0.6 & 5.5 8 & 360 410 & 12 20 & 800 1200 & 3.2 4.6 + lhc & 0.2 & 13 23 & 710 850 & 170 350 & 2000 3500 & 21 35 + |
throughout the paper , @xmath0 will be a compact oriented surface .
the results automatically extend to non - orientable surfaces by considering their orientation coverings , and also extend to surfaces with cusps through minor rephrasings of the arguments . as indicated in the introduction ,
a hyperbolic metric @xmath1 on @xmath0 defines a transverse measure @xmath7 for the geodesic foliation @xmath5 of @xmath3 .
recall that a transverse measure is the assignment of a radon measure on each surface @xmath23 transverse to @xmath5 , in a way which is invariant under the holonomy of @xmath5 .
equivalently , a transverse measure locally is a measure on the space of leaves of @xmath5 in local charts for @xmath0 , so that these measures agree when charts overlap .
globally , the space of leaves of @xmath0 is not hausdorff , so it is convenient to lift the situation to the universal covering @xmath42 of @xmath0 .
lift the hyperbolic metric @xmath1 to a hyperbolic metric on @xmath43 , which we will still denote by @xmath1 .
the geodesic foliation @xmath44 of @xmath45 defined by this metric projects to the geodesic foliation @xmath5 of @xmath3 . because @xmath42 is simply connected and @xmath1 has negative curvature , every geodesic is globally length minimizing , and it follows that @xmath44 has no recurrent leaves . as a consequence , the space of leaves of @xmath46 is hausdorff .
this space of leaves is also the space @xmath47 of all oriented bi - infinite geodesics of @xmath42 . to work with @xmath48 , it is convenient to consider the _ circle at infinity _ @xmath49 of @xmath42 .
recall that @xmath49 is the quotient of the set of geodesic rays of @xmath42 under the equivalence relation which identifies two rays when they stay at bounded distance from each other .
if we pick a base point @xmath50 in @xmath42 , each equivalence class has a unique representative among the geodesic rays issued from @xmath50 , which identifies @xmath49 to the unit tangent space @xmath51 .
in particular , this identification @xmath52 defines a metric on @xmath49 , the _ angle metric _ based at @xmath50 .
every geodesic of @xmath43 has two distinct end points on the circle at infinity @xmath49 , and any couple of distinct points of @xmath49 is obtained in this way .
this identifies the space @xmath47 of oriented geodesics of @xmath42 to the open annulus @xmath53 , where @xmath54 denotes the diagonal . a _ measure geodesic current _ on @xmath0 , namely a transverse measure for the geodesic foliation @xmath5 in @xmath3 , lifts to a transverse measure for the geodesic foliation @xmath46 in @xmath55 .
it globally defines a measure on the space @xmath56 of leaves of @xmath46 , since no leaf of @xmath57 is recurrent .
note that this measure is invariant under the natural action of the fundamental group @xmath58 .
this identifies the space @xmath10 of measure geodesic currents to the space of @xmath58invariant measures on @xmath47 .
if one replaces the metric @xmath1 by another hyperbolic metric @xmath4 on @xmath0 and the base point @xmath50 by another point @xmath59 , every @xmath4geodesic ray issued from @xmath60 in @xmath42 is quasi - geodesic for @xmath1 , and therefore stays at bounded distance from a unique @xmath1geodesic ray issued from @xmath50 .
this shows that the circle at infinity @xmath49 is actually independent of the hyperbolic metric @xmath1 .
in addition , from the quasi - geodesic property and from easy curvature estimates , the two angle metrics @xmath61 and @xmath62 on @xmath49 which are respectively associated to @xmath1 and the base point @xmath50 and to @xmath4 and the base point @xmath60 are _ hlder equivalent _ in the following sense : there exists constants @xmath63 and @xmath64 such that @xmath65 for every @xmath66 , @xmath67 .
see also @xcite or @xcite .
in particular , @xmath56 inherits from the angle metric of @xmath49 a natural metric well - defined up to hlder equivalence , and for which the action of the fundamental group @xmath58 is hlder continuous . if @xmath68 is a compact subset of @xmath47 and if @xmath64 , let @xmath69 be the set of all hlder continuous functions @xmath70 with support contained in @xmath68 and with hlder exponent @xmath71 .
recall that @xmath21 is _ hlder continuous with hlder exponent _ @xmath720,1\right]$ ] when the norm @xmath73 is finite .
let @xmath74 be the set of all the hlder continuous functions with compact support on @xmath47 , namely @xmath74 is the union of all the @xmath69 as @xmath68 ranges over all compact subsets of @xmath47 and @xmath71 ranges over all numbers in @xmath750,1\right]$ ] .
hlder geodesic current _ on @xmath0 is a linear map @xmath76 which is invariant under the action of @xmath58 , and which is continuous in the sense that its restriction to each @xmath69 is continuous for the norm @xmath77 .
note that this notion is independent of the choice of the negatively curved metric @xmath1 on @xmath0 , because the distance @xmath61 on @xmath47 is well - defined up to hlder equivalence .
let @xmath25 denote the space of hlder geodesic currents , endowed with the weak * topology defined by the family of semi - norms @xmath78 as @xmath21 ranges over all the elements of @xmath74 .
let @xmath79 be a hyperbolic metric on @xmath0 . then there is an orientation - preserving isometry @xmath80 between the universal covering @xmath43 , endowed with the lift of @xmath1 , and the hyperbolic plane @xmath81 .
this isometry induces hlder bicontinuous homeomorphisms @xmath82 and @xmath83 which , to avoid the proliferation of symbols , we will also denote by @xmath84 . then the liouville geodesic current is the measure @xmath85 on @xmath47 which is the pull back by @xmath84 of the liouville geodesic current @xmath86 on @xmath87 . to compute @xmath86 , consider the poincar disk model for the hyperbolic plane @xmath81 , namely the open unit disk @xmath88 with the metric which at @xmath89 is @xmath90 times the euclidean metric .
its circle at infinity @xmath91 then is identified to the unit circle @xmath92 .
if @xmath66 , @xmath93 , let @xmath94 be the oriented geodesic of @xmath95 going from @xmath66 to @xmath96 .
then a simple computation ( see ( * ? ? ?
* a.2 - 3 ) ) shows that the liouville measure @xmath97 on @xmath98 coincides with the measure which is locally @xmath99 at @xmath94 . in the upper half - space model @xmath100 for @xmath81
, the circle at infinity @xmath101 corresponds to @xmath102 .
then , the liouville measure on @xmath87 coincides with the measure which is @xmath103 at the geodesic @xmath94 going from @xmath66 to @xmath104 . the reader should beware of a few minor differences with the setting of @xcite .
the liouville geodesic current @xmath7 considered here is twice that of @xcite , to make it more directly related to the volume form of @xmath3 .
also , @xmath47 denotes here the space of _ oriented _ geodesics of @xmath42 .
we summarize here the basic properties of the shearing coordinates for teichmller space which we will need .
we refer to @xcite for details .
these shearing coordinates are associated to the choice of a maximal geodesic lamination @xmath105 on @xmath0 .
recall that , for an arbitrary metric of negative curvature on @xmath0 , a _
geodesic lamination _ is a closed subset of @xmath0 which can be decomposed as a union of disjoint complete geodesics which have no self - intersection points .
such a notion is actually a topological object , independent of the metric , in the sense that there is a natural identification between @xmath1geodesic laminations and @xmath4geodesic laminations for any two negatively curved metrics @xmath1 and @xmath4 . a geodesic lamination is _ maximal _ if it is maximal for inclusion among all geodesic laminations , which is equivalent to the property that the complement @xmath106 consists of finitely many infinite triangles .
a fundamental example of a maximal geodesic lamination is obtained as follows .
start with a family @xmath107 of disjoint simple closed geodesics decomposing @xmath0 into pairs of pants .
each pair of pants can be divided into two infinite triangles by three infinite geodesics spiralling around some boundary components .
the union of @xmath107 and of these spiralling geodesics forms a maximal geodesic lamination @xmath105 .
the shearing coordinates belong to the space @xmath108 of _ transverse cocycles _ for the maximal geodesic lamination @xmath105 , namely of finitely additive signed transverse measures ( valued in @xmath19 ) for @xmath105 .
the vector space @xmath108 has finite dimension @xmath109 ; see ( * ? ? ?
* theorem 15 ) .
the analogy in notation between the space @xmath110 of transverse cocycles for the geodesic lamination @xmath105 and the space @xmath25 of hlder geodesic currents on the surface @xmath0 is not a coincidence , because it is proved in @xcite that @xmath110 is in a natural way a subspace of @xmath25 ; however , this is irrelevant here .
the shearing coordinates for teichmller space define an open embedding @xmath111 which associates to a hyperbolic metric @xmath14 its _ shearing cocycle _
@xmath112 @xcite .
it is proved in ( * ? ? ?
* theorem a ) that this map defines a real analytic diffeomorphism from @xmath113 to an open subset of @xmath110 .
we will use a relatively explicit description of the inverse of the map @xmath114 , as constructed in @xcite .
let @xmath115 , let @xmath1 be another hyperbolic metric which is near @xmath116 , and let @xmath117 , @xmath112 be their respective shearing cocycles .
then the isometries @xmath118 and @xmath119 considered in section [ sect : liouv ] can be chosen so that the induced map @xmath120 is of the type which we now describe .
unfortunately , this will require some care in the definitions as well as relatively cumbersome notation .
use the isometry @xmath121 to identify the preimage @xmath122 of @xmath105 in @xmath42 to a geodesic lamination of @xmath81 , which we will also denote by @xmath122 .
pick a base point @xmath123 which is disjoint from @xmath124 .
this defines a partial order on the set of components of @xmath125 which are different from the component @xmath126 containing @xmath50 where , for two such components @xmath127 and @xmath128 , @xmath129 exactly when @xmath127 separates @xmath50 from @xmath128 . : @xmath130 , @xmath131 and @xmath132 are not comparable , and @xmath133 is a spanning family . ]
orient the leaves of @xmath124 to the left , as seen from the base point @xmath50 .
for such a leaf @xmath134 and for @xmath135 , let the _ elementary earthquake _ , or _ elementary shearing map _ @xmath136 coincide with the identity on the component of @xmath137 that contains @xmath50 and with the hyperbolic translation of @xmath138 along the oriented geodesic @xmath134 on the other component of @xmath139 .
note that , although @xmath140 does not continuously extend to @xmath81 , it has a continuous extension to a _
@xmath141 , which we also denote by @xmath140 . for a component @xmath127 of @xmath142 which is not the component @xmath126
containing @xmath50 , the boundary of @xmath127 consists of three leaves @xmath143 , @xmath144 and @xmath145 of @xmath124 since @xmath105 is a maximal geodesic lamination . choose the indexing so that @xmath143 , @xmath144 and @xmath145 occur in this order as one goes counterclockwise around @xmath146 , and so that @xmath145 is in the boundary of the component of @xmath147 that contains @xmath50 . for @xmath135 , define @xmath148 .
set @xmath149 .
note that @xmath150 also defines a transverse cocycle for @xmath124 .
if @xmath127 is a component of @xmath142 , let @xmath151 be the number associated by this transverse cocycle to a geodesic arc joining the base point @xmath50 to an arbitrary point of @xmath127 .
spanning family _ of components of @xmath142 be a finite family @xmath152 of components of @xmath142 different from the component @xmath126 containing @xmath50 such that no two @xmath153 are comparable for @xmath154 , namely such that no @xmath153 separates @xmath50 from another @xmath155 .
we will write @xmath156 when @xmath157 for some @xmath158 .
for such a spanning family @xmath159 , consider the infinite non - commutative product @xmath160 defined as the limit of @xmath161 as the finite family @xmath162 tends to the set of all @xmath163 with @xmath156 and where , for every @xmath164 , either @xmath165 or @xmath166 and @xmath167 are not comparable ( namely @xmath168 ) . note that @xmath169 and @xmath170 commute when @xmath166 and @xmath167 are not comparable , so that the order does not matter in this case . for the same reason ,
the order is irrelevant in the product of the ( finitely many ) @xmath171 with @xmath172 .
the arrow on top of the first product symbol serves as a reminder that the terms must appear in an order compatible with the partial order @xmath154 . the fact that the limit exists is proved in @xcite .
the set up in that paper is slightly different , but the restriction of @xmath173 to a component @xmath127 of @xmath174 coincides with what is called @xmath175 in @xcite if @xmath156 , with @xmath176 if there is an @xmath177 such that @xmath178 , and with the identity otherwise .
we now let the spanning family @xmath159 _ uniformly tend to infinity _ , in the sense that the set @xmath179 converges to the set of all components @xmath163 of @xmath142 .
then @xmath180 converges to a map @xmath181 which coincides with a hyperbolic isometry on each component of @xmath142 .
the convergence is immediate from the fact that @xmath180 and @xmath182 coincide on the component @xmath127 when @xmath156 and @xmath183 .
the map @xmath184 is the _ shear map _ associated to @xmath124 and to the transverse cocycle @xmath150 .
in general , the shear map @xmath184 has no continuous extension to @xmath81 .
however , it conjugates the action of @xmath58 on @xmath185 to a group @xmath186 of isometries of @xmath81 whose action is free and properly discontinuous ; see @xcite . as a consequence
, @xmath184 continuously extends to a homeomorphism @xmath187 .
the fact that @xmath188 is the shearing cocycle of the metric @xmath29 means that the isomorphism @xmath189 defined by @xmath184 can be realized by an isometry between @xmath190 and the quotient @xmath191 .
in other words , the isometry @xmath84 can be chosen so that the homeomorphisms @xmath184 and @xmath120 coincide .
in addition , as @xmath159 uniformly converges to infinity , the homeomorphism @xmath192 uniformly converges to @xmath187 .
the homeomorphisms @xmath193 , @xmath194 , @xmath195 , @xmath184 of @xmath196 induce homeomorphisms of @xmath197 , which we will denote by the same symbols .
let @xmath198 , @xmath34 -\varepsilon , \varepsilon\right[$ ] , be a differentiable curve in the teichmller space @xmath9 .
we want to show that the derivative @xmath199 exists as a hlder geodesic current .
namely , for a hlder continuous function @xmath200 with compact support , we want to show that the derivative @xmath201 exists and depends continuously on @xmath21 .
let @xmath202 be the shearing cocycle associated to the hyperbolic metric , and set @xmath203 .
then , @xmath204 where , as in section [ sect : shearing ] , the homeomorphism @xmath205 is defined as the limit of the homeomorphisms @xmath206 as the spanning family @xmath159 uniformly tends to infinity .
for a fixed component @xmath127 of @xmath207 , the number @xmath208 is a linear function of the transverse cocycle @xmath202 .
since the shearing cocycle @xmath209 depends differentiably on the hyperbolic metric @xmath210 , it follows that the derivative @xmath211 exists and depends only on @xmath127 and on the tangent vector @xmath212 .
if we formally differentiate the infinite product of equation [ eqn : defpartialshear ] , we can expect that @xmath213 where the question mark on the @xmath214 sign is here to remind us that this is only a conjectural formula , where the contribution of a geodesic @xmath215 is @xmath216 and where the contribution of a component @xmath127 of @xmath217 is @xmath218 since @xmath148 by definition .
we will see that , as we let the spanning family @xmath159 uniformly tend to infinity , the last term @xmath219 of equation [ eqn : conjpartialshear ] vanishes in the limit .
therefore , the expected formula for the derivative of @xmath220 is @xmath221 where the sum is over all components @xmath127 of @xmath217 that are different from the component @xmath126 containing the base point @xmath50 .
the main technical point of our arguments is to prove that the infinite sums of equations [ eqn : conjfullshear ] and [ eqn : conjpartialshear ] do converge and , unlike the above formal computations , will strongly use the hypothesis that the function @xmath21 is hlder continuous .
we first compute the derivative @xmath222 corresponding to the elementary earthquake along a single geodesic @xmath134 .
this will not require much regularity for @xmath21 .
[ thm : elemshear ] for an oriented geodesic @xmath215 , let @xmath222 be defined by equation [ eqn : defc(g ) ] . then , for every continuous function @xmath223 with compact support , @xmath224 where @xmath225 denotes the angle from @xmath134 to @xmath226 at their intersection point , measured counterclockwise for the metric @xmath116 and without taking the orientation of the geodesics into account , and where by convention @xmath227 if @xmath134 and @xmath226 do not meet . before going any further
, we will introduce a restriction on the function @xmath21 which will somewhat simplify the exposition .
the liouville measures @xmath7 are invariant under the involution @xmath228 of @xmath47 defined by reversing the orientation of the geodesics . as a consequence ,
the integral of @xmath21 for @xmath7 is equal to the integral of the symmetrized function @xmath229 .
the same applies to the integral of equation [ eqn : elemshear ] .
we can therefore restrict attention to those functions @xmath21 which are _ balanced _ in the sense that they are invariant under the involution @xmath228 .
choose the isometric identification @xmath230 so that , in the upper half - space model for @xmath81 , the oriented geodesic @xmath134 goes from the point @xmath231 to the point @xmath232 .
then , on @xmath233 , the homeomorphism @xmath193 is such that @xmath234 for @xmath235 and @xmath236 for @xmath237 .
since only those geodesics which cross @xmath134 contribute to the derivative , it follows that @xmath238 where @xmath94 denotes the geodesic joining @xmath66 to @xmath96 . here
the factor @xmath239 has two origins : one factor @xmath240 comes from the expression of the liouville geodesic current in section [ sect : liouv ] ; another @xmath240 comes from the fact that we have to consider the two possible orientations for each geodesic crossing @xmath134 , and that each orientation contributes the same amount by our assumption that @xmath21 is balanced .
an elementary computation in hyperbolic geometry shows that the term @xmath241 is equal to the cosine of the angle from the geodesic @xmath134 joining @xmath231 to @xmath232 to the geodesic @xmath94 joining @xmath66 to @xmath96 .
this concludes the proof .
throughout this section , @xmath0 is endowed with the hyperbolic metric @xmath116 , and @xmath105 is realized by an @xmath116geodesic lamination .
we pick a base point @xmath242 which is disjoint from the preimage @xmath124 of @xmath105 .
endow the circle at infinity @xmath49 with the angle metric as seen from the point @xmath50 , and the space @xmath243 of oriented geodesics of @xmath42 with the corresponding product metric . for a hlder continuous function @xmath244
, its hlder norm @xmath245 will be computed with respect to this metric on @xmath47 .
consider a component @xmath127 of @xmath246 which is not the component @xmath126 containing the base point @xmath50 .
let @xmath247 denote the distance from @xmath50 to @xmath127 , namely the distance from @xmath50 to the point @xmath248 of the geodesic @xmath145 that is closest to @xmath50 .
let @xmath249 be the orthogonal projection to @xmath145 of the vertex of the triangle @xmath127 that separates @xmath143 from @xmath144 .
finally , let @xmath250 be the signed distance from @xmath248 to @xmath249 in the oriented geodesic @xmath145 , namely @xmath251 is equal to the distance from @xmath248 to @xmath249 and @xmath252 exactly when @xmath249 is to the right of @xmath248 as seen from @xmath50 . and @xmath253 .
] the key ingredient of the proof of theorem [ thm : mainthm ] is the following estimate for the quantity @xmath254 defined by equation [ eqn : defc(t ) ] .
we write @xmath255 when the ratio @xmath256 is bounded above by a constant , and we write @xmath257 when both @xmath255 and @xmath258 .
[ thm : estc(t ) ] let @xmath259 be a hlder continuous function with hlder exponent @xmath71 , and such that every geodesic @xmath260 in the support of @xmath21 meets the ball @xmath261 .
then , @xmath262 where the constant hidden in the symbol @xmath263 depends only on the radius @xmath264 and on a number @xmath265 such that @xmath124 is disjoint from @xmath266 .
we can rephrase the estimate of proposition [ thm : estc(t ) ] as follows .
the hyperbolic triangle @xmath127 has a natural _ center _
@xmath267 , defined for instance as the point that is at equal distance from the three sides of @xmath127 .
[ thm : esto_t ] @xmath268 by elementary hyperbolic geometry , the distance @xmath269 is within @xmath270 of @xmath271 which , by consideration of the right angled hyperbolic triangle @xmath272 , is itself within @xmath273 of @xmath274 . the estimate of proposition [ thm : estc(t ) ] can therefore be rewritten as @xmath275 we now begin the proof of proposition [ thm : estc(t ) ] , which will occupy most of this section .
we split the argument in two cases , according to whether the distance @xmath247 from @xmath50 to @xmath127 is larger than @xmath276 or not .
choose the isometry @xmath121 between @xmath42 , endowed with the metric @xmath116 , and the poincar disk model for the hyperbolic space @xmath81 , bounded by the unit circle @xmath92 in @xmath277 , so that @xmath121 sends the base point @xmath242 to the center @xmath278 of the disk .
we use @xmath121 to identify @xmath42 to @xmath81 so that , in this identification @xmath279 , the map @xmath121 is now the identity
. we will consequently use the same symbol to represent an object in @xmath42 and the corresponding object in @xmath95 . as in the proof of lemma [ thm : elemshear ]
, we can assume without loss of generality that @xmath21 is balanced , namely invariant under the involution @xmath228 of @xmath47 defined by reversing the orientation of geodesics .
every geodesic @xmath226 which contributes to the integrals in @xmath254 must be in the support of @xmath21 and must cross at least one of the geodesics @xmath280 . because of our assumption that @xmath281 and by choice of @xmath264 , such an @xmath226 will necessarily cross @xmath145 .
the vertices of the ideal triangle @xmath127 decompose the circle @xmath282 into three intervals @xmath283 , @xmath284 and @xmath285 , where @xmath286 has the same end points as the geodesic @xmath280 . if we combine the definition of @xmath254 in equation [ eqn : defc(t ) ] , the computation of lemma [ thm : elemshear ] for @xmath222 , the expression for the liouville geodesic current @xmath86 given in section [ sect : liouv ] , and the fact that @xmath21 is balanced , it follows that @xmath287 where @xmath94 denotes the geodesic with end points @xmath66 , @xmath93 .
it therefore suffices to bound the quantity @xmath288 for every @xmath289 .
( the factor @xmath240 is here to avoid dragging cumbersome constants in the computations below . )
we now switch to the upper - half space model for @xmath81 , bounded by the real line @xmath290 in @xmath277 , in such a way that the base point @xmath50 now corresponds to the point @xmath291 and in such a way that the point @xmath96 is sent to @xmath232 .
the measure @xmath292 on @xmath92 is then sent to the measure @xmath293 on @xmath290 . in this context , this gives the expression @xmath294 where , as before , @xmath295 is the geodesic going from @xmath66 to @xmath296 , namely the vertical euclidean line passing through @xmath66 , and where @xmath138 and @xmath297 correspond to the end points of @xmath285 , with @xmath298 .
let @xmath299 correspond to the third vertex of @xmath127 .
if we use elementary euclidean geometry to compute @xmath300 and if we write @xmath301 , then @xmath302 note that @xmath303 therefore , @xmath304 if we set @xmath305 . integrating in @xmath66 ,
we conclude that @xmath306}\left\vert\eta(x)\right\vert\ ] ] we now use the fact that @xmath307 is hlder continuous with respect to the angle metric on @xmath92 , with hlder exponent @xmath720,1\right]$ ] and with hlder norm @xmath245 .
as we switch from the disk model to the upper half - space model for @xmath81 by an isometry sending @xmath231 to @xmath308 , the angle metric on @xmath92 becomes the metric @xmath309 on @xmath19 .
it follows that the function @xmath310 is hlder continuous with respect to the usual metric @xmath311 of @xmath19 , with hlder exponent @xmath71 and with hlder norm @xmath312 . in particular , because @xmath313 , we have that @xmath314 for every @xmath66 .
it follows that @xmath315}\left\vert\eta(x)\right\vert$ ] is bounded by @xmath316 . combining with equation [ eqn : esti(infty ) ]
, we obtain that @xmath317 we now estimate @xmath247 and @xmath253 in terms of @xmath138 , @xmath318 and @xmath319 .
[ thm : estu_tandd_t ] if @xmath320 and if @xmath127 is at distance at least @xmath276 from the base point @xmath50 , then : 1 .
@xmath321 2 .
@xmath322 and @xmath323 when @xmath324 3 .
@xmath325 and @xmath326 when @xmath327 . on the geodesic @xmath145 , with end points @xmath138 and @xmath318
, we already have two preferred points , namely the orthogonal projection @xmath248 of @xmath328 and the orthogonal projection @xmath249 of the vertex @xmath319 . recall that @xmath247 is the distance from @xmath50 to @xmath248 , and that @xmath253 is the signed distance from @xmath248 to @xmath249 in the oriented geodesic @xmath145 .
consider a third point , namely the orthogonal projection @xmath329 of the point @xmath232 to @xmath145 . in euclidean terms , @xmath329 is the point of the euclidean semi - circle @xmath330 that has the largest imaginary part , namely @xmath331 . , with @xmath252 . ]
the points @xmath332 such that the geodesic joining @xmath66 to @xmath232 meets the disk @xmath333 form an interval @xmath334 $ ] , where @xmath335 depends only on @xmath264 .
if @xmath320 , then the intervals @xmath336 $ ] and @xmath334 $ ] have a non - trivial intersection . since , in addition , @xmath337 is disjoint from the interior of the ball @xmath338 , we conclude that @xmath138 and @xmath318 stay in a closed interval @xmath339 $ ] where , again , @xmath340 depends only on @xmath264 .
consider the horocycle centered at @xmath232 passing through @xmath50 , namely the euclidean horizontal line passing through @xmath328 .
let @xmath341 be the point of this horocycle which lies on the same vertical line as @xmath329 , namely @xmath342 .
the piece of horocycle joining @xmath50 to @xmath341 has length @xmath343 .
it follows that the distance from @xmath50 to @xmath341 is bounded by @xmath340 .
also , the orthogonal projection of @xmath341 to @xmath145 is equal to @xmath329 , the orthogonal projection of @xmath50 to @xmath145 is equal to @xmath248 , and the orthogonal projection map is distance non - increasing ; it follows that the distance from @xmath248 to @xmath329 is also bounded by @xmath340 . by definition
, @xmath247 is equal to the distance from @xmath50 to @xmath248 . from the above observations and the triangle inequality
, it follows that @xmath247 is within @xmath344 of the distance from @xmath342 to @xmath331 . since this distance is equal to @xmath345 , and
since @xmath346 this proves that @xmath347 .
similarly , @xmath253 is equal to the signed distance from the point @xmath248 to @xmath249 in @xmath145 , and is therefore within @xmath340 of the signed distance @xmath348 from @xmath329 to @xmath249 . consider the isometry @xmath349 of @xmath81 sending @xmath138 , @xmath318 , @xmath232 to @xmath232 , @xmath231 , @xmath350 , respectively .
then @xmath349 sends @xmath319 to @xmath351 . from the invariance of cross - products under @xmath349
, we conclude that @xmath352 namely @xmath353 . it follows that @xmath354 if @xmath355 , then @xmath348 admits a uniform lower bound and @xmath356 admits a uniform upper bound , so that @xmath357 .
it follows that @xmath322 .
this implies @xmath358 similarly , when @xmath327 , then @xmath359 and @xmath360 this concludes the proof of lemma [ thm : estu_tandd_t ] .
we now return to our estimate @xmath361 from equation [ eqn : est(infty ) ] .
when @xmath355 , lemma [ thm : estu_tandd_t ] and the fact that @xmath362 imply @xmath363 the same estimate holds when @xmath327 by a symmetric argument .
if we go back to the beginning of the proof , this shows that @xmath364 for every @xmath365 .
then , @xmath366 this concludes the proof of proposition [ thm : estc(t ) ] when @xmath127 is at distance at least @xmath276 from the base point @xmath50 .
note that , in this case , we only need to show that @xmath367 . also , we again can assume that @xmath21 is balanced , without loss of generality . in this case
, we have an additional term because there can be geodesics @xmath226 which contribute to @xmath254 but which do not meet the geodesic @xmath145 .
these additional geodesics must cross both @xmath143 and @xmath144 , and therefore must have one end point in the interval @xmath283 and the other one in @xmath284 .
so , @xmath254 splits as a sum @xmath368 with @xmath369 and @xmath370 using the fact that @xmath21 is balanced . as in the first case ,
to bound the first integral @xmath371 , it suffices to bound the quantity @xmath372 for every @xmath289 . for future reference ,
let us note the following easy fact .
[ thm : i(y)bounded ] @xmath373 if @xmath66 contributes to @xmath374 , then @xmath94 meets the disk @xmath333 , so that @xmath375 is bounded away from 0 .
let us switch again to the upper half - space model for @xmath95 , in such a way that @xmath96 corresponds to @xmath232 and @xmath50 corresponds to @xmath308 .
then @xmath374 becomes @xmath376 with @xmath377 with @xmath301 ( no need to introduce a function @xmath378 now ) . if we again note that @xmath379 for every @xmath66 and if we integrate in @xmath66 , it follows that @xmath380 let us relate this estimate to @xmath381 .
[ thm : estu_tandd_tbis ] if @xmath127 is at distance at most @xmath276 and at least @xmath265 from the base point @xmath50 , 1 .
@xmath382 when @xmath324 2 .
@xmath383 when @xmath327 .
consider the case where @xmath384 .
the other case is similar .
let @xmath385 be the hyperbolic rotation around the point @xmath328 that sends @xmath138 to 0 .
namely , @xmath386 .
the distance from @xmath50 to the geodesic @xmath387 is equal to the distance from @xmath50 to @xmath145 and , by hypothesis , is therefore bounded between the two positive constants @xmath228 and @xmath276 . since the end points of @xmath387 are @xmath388 and @xmath389 , it follows that there are two constants @xmath390 such that @xmath391 $ ] .
namely @xmath392 .
since @xmath393 and @xmath389 stay bounded , we can apply to @xmath394 the arguments of the proof of lemma [ thm : estu_tandd_t ] .
note that @xmath395 is equal to the distance from @xmath396 to @xmath397 . on the euclidean semi - circle @xmath387 ,
let @xmath329 be the point that has the largest imaginary part and let @xmath348 be the signed distance from @xmath397 to @xmath329 in @xmath387 . then , as in the proof of lemma [ thm : estu_tandd_t ]
, @xmath253 is within a uniformly bounded distance of @xmath348 and @xmath398 since @xmath388 . because @xmath327 , it follows that @xmath399 , and therefore that @xmath400 .
this proves @xmath401 this concludes the proof of lemma [ thm : estu_tandd_tbis ] when @xmath327 .
the other case is similar , using the hyperbolic rotation around @xmath50 that sends @xmath318 to 0 .
we now return to @xmath376 .
we want to prove that @xmath402 .
the support of @xmath403 is contained in the interval @xmath334 $ ] consisting of those @xmath66 such that the geodesic joining @xmath66 to @xmath232 meets the disk @xmath333 .
therefore , if @xmath320 , the interval @xmath336 $ ] must meet @xmath334 $ ] .
also , if @xmath336 $ ] contains @xmath404 $ ] , its length @xmath405 can not be too large , as @xmath328 would otherwise be on the wrong side of the oriented geodesic @xmath145 going from @xmath138 to @xmath318 .
therefore , there exists @xmath406 , depending only on @xmath264 , such that at least one of @xmath138 , @xmath318 belongs to the interval @xmath339 $ ] .
let us restrict attention , without loss of generality , to the case where @xmath138 is in @xmath339 $ ] .
we now split the argument in two cases , according to whether @xmath407 or @xmath355 .
first consider the case where @xmath407 .
since @xmath376 is bounded by lemma [ thm : i(y)bounded ] , we can restrict attention to the case where @xmath395 is uniformly large , namely where , by lemma [ thm : estu_tandd_tbis ] , @xmath408 is uniformly small .
note that @xmath409 so that @xmath410 if @xmath411 ( since @xmath412 ) . by equation [ esti(infty)bis ] and lemma [ thm : estu_tandd_tbis ] , it then follows that @xmath413 which concludes the proof that @xmath414 in this case .
now consider the case where @xmath355 .
again , by lemmas [ thm : i(y)bounded ] and [ thm : estu_tandd_tbis ] , we can restrict attention to the case where @xmath395 is uniformly large , namely where @xmath415 is uniformly small .
if @xmath318 is uniformly bounded , say @xmath416 , the same argument as above gives that @xmath417 if @xmath418 , so that @xmath419 in the other case where @xmath420 , then @xmath421 since @xmath422 .
note that @xmath423 if @xmath424 .
( for the first inequality , note that the function @xmath425 is monotone increasing ) .
we now need to go back to the original expression for @xmath376 in equation [ eqn : formi(infty ) ] . if we remember that the support of @xmath403 is contained in @xmath404 $ ] , @xmath426 by lemma [ thm : estu_tandd_tbis ] and using the fact that @xmath427 is uniformly bounded when @xmath428 .
this proves that @xmath429 in all cases .
therefore @xmath430 for every @xmath365 , and @xmath431 we now consider the other term @xmath432 since @xmath433 is bounded away from 0 when the geodesic @xmath434 belongs to the support of @xmath21 .
[ thm : lengthest ] 1 .
@xmath435 and @xmath436 when @xmath437 2 .
@xmath438 and @xmath439 when @xmath440 . by symmetry
, we can restrict attention to the case where @xmath407 without loss of generality .
the length of @xmath441 is equal to the angle under which @xmath283 is seen from @xmath50 .
since the projection @xmath248 of @xmath50 to @xmath145 stays in the ball @xmath338 , this angle is comparable to the angle @xmath442 under which @xmath283 is seen from @xmath248 , namely @xmath443 .
considering the right - angled hyperbolic triangle whose vertices are @xmath248 , the vertex @xmath319 of @xmath127 that separates @xmath283 from @xmath284 , and the projection @xmath249 of @xmath319 to @xmath145 , elementary hyperbolic trigonometry gives @xmath444 since @xmath395 is equal to the hyperbolic distance from @xmath248 to @xmath249 by definition . because the distance from @xmath50 to @xmath145 is bounded , @xmath445 , so that @xmath436 .
this concludes the proof of lemma [ thm : lengthest ] . from lemma [ thm : lengthest ] and our prior estimate
, we obtain @xmath446 since we already showed that @xmath447 , this proves @xmath448 which concludes the proof of proposition [ thm : estc(t ) ] when @xmath449 , and therefore in all cases .
we will need a refinement of proposition [ thm : estc(t ) ] , which is automatically given by the above proof .
indeed , it did not use the full force of the hypothesis that @xmath21 is hlder continuous .
the hlder continuity of @xmath21 did not come up at all in the case where @xmath449 , where only the @xmath450 norm @xmath451 was used . in the case where @xmath452 , the only occurrence of the hlder continuity was just above equation [ eqn : est(infty ) ] , and more precisely in the statement that @xmath453 for every @xmath454 $ ] , with the notation used there .
because of the estimate that @xmath455 from lemma [ thm : estu_tandd_t ] , we can replace this inequality by the stronger ( up to a constant ) statement that @xmath456 where , for a geodesic @xmath457 which is disjoint from the base point @xmath50 , @xmath458 is defined by @xmath459 where the second supremum is taken over all geodesics @xmath226 , @xmath460 which both cross the geodesic @xmath134 and have the same end point on the side of @xmath134 facing the base point @xmath50 , and where @xmath461 is the distance from @xmath50 to @xmath134 .
then the proof of proposition [ thm : estc(t ) ] automatically gives : [ thm : estc(t)bis ] let @xmath259 be a hlder continuous function with hlder exponent @xmath71 , and such that every geodesic @xmath260 in the support of @xmath21 meets the ball @xmath261 .
then , @xmath462 where the constant hidden in the symbol @xmath263 depends only on the radius @xmath264 and on a number @xmath265 such that @xmath124 is disjoint from @xmath266 .
we will need a similar estimate for the quantity @xmath463 associated to the geodesic @xmath464 as in equation [ eqn : defc(g ) ] .
again , let @xmath461 be the distance from @xmath134 to the base point @xmath50 , measured in the metric @xmath116 .
[ thm : estc(g ) ] let @xmath259 be a hlder continuous function with hlder exponent @xmath71 , and such that every geodesic @xmath260 in the support of @xmath21 meets the ball @xmath261 .
then , @xmath465 where the constant hidden in the symbol @xmath263 depends only on the radius @xmath264 .
the proof is very similar to , and much simpler than , the first case of the proof of proposition [ thm : estc(t ) ] .
we consequently omit it .
this section is devoted to proving that the conjectural formula of equation [ eqn : conjfullshear ] really makes sense .
the arguments of section [ sect : mainest ] were exclusively based in the universal cover @xmath466 . in this section , we will strongly use the action of @xmath58 and the fact that the quotient space @xmath0 of @xmath43 under this action is compact .
[ thm : sumcv ] the sum @xmath467 is convergent , and its absolute value is bounded by @xmath468 for a constant @xmath319 depending only on the hlder exponent @xmath71 and on a compact subset containing the support of @xmath21 ( as well as on the metric @xmath469 and on the transverse cocycle @xmath470 ) . by proposition [ thm : estc(t ) ] , it suffices to show that @xmath471 converges .
let @xmath472 be a train track carrying @xmath105 . recall that a _ ( fattened ) train track _
@xmath473 on the surface @xmath0 is a family of finitely many ` long ' rectangles @xmath474 , ,
@xmath475 which are foliated by arcs parallel to the ` short ' sides and which meet only along arcs ( possibly reduced to a point ) contained in their short sides .
in addition , a train track @xmath473 must satisfy the following : 1 .
each point of the ` short ' side of a rectangle also belongs to another rectangle , and each component of the union of the short sides of all rectangles is an arc , as opposed to a closed curve ; 2 .
note that the closure @xmath476 of the complement @xmath477 has a certain number of ` spikes ' , corresponding to the points where at least 3 rectangles meet ; we require that no component of @xmath476 is a disc with 0 , 1 or 2 spikes or an annulus with no spike .
the rectangles are called the _
edges _ of @xmath473 .
the foliations of the edges of @xmath473 induce a foliation of @xmath473 , whose leaves are the _ ties _ of the train track .
the finitely many ties where several edges meet are the _ switches _ of the train track @xmath473 . a tie which is
not a switch is _
generic_. the geodesic lamination @xmath478 is _ carried _ by the train track @xmath473 if it is contained in the interior of @xmath473 and if its leaves are transverse to the ties of @xmath473 .
there are several constructions which provide a train track @xmath473 carrying @xmath105 ; see for instance @xcite@xcite .
let @xmath479 be the preimage of @xmath473 in the universal covering @xmath42 .
choose an orientation for the ties of each edge of @xmath473 ( with no requirement that these orientations match at the switches ) , and lift this orientation to all the edges of @xmath479 .
for each edge @xmath480 of @xmath479 , let @xmath481 be the component of @xmath246 that contains the negative end points of the ties of @xmath482 , for this choice of orientation .
recall that , for a component @xmath163 of @xmath207 , @xmath248 is the point of the closure of @xmath127 that is closest to the base point @xmath50 . to simplify the exposition
, we can arrange by a small perturbation of @xmath473 that no @xmath248 in contained in a switch of @xmath479 . as a consequence , each @xmath248
is contained in a unique edge @xmath480 of @xmath479 .
then , @xmath483 where the sums @xmath484 are over all edges of @xmath479 and where the sums @xmath485 are over those components @xmath127 of @xmath246 for which @xmath248 is in @xmath482 .
the distance @xmath247 is also the distance from @xmath248 to the base point @xmath50 .
it follows that , if @xmath486 , then @xmath487 where @xmath488 is the distance from @xmath50 to @xmath482 and where the constants hidden in the symbol @xmath489 depend only on the diameters of the edges @xmath480 , which are uniformly bounded by equivariance under the action of @xmath58 . if @xmath486 , consider the _ divergence radius _ @xmath490 of the component @xmath491 of @xmath492 , defined as follows . if @xmath491 is one of the two components of @xmath493 that touch the boundary @xmath494 , then @xmath495
otherwise , @xmath491 is delimited by two geodesic arcs contained in two of the three leaves @xmath280 , @xmath496 , of @xmath124 bounding @xmath127 .
these two leaves are asymptotic on one side of @xmath480 . on the other side , they cross the same succession of @xmath228 edges of @xmath497 before taking diverging routes at some switch of @xmath497 .
then , @xmath490 is defined as this number @xmath228 of edges which the two leaves cross together before diverging . a consequence of the combinatorics of transverse cocycles , proved in ( * ? ? ?
* lemma 6 ) , is that @xmath498 where @xmath499 and where we absorb this quantity @xmath500 in the constant of the second @xmath263 .
the complement @xmath106 has finitely many components ; see for instance @xcite@xcite . as a consequence ,
there are only finitely many @xmath127 modulo the action of @xmath58 .
it follows that , when we consider the point @xmath249 projection to @xmath145 of the third vertex of @xmath127 , the distance from @xmath249 to @xmath501 is uniformly bounded . from the definition of @xmath490 and the definition of @xmath253 as the signed distance from @xmath248 to @xmath249 , we conclude that @xmath502 , where the constants in the symbol @xmath489 depend only on the ( suitably defined ) lengths of the edges of @xmath473 .
it follows that there is an @xmath503 such that @xmath504 finally , using a compact fundamental domain for the action of @xmath58 on @xmath42 , one can go from @xmath50 to @xmath505 by a path which crosses at most @xmath506 edges of @xmath479 , or more precisely which is made up of @xmath507 ties and of arcs in @xmath508 .
it follows that @xmath509 combining the estimates of equations [ equn:11.1][equn:11.5 ] , we obtain @xmath510 for arbitrary @xmath511 and @xmath512 . given @xmath480 and an integer @xmath513 , the number of @xmath127 meeting @xmath480 ( or , equivalently of components @xmath491 of @xmath514 ) with @xmath515 is uniformly bounded by a number depending only on the topology of @xmath0 ; see ( * ? ? ?
* lemma 4 ) .
it follows that @xmath516 and therefore that @xmath517 now , lift the edges @xmath518 , @xmath519 ,
, @xmath520 , of @xmath473 to edges @xmath521 of @xmath479 , so that every edge @xmath480 is of the form @xmath522 for some @xmath523 and for one of these finitely many @xmath524 .
if @xmath525 , then @xmath526 , so that @xmath527 by convergence of the poincar series @xmath528 for @xmath529 , this completes the proof of the convergence of the series @xmath530 .
this also shows that @xmath531 , which completes the proof of proposition [ thm : sumcv ] .
we will need at some point a slightly stronger result : [ thm : sumcv : improv ] for @xmath18 sufficiently small , the series @xmath532 converges , and this uniformly in @xmath18 .
we again split the sum as @xmath533 if @xmath486 , it again follows from the combinatorics of transverse cocycles @xcite that @xmath534 and @xmath535 because @xmath536 and by existence of the derivative cocycle @xmath537 , we have that @xmath538 . also , @xmath539 and @xmath540 for some @xmath503 , as in equations [ equn:11.2 ] and [ equn:11.4 ]
. therefore , there are constants @xmath138 , @xmath541 such that @xmath542 for @xmath543 and @xmath544 , if @xmath18 is chosen sufficiently small that @xmath545 .
we can then conclude the proof as for proposition [ thm : sumcv ] .
the main theorem of the article , namely theorem [ thm : mainthm ] in the introduction , is an immediate consequence of theorems [ thm : compderiv ] and [ thm : derlin ] below .
[ thm : compderiv ] let @xmath198 , @xmath34 -\varepsilon , \varepsilon\right[$ ] , be a differentiable curve in @xmath9 , let @xmath105 be a maximal geodesic lamination in @xmath0 , and let @xmath124 be its preimage in the universal covering @xmath42 .
then , for every hlder continuous function @xmath200 with compact support , @xmath546 where : the sum is over all components @xmath127 of @xmath207 which are different from the component @xmath126 containing the base point @xmath242 ; @xmath547 where the boundary of @xmath127 consists of the geodesics @xmath143 , @xmath144 , @xmath145 with @xmath145 closest to @xmath50 , @xmath548 is the cosine of the angle from the geodesic @xmath134 to the geodesic @xmath226 at their intersection point , measured counterclockwise for the metric @xmath116 , and @xmath549 by convention when @xmath134 and @xmath226 do not meet ; @xmath550 where @xmath208 is the number associated by the shearing cocycle @xmath202 of @xmath551 to an arbitrary geodesic arc joining the base point @xmath50 to @xmath127 .
we are now ready to justify the formal computations of section [ sect : formalcomp ] .
we will use the definitions and notation of that section .
again , we use the isometry @xmath121 associated to @xmath116 to identify @xmath42 to @xmath81 . in this identification @xmath279 ,
the map @xmath121 is now the identity and we use the same symbol to represent an object in @xmath42 and its image under @xmath121 in @xmath81 .
note that , for two transverse cocycles @xmath150 , @xmath226 , @xmath555 where @xmath556 and @xmath557 .
since this product involves only finitely many earthquakes , it then follows from lemma [ thm : elemshear ] that , for every @xmath18 , @xmath558 as @xmath559 converges to @xmath564 , the homeomorphism @xmath561 of @xmath565 uniformly converges to @xmath566 . in particular , for a fixed @xmath127 , @xmath567 converges to @xmath568 .
similarly , for @xmath172 , @xmath569 converges to @xmath570 . to justify equation [ eqn : derivcompshear ]
, we need a uniform estimate for the convergence of the sums in the right hand sides of equations [ eqn : derivfiniteshear ] and [ eqn : derivcompshear ] .
if @xmath577 , we can arrange the indexing so that those @xmath166 with @xmath578 are exactly those with @xmath579 ( while keeping the property that @xmath554 ) .
then , @xmath580 where @xmath581 is the one among the two geodesics @xmath582 and @xmath583 that separates @xmath127 from the base point @xmath50 , and where @xmath584 is the isometry of @xmath81 that acts by translation of @xmath135 along the oriented geodesic @xmath134 .
let @xmath585 . then @xmath586 and @xmath587 .
on the other hand , it is proved in @xcite that , if @xmath588 is sufficiently small ( depending on the metric @xmath116 ) and therefore if @xmath18 is sufficiently small , @xmath589 remains bounded in the isometry group of @xmath81 ( the bound depends on @xmath159 , but not the condition on @xmath18 ) .
as @xmath559 tends to @xmath560 , the homeomorphism @xmath561 of @xmath565 uniformly converges to @xmath566 . in particular , for a fixed @xmath127 , @xmath602 converges to @xmath603 and @xmath604 converges to @xmath605 . combining equation [ eqn : estfinite earth ] and lemma [ thm : sumcv : improv ]
, we conclude by dominated convergence that @xmath606 for @xmath18 sufficiently small .
lemma [ thm : dercomshear ] then follows by differentiating the two sides of equation [ eqn : intcompshear ] .
we now let the spanning family @xmath159 uniformly tend to infinity . recall that this means that the set @xmath607 converges to the set of all components @xmath163 of @xmath142 .
we will do this in a controlled way .
recall that @xmath267 denotes the center of the ideal triangle @xmath127 . for @xmath608 ,
let @xmath609 denote the set of those components @xmath163 of @xmath610 such that @xmath611 .
note that the set of all @xmath267 consists of finitely many orbits of the action of @xmath58 on @xmath42 , one orbit for each component of @xmath106 , and is consequently locally finite .
it follows that @xmath609 is finite .
let @xmath612 consist of all elements of @xmath609 which are maximal for @xmath154 . by construction , @xmath613 is a spanning family . by proposition [ thm : estc(g ) ] , @xmath620 .
the angle under which @xmath621 is seen from the base point @xmath50 is @xmath622 . because the @xmath623 are pairwise non comparable , the angular sectors under which the corresponding @xmath621 are seen from @xmath50 are disjoint .
we conclude that @xmath624 .
therefore , @xmath625 if @xmath626 is the infimum of the distance @xmath627 over @xmath623 .
if @xmath54 is the diameter of @xmath0 for the metric @xmath116 , every point of @xmath279 is at distance at most @xmath54 from some @xmath267 .
it follows that , for every @xmath127 , there is a @xmath628 such that @xmath629 . since @xmath613 consists of the maximal elements of @xmath630 , we conclude that @xmath631 . as a consequence , @xmath632 which tends to @xmath231 as @xmath520 tends to @xmath232 . for a general @xmath18 , note that @xmath633 if @xmath623
the homeomorphism @xmath634 of @xmath635 is hlder continuous with uniformly bounded hlder exponent and norm ( depending on a compact subset of teichmller space containing the @xmath210 ) .
since the angle under which @xmath636 is seen from @xmath50 is @xmath637 , where @xmath638 is the distance from @xmath50 to @xmath636 , it follows that @xmath639 for some constant @xmath640 . in particular , the infimum @xmath641 of all @xmath638 with @xmath623 is such that @xmath642 .
proposition [ thm : estc(g ) ] gives that @xmath643 by definition of @xmath644 , we have @xmath645 decomposing @xmath613 as the union of the set @xmath646 of those @xmath647 with @xmath648 and the set @xmath649 of those @xmath650 with @xmath651 , we conclude that @xmath652 since @xmath653 , as in the case where @xmath617 .
this completes the proof of lemma [ thm : nobdryterm ] .
note that @xmath658 when @xmath156 , and @xmath659 otherwise .
in particular , in the first case , @xmath660 is the @xmath210distance from the base point @xmath50 to the @xmath210geodesic triangle in @xmath42 corresponding to @xmath127 . as a consequence , @xmath661 in both cases , and we conclude that @xmath662 for some @xmath529 depending on @xmath71 and on a compact subset of @xmath9 containing the image of the curve @xmath198 .
therefore , @xmath663 is bounded by the sum of @xmath664 and @xmath665 , where @xmath127 ranges over all components of @xmath666 that are different from @xmath126 , and where @xmath667 in the second sum .
these two sums converge by lemma [ thm : sumcv : improv ] , applied to the metric @xmath210 instead of @xmath116 for the second sum .
this proves lemma [ thm : cvpartialearth ] .
[ thm : derlin ] with the data of theorem [ thm : compderiv ] , the map @xmath670 which to a hlder continuous function @xmath200 with compact support associates @xmath671 is a hlder geodesic current @xmath672 which depends only on the tangent vector @xmath673 .
the map @xmath674 so defined is linear , and depends continuously on @xmath116 . by theorem [ thm : compderiv ] and
proposition [ thm : sumcv ] , @xmath670 defines a hlder distribution on the space @xmath48 . also , since @xmath675 is invariant under the action of @xmath58
, so is its derivative @xmath670 .
( note that it is less obvious to see the invariance directly from the formula @xmath676 ) . as a consequence ,
@xmath677 is a hlder geodesic current .
since the map which to a hyperbolic metric @xmath14 associates its shearing cocycle @xmath678 is a local diffeomorphism , @xmath679 depends linearly on the tangent vector @xmath680 .
it follows that @xmath676 depends linearly on @xmath681 . in other words ,
the map @xmath682 is linear .
finally , for two geodesics @xmath134 , @xmath683 , the cosine @xmath684 depends continuously on the metric @xmath116 used to measure the angle @xmath685 . by bounded convergence
, it follows that each @xmath254 is a continuous function of @xmath116 .
in addition , the arguments of the proof of proposition [ thm : sumcv ] show that the convergence of the series @xmath686 is uniform as @xmath116 stays in a compact subset of @xmath113 . consequently , for every @xmath21
, @xmath676 depends continuously on @xmath116 . in other words
, @xmath687 is a continuous function of @xmath116 and @xmath681 .
richard d. canary , david b. a. epstein , paul green , _ notes on notes of thurston _
, in : _ analytical and geometric aspects of hyperbolic space ( coventry / durham , 1984 ) _ , 392 , london math .
lecture note ser
. vol . * 111 * , cambridge univ . press ,
cambridge , 1987 . | for a compact riemannian manifold of negative curvature , the geodesic foliation of its unit tangent bundle is independent of the negatively curved metric , up to hlder bicontinuous homeomorphism .
however , the riemannian metric defines a natural transverse measure to this foliation , the liouville transverse measure , which does depend on the metric . for a surface @xmath0
, we show that the map which to a hyperbolic metric on @xmath0 associates its liouville transverse measure is differentiable , in an appropriate sense .
its tangent map is valued in the space of transverse hlder distributions for the geodesic foliation .
one of the very basic examples of measure preserving dynamical system is the geodesic flow of a riemannian manifold @xmath0 .
the metric @xmath1 of @xmath0 has a natural lift to a riemannian metric @xmath2 on the unit tangent bundle @xmath3 , and liouville observed that the volume form defined by @xmath2 on @xmath3 is invariant under the geodesic flow .
we want to analyze what happens in this situation as we vary the metric @xmath1 . in general , as we modify the metric @xmath1 , the topology of the geodesic flow can dramatically change . however
, if @xmath0 is compact and if @xmath1 has negative curvature , a fortunate phenomenon occurs : the _ geodesic foliation _ of @xmath3 , whose leaves are the orbits of the geodesic flow , is independent of the metric @xmath1 .
more precisely , if @xmath1 and @xmath4 are two negatively curved riemannian metrics on the compact manifold @xmath0 , and if @xmath5 and @xmath6 are the corresponding geodesic foliations of @xmath3 , there exists a hlder bicontinuous homeomorphism of @xmath3 which sends @xmath5 to @xmath6 and which is isotopic to the identity ; see for instance @xcite .
in general , this homeomorphism does not respect the parametrization of the leaves of @xmath5 and @xmath6 provided by the geodesic flows @xcite . as we move from the geodesic flow to the geodesic foliation
, there is a one - to - one correspondence between measures on @xmath3 which are invariant under the geodesic flow and transverse measures for the geodesic foliation @xmath5 .
indeed , any measure which is invariant under the geodesic flow locally is the product of some transverse measure for @xmath5 and of the measure induced by the flow on the leaves of @xmath5 . in particular , the liouville measure of
the metric @xmath1 defines a transverse measure @xmath7 for the geodesic foliation @xmath5 of @xmath1 . as indicated above
, the geodesic foliation @xmath5 does not depend on the negatively curved metric @xmath1 ; however , as a transverse measure for this fixed foliation , the liouville transverse measure @xmath7 does depend on @xmath1 . in this paper
, we restrict attention to the case where @xmath0 is a compact surface and where @xmath1 is a hyperbolic metric , namely has constant curvature @xmath8 .
let @xmath9 be the _ teichmller space _ of @xmath0 , consisting of all isotopy classes of hyperbolic metrics on @xmath0 , and let @xmath10 denote the space of all transverse measures for the geodesic foliation @xmath11 of @xmath3 .
the elements of @xmath12 are called _ measure geodesic currents _ as they are a special type of de rham currents on @xmath3 . the above
construction defines a map @xmath13 which to a hyperbolic metric @xmath14 associates its liouville current @xmath7 ; indeed , @xmath7 depends only on the isotopy class of @xmath1 .
this map @xmath15 was the focus of the article @xcite ( see also @xcite ) where , in particular , it was shown to be a proper embedding if @xmath10 is endowed with the weak * topology .
the current paper is devoted to proving that the map @xmath15 is differentiable , provided that we suitably extend its range .
the case of the map which to @xmath16 associates the dirac measure @xmath17 at the point @xmath18 on @xmath19 suggests that , when considering tangent vectors to curves in a space of measures , one should expand one s scope to distributions . indeed , in this case , the derivative @xmath20 is the distribution which to the differentiable function @xmath21 associates its derivative @xmath22 . for liouville currents ,
one encounters an additional difficulty : although a hyperbolic metric @xmath1 defines a differentiable transverse structure for the geodesic foliation @xmath5 , this transverse differentiable structure depends on the metric @xmath1 . as a consequence
, there is no intrinsic general notion of transverse distribution for @xmath5 .
however , @xmath5 admits an intrinsic transverse hlder structure , and we can consider _ transverse hlder distributions _ for @xmath5 .
such a transverse hlder distribution assigns to each surface @xmath23 transverse to @xmath5 in @xmath3 a continuous linear form on the space of compactly supported hlder continuous functions @xmath24 , and this assignment is invariant under the holonomy of @xmath11 . in particular , a transverse hlder distribution is a special type of transverse distribution , and defines a foliated de rham current of dimension 1 in @xmath3 , as in @xcite .
let @xmath25 denote the space of all transverse hlder distributions for the geodesic foliation @xmath5 , also called _ hlder geodesic currents_. when endowed with the weak * topology , @xmath25 is an infinite dimensional topological vector space , which contains the space @xmath12 of measure geodesic currents .
recall that the teichmller space @xmath9 is a differentiable manifold of dimension @xmath26 , where @xmath27 is the euler characteristic of @xmath0 .
[ thm : mainthm ] the liouville map @xmath28 is differentiable at each @xmath29 , in the following sense : there exists a linear map @xmath30 such that @xmath31 for every tangent vector @xmath32 and for every differentiable curve @xmath33 , @xmath34-\varepsilon , + \varepsilon\right[$ ] , with @xmath35 and @xmath36 .
in addition , the tangent map @xmath37 depends continuously on @xmath1 . by definition of the weak * topology on @xmath25 ,
the statement @xmath31 means that , for every hlder continuous function with compact support @xmath38 defined on a surface @xmath23 transverse to @xmath5 in @xmath3 , the number @xmath39 associated to @xmath21 by the hlder transverse distribution @xmath40 is equal to @xmath41 in particular this derivative exists .
hlder geodesic currents already appeared in the paper @xcite , where they occurred as tangent vectors to the space of measured geodesic laminations on the surface @xmath0 ; see in particular @xcite for a few applications of this analytic point of view .
it seems remarkable that a notion which was originally introduced because of consistency considerations , namely because it was the only type of transverse distributions to @xmath5 which seemed to make sense , could actually end up being the correct background to compute the tangent maps of various geometric functions .
we prove theorem [ thm : mainthm ] by an explicit computation using the shearing coordinates for teichmller space developed in @xcite ( dual to the length coordinates appearing in @xcite ) .
in particular , we give in theorem [ thm : compderiv ] an explicit formula for the tangent map @xmath37 in terms of these shearing coordinates .
the main part of the argument is to prove the convergence of the series involved in this formula .
presumably , one could alternatively use more classical coordinates for @xmath9 , such as the fenchel - nielsen coordinates .
however , this seems harder to do and the shearing coordinates have the definite technical advantage that one only has to deal with shearing as opposed to mixing lengths and shearing .
dragomir ari has recently developed a new approach @xcite to the results of this article .
it is somewhat more direct , and has the additional advantage of extending to ( infinite dimensional ) teichmller spaces of surfaces with infinite area , and in particular to the universal teichmller space .
to some extent , the occurrence of hlder geodesic currents in this infinite dimensional setting is somewhat surprising , as one could have expected distributions with a lower level of regularity .
we are grateful to the referee for a critical reading of the manuscript and for many insightful suggestions . |
the measured flake was obtained from the mechanical exfoliation of natural graphite .
it was then transferred non - deterministically on top of a undoped sio@xmath40(90nm)/si substrate on top of which an array of holes ( @xmath41 m diameter ) was etched by means of optical lithography techniques .
room temperature raman characterization is performed on the flake , using a helium - neon laser source of @xmath42 nm ( i.e. , @xmath43 ev ) . to probe the landau levels dispersion in our abc n - lg flake
, we used an experimental set - up comprising a homemade micro - raman probe that operates at low temperatures and high magnetic fields .
the excitation source was provided by a solid state titanium doped sapphire laser . the excitation wavelength fixed at @xmath44 nm ( i.e. , @xmath45 ev )
is brought to the sample by a @xmath46 m core mono - mode optical fiber .
the end of the raman probe hosts a miniaturized optical table comprising a set of filters and lenses in order to clean and focalize the laser spot .
a high numerical aperture lens is used to focalize the laser light on the sample , which is mounted on x@xmath47y@xmath47z piezo stages , allowing us to move the sample relative to the laser spot with sub-@xmath48 m accuracy .
the non - polarized backscattered light is then injected into a 50 @xmath48 m multi - mode optical fiber coupled to a mono - chromator equipped with a liquid nitrogen cooled charge coupled device ( ccd ) array .
the excitation laser power was set to @xmath49 1 mw and focused onto a @xmath50 m diameter spot .
the resulting intensity is sufficiently low to avoid significant laser induced heating and subsequent spectral shifts of the raman features .
the probe is then placed in a resistive magnetic equipped with a liquid he cryostat at 4 k. the evolution of the raman spectrum with magnetic field was then measured on the freestanding parts by sweeping the values of the magnetic field .
each spectra was recorded for a @xmath51 t in order to avoid any significant broadening of the magnetic dependent features line - widths .
the details of the theoretical description of the lls and the magneto - raman spectrum are given in the supplementary information . | graphene layers are known to stack in two stable configurations , namely aba or abc stacking , with drastically distinct electronic properties . unlike the aba stacking , little has been done to experimentally investigate the electronic properties of abc graphene multilayers . here , we report the first magneto optical study of a large abc domain in a graphene multilayers flake , with abc sequences exceeding 17 graphene sheets . the abc - stacked multilayers can be fingerprinted with a characteristic electronic raman scattering response , which persists even at room temperatures .
tracing the magnetic field evolution of the inter landau level excitations from this domain gives strong evidence to the existence of a dispersionless electronic band near the fermi level , characteristic of such stacking .
our findings present a simple yet powerful approach to probe abc stacking in graphene multilayer flakes , where this highly degenerated band appears as an appealing candidate to host strongly correlated states .
tailoring the electronic and optical properties of layered materials by controlling the layer orientation or their stacking order is an important possibility offered by the physics of two dimensional systems .
graphene is the first isolated two dimensional crystal @xcite and the properties of graphene stacks , from mono- to multi - layers ( n - lg ) , have been intensively investigated in the last decade @xcite .
the thermodynamically stable stacking of multilayer graphene is the bernal stacking , where the a sublattice in one layer comes right below the b sublattice in the other layer @xcite .
it was not until recently that experiments on rhombohedral stacking , with an abc layer sequence , have been reported .
abc tri - layer graphene , the simplest rhombohedral n - lg , have been successfully isolated and presents a tunable band gap @xcite as well as chiral quasi - particles as evidenced from their unconventional quantum hall effect @xcite . to our knowledge , there are yet no investigations of abc graphene multilayers .
this is mainly because of the low abundance of this polytype in real samples .
hence , tracing the evolution of the electronic properties of abc - stacked multilayers when increasing the number of layers remains challenging .
one of the intriguing predictions about electronic properties of abc - stacked multilayers is that they are expected to host surface states ( localized mainly on the top and bottom layers ) with a flat dispersion near the corners of the brillouin zone , and bulk states with a band gap .
when increasing the number of abc - stacked layers , the extend of the flat dispersion of the low energy bands increases , while the size of the bulk band gap decreases @xcite .
notably , the surface states of abc - stacked graphene multilayers have been reported as topologically protected , due to the symmetry of this material @xcite , and may thus resemble those well - known surface states of 3d topological insulators @xcite .
interestingly , due to weak spin - orbit interaction in carbon - based systems , an even closer analogy is found with the surface states of topological crystalline insulators @xcite .
nevertheless , in contrast to these two families of topological insulators , the bulk band gap of abc - stacked graphene multilayers gradually closes with the increasing number of atomic sheets and bulk rhombohedral graphite may behave as a 3d topologically protected semimetal @xcite .
the expected high degeneracy of the flat bands could lead to exotic electronic ground states , such as magnetically ordered phases or surface superconductivity @xcite .
experimental signatures of this flat bands has been recently observed in scanning tunnelling spectroscopy ( sts ) and angular - resolved photo emission spectroscopy ( arpes ) investigations of nanometer scale domains of abc - stacked graphene multilayers ( 5 layers ) grown on 3c - sic @xcite . in this letter , we demonstrate , using raman scattering techniques , that abc - stacked multilayers ( @xmath0 ) can be found within exfoliated n - lg flakes .
they have characteristic signatures in their raman scattering response , which allow for their identification at room temperature .
these signatures include a change of the 2d band feature line shape @xcite , together with an additional raman scattering feature which we attribute to electronic raman scattering ( ers ) across the band gap in the bulk .
abc - stacked domains can thus be identified by spatially mapping the raman scattering response of different flakes .
experiments performed with an applied magnetic field ( @xmath1 ) reveal two distinct series of electronic inter landau level excitations , involving the low energy surface flat bands , and the bulk gapped states , respectively .
experimental results are explained in the frame of a tight binding ( tb ) model of the electronic band structure of rhombohedral n - layer thin graphite layers @xcite .
the 7 parameters tb model developed by slonczewski - weiss and mcclure @xcite for bulk graphite can be simplified by only considering the two first intra- and inter- nearest neighbors hopping parameters @xmath2 and @xmath3 , respectively ( figure [ fig : bs]f - i ) . the calculated low energy band structure of abc n - lg is shown in figure [ fig : bs]a - c , for @xmath4 and @xmath5 layers .
two characteristic evolutions when increasing the number of abc - stacked layers arise from these calculations : ( _ i _ ) the flat part of the low energy @xmath6 bands ( red bands in figure [ fig : bs]a - c ) extends over a larger @xmath7-space region , and ( _ ii _ ) the energy separation @xmath8 between the @xmath9 ( green bands in figure [ fig : bs]a - c ) , decreases and ultimately closes for rhombohedral graphite ( see figure [ fig : bs]d ) .
a natural way of exploring electronic band structures of solids is to apply a magnetic field in order to induce landau quantization , and to trace the evolution of inter landau level excitations with a magneto - spectroscopy techniques , such as magneto - raman scattering spectroscopy .
the evolution with magnetic field of the four lowest in energy bands of a @xmath10 abc stacked sequence is presented in figure [ fig : bs]e .
one can note that landau levels are formed from the flat bands ( red lines ) , and that their energies grow nearly linearly with the magnetic field , but starting from a finite onset magnetic field . for @xmath10 ,
the onset field is close to @xmath11 t , as indicated by blue bar in figure [ fig : bs]e . as a consequence , inter - landau level excitations within the flat bands ( red arrows ) , in the first approximation ,
evolve linearly with the applied magnetic field , however with a rather unusual extrapolated negative energy offset ( see supplementary information ) .
we anticipate that this will be the magneto - raman scattering signature of electronic states with such dispersion .
landau levels in the @xmath12 and @xmath13 bands ( green lines in figure [ fig : bs]e ) also show a non trivial evolution when increasing the magnetic field . from @xmath14
ev at the k - point , their energies first decrease ( increase ) with the magnetic field until they reach @xmath15 , respectively . for higher magnetic fields ,
their energies increase ( decrease ) with a quasi linear evolution .
thus , inter landau level excitations involving these states ( green arrows ) are expected to first decrease in energy down the gap value , and to grow in a linear way for higher magnetic fields .
an optical photograph of the investigated flake is presented in figure [ fig : zero_b]a .
the flake is produced from the mechanical exfoliation of bulk graphite and transferred on a sio@xmath16(90nm)/si substrate patterned with holes in the oxide ( see methods ) .
the raman scattering signal from the suspended parts is enhanced due to an optical interference effect in the substrate @xcite .
the n - lg flake covers five different holes labelled h0 to h4 , where it is suspended .
the flake has been characterized by atomic force microscopy ( afm ) measurements which indicate a thickness of 15 to 17 layers ( see supplementary information ) .
raman scattering spectra are then measured at room temperature with 50x objective and a @xmath17 nm laser excitation , or at liquid helium temperature in a solenoid using a homemade micro - magneto - raman scattering ( mmrs ) spectroscopy set - up with a @xmath18 nm laser excitation ( see methods ) . in both experiments ,
piezo stages are used to move the sample with respect to the laser spot and to spatially map the raman scattering response of the flake , or to investigate specific locations .
two characteristic raman scattering spectra measured at h0 ( black curve ) and at h4 ( orange curve ) are presented in figure [ fig : zero_b]b .
one can recognize on these spectra the characteristic phonon response of sp@xmath19 carbon , including the g band feature around @xmath20 @xmath21 , and the 2d band feature observed around @xmath22 @xmath21 when measured with @xmath23 nm excitation .
if the spectrum measured at h0 corresponds to that of thin ab - stacked layers @xcite , the spectrum measured at h4 is notably different , even though the thicknesses of the layer at these two locations are similar : i ) the g band energy at h4 is blue shifted by @xmath24 @xmath21 with respect to that measured at h0 ( figure [ fig : zero_b]c , ii ) the 2d band line shape is more complex with two additional contributions indicated by arrows in [ fig : zero_b]d , and iii ) an additional broad feature ( figure [ fig : zero_b]e is observed at h4 around 1805 @xmath21 with a full width at half maximum ( fwhm ) of @xmath25 @xmath21 .
these are the three main raman scattering signatures of abc - stacked graphene multilayers . to study the homogeneity of the n - lg flake
, we have performed spatial mappings of the raman scattering response at room temperature , with @xmath26 m spatial steps .
when scanning the surface of the flake , it appears that the broad raman scattering feature is observed in a large domain and that its central position can change from one location to another ( see figure [ fig : zero_b]f ) .
as will be clarified in the following by the magneto - raman measurements , we interpret this feature as arising from an electronic excitation between two @xmath9 bands of the band structure of abc n - lg .
similar to metallic carbon nanotubes @xcite and bulk graphite @xcite , electronic excitations do contribute to the room temperature raman scattering response of abc - stacked n - lg .
the different locations on the flake where this ers signal is observed are presented in the form of a false color spatial map in figure [ fig : zero_b]h .
the modified line shape of the 2d band is also observed at different locations on our flake , shown in figure [ fig : zero_b]g .
the correlation between these two false color maps is a strong indication that the observation of the ers feature and of the modified 2d band line shape are signatures of the same , abc stacking .
this flake is hence composed of two distinct domains : the first is aba - stacked , and is extending over h0 , h1 and h2 , while the other is abc - stacked , extending over h3 and h4 .
the energy of this ers ( i.e. , the bulk band gap in figure [ fig : bs]a - c depends on the number of abc - stacked layers , and , taking the extreme values of the ers energy from the measured raman spectra all over the abc domain , we estimate the number of abc - stacked graphene layers to vary from @xmath27 to @xmath28 over the whole flake .
this feature is also observed at low temperature , as can be seen in figure [ fig : b_spec]a .
characteristic mmrs spectra measured at h4 are presented in figure [ fig : b_spec]b , for different values of b. the central position of the broad ers feature increases with increasing magnetic field and its intensity seems to vanish for @xmath29 t. the fact that magnetic fields can influence so much the energy and amplitude of this feature is in line with an electronic origin for this excitation . above @xmath30
t a series of sharp features , dispersing with the magnetic field , appears in the raman scattering response .
they have a symmetric line shape , in contrast to electronic excitations observed in bulk graphite @xcite . the evolution of the raman scattering response with magnetic field , measured at h3 and h4 ,
are presented in figures [ fig : b_spec]c - d , respectively , in the form of false color maps of the scattered intensity .
these two locations show a similar response .
the observed electronic excitation spectrum in magnetic field is composed of a series of features linearly dispersing with increasing magnetic field and of the ers observed at @xmath31 , that acquires a fine structure at finite @xmath1 .
the linearly dispersing features , in line with magneto - raman scattering selection rules in graphene @xcite or in graphite @xcite , can be attributed to symmetric ( @xmath32 , where n is the landau level index ) inter landau level excitations within the @xmath33 bands and they represent the strongest contribution to the electronic raman scattering spectrum in magnetic field .
optical - like excitations ( @xmath34 ) are not directly seen but they effectively couple to the g band .
they give rise to the magneto - phonon resonance and to the associated anti - crossings when they are tuned in resonance with the g band energy @xcite .
such anti - crossings are indicated by red arrows in the b - differentiated color map of h4 , presented in figure [ fig : b_spec]f .
the detailed analysis of the magneto - phonon resonance in abc n - lg is beyond the scope of this paper . of much weaker intensity , raman scattering features with an energy decreasing when increasing the magnetic field
are observed ( red marked regions in figure [ fig : b_spec]d - e .
these features are better seen in figure [ fig : b_spec]e , that shows the @xmath31 subtracted false color map of h3 .
they correspond to symmetric ( @xmath32 ) inter landau level excitations within the lowest @xmath35 bands in the bulk .
the energy of such excitations first decreases with the magnetic field as long as the landau levels are confined in the cone of the corresponding band structure at @xmath31 ( figure [ fig : bs]c ) and then , when the inter landau level energy spacing reaches the value of the energy gap in the bulk , they merge with the broad ers feature .
the evolution of these two families of electronic excitations with increasing magnetic field is grasped by our tight - binding analysis .
we first determined the central positions of the different observed excitations using lorentzian functions , and searched for the parameters entering the model to best describe our data . from this modelling
, we can determine the number of layers to be @xmath36 and @xmath5 for h3 and h4 , respectively ( figures [ fig : fits]a - b ) , while we set @xmath37 ev and @xmath38 ev , as observed in bulk graphite @xcite . in a simple approach ( see supplementary information ) , we calculated the electronic excitation spectrum @xcite and its evolution when applying a magnetic field .
these results are compared to the experimental evolution in figures [ fig : fits]c - d for h4 .
the model reproduces the main observed features .
in particular , the electronic excitation observed at @xmath31 is reproduced and arises indeed from an inter band electronic excitation that loses its spectral weight when the magnetic field is increased , transforming into inter landau level excitations with their characteristic negative energy dispersion .
we report the observation of electronic excitations in n - lg system , which includes a large domain of @xmath39 abc - stacked layers . the analysis of its low energy electronic excitations with magnetic field can be understood in the frame of a tight - binding model with three parameters , the number of abc - stacked layers and the intra- and inter - layer nearest neighbors hopping integrals @xmath2 and @xmath3 .
such stacking has a unique signature in its raman scattering response at @xmath31 , in the form of a low energy electronic excitations across the band gap in the bulk .
the ers response is also observed at room temperature .
the central position of the ers is related to the number of abc - stacked layers which can then be deduced using simple raman scattering spectroscopy .
our findings underscore the rich physics hidden in graphene multilayers with abc stacking , namely the existence of electronic bands with a flat dispersion ( diverging density of states ) localized on the surface , and of electronic states in the bulk with an energy gap that depends on the number of layers .
these results represent an impetus for other studies targeting the highly correlated surface states , which may lead to emergent exotic electronic ground states on this system , such as magnetic order or superconductivity .
this work has been supported by the erc advanced grant momb ( contract no .
320590 ) and the ec graphene flagship ( project
no .
604391 ) .
h.p.o.c .
, g.u . and c.a.b .
acknowledge financial support from picts 2013 - 1045 and bicentenario 2010 - 1060 from anpcyt , pip 11220110100832 from conicet and grant 06/c415 from secyt - unc . |
the subaru deep survey is a systematic project of the 8.2 m subaru telescope to study the deep extragalactic universe .
the subaru deep field was selected near the north galactic pole , avoiding large galactic extinction and nearby galaxy clusters , and the airmass of this field is smaller than the hubble deep field at mauna kea ( maihara et al .
the wide field near - infrared ( nir ) camera cisco took a very deep nir @xmath9 iamge in @xmath10 and @xmath11 bands , with 5@xmath3 magnitude limits of 25.1 and 23.5 .
this is the deepest image in the @xmath0 band taken so far , providing a unique @xmath0-selected sample of galaxies which should be useful for study of faint , high-@xmath1 galaxies .
the field was also deeply followed - up by optical instruments of focas and suprime - cam . here
we review some interesting implications obtained by these data set , focusing on nir galaxy counts , colors , and photometric - redshift distribution , compared with some theoretical models of galaxy formation and evolution .
although omitted here , some interesting results for the clustering of lyman break galaxies and lyman alpha emitters at @xmath12 have been obtained in the sdf and another project of the subaru / xmm - newton deep survey , thanks to the very wide field of the suprime - cam .
see ouchi et al .
( 2001 , 2002 ) for these .
figure 1 shows @xmath0 band sdf galaxy counts , compared with those estimated by other observations ( totani et al .
2001a ) . here , we plot counts multiplied by flux , rather than count itself , to show the contribution to the cosmic background radiation ( cbr ) per magnitude .
both the raw and corrected counts assuming point sources are showing very flat faint - end slope , with rapidly decreasing contribution to cbr beyond @xmath13 .
therefore the extrapolation of the galaxy counts into fainter magnitudes does not significantly increase ebl but converges to a finite ebl flux , and this means that the bulk of ebl from galactic light has already been resolved into discrete galaxies .
these results require that the diffuse ebl in nir bands should not be different from the count integrations , provided that the ordinary galactic light is the dominant source of the ebl in these bands , as generally believed . however , a few recently reported detections of diffuse ebl in these bands suggest that the diffuse cbr flux in @xmath0 bands is consistently higher than the count integrations by a factor of @xmath14 : @xmath15 ( cambrsy et al . 2001 ) , @xmath16 ( matsumoto 2000 ) , and @xmath17 ( wright 2001 ) , which should be compared with the integration of @xmath0 counts ( @xmath18 )
. if the discrepancy between the diffuse ebl and count integration is real , it might suggest the existence of very diffuse component which is different from normal galaxies . before deriving this extraordinary conclusion ,
however , all possible systematic uncertainties in the above estimates must extensively be checked .
one of such systematics is the contribution to ebl by the galaxies missed in deep galaxy surveys .
since galaxies are extended sources , the detectability near the detection limit is not as simple as point sources .
furthermore , the well - known effect of the cosmological dimming of surface brightness [ @xmath19 should make high-@xmath1 galaxies very difficult to detect , while such objects may have a significant contribution to ebl .
the photometry scheme could also be a problem , because there is considerable uncertainty in the estimate of the magnitude of faint galaxies because of ` growing ' the photometry beyond the outer detection isophotes of galaxies .
therefore we estimated the contribution to cbr by galaxies missed in sdf , based on a realistic theoretical model of galaxy counts which includes all known physical or observational selection effects and incompleteness , based on the method presented in yoshii ( 1993 ) and totani & yoshii ( 2000 ) .
first , we construct a model of galaxy counts which best fits to the observed raw ( i.e. , uncorrected ) counts , taking into account all the above selection effects .
then we can calculate the true galaxy counts and ebl flux using the same model without selection effects , and comparison between the true counts and observed counts gives an estimate of contribution by missing galaxies .
the results are shown in fig .
1 , where the dashed line is the best - fit model to raw sdf counts taking into account the selection effects in theoretical calculation , while the solid line is the same model but without the selection effects . by this procedure
, we found that the correction by the incompleteness would increase the cbr from galaxies by at most 10% , which is too small to reconcile the count integrations and diffuse cbr measurements ( totani et al .
therefore we conclude that there must be a new source of cbr in the nir band , which must be very different and distinct populations from known galaxies , unless some unkown systematics have affected the diffuse cbr measurements significantly .
the model shown in the dashed and solid lines is a so - called pure - luminosity - evolution ( ple ) model without number evolution , including five types of galaxies ( e / s0 , sab , sbc , scd , and sdm ) , whose detail is given in totani & yoshii ( 2000 ) .
in fact , all the sdf nir data of counts , @xmath20 colors , and size distributions are well described by this rather simple model in the popular @xmath21-dominated flat universe , without any indication of number evolution .
this is somewhat in contrast to the result of the same model against hdf galaxies , where a modest number evolution of @xmath22 is required to fit the data [ @xmath23 , @xmath24 ( totani & yoshii 2000 ) .
this is the most naturally explained by different merging histories for different galaxy types ; longer wavelengths are dominated by earlier types .
therefore , these results indicate that elliptical galaxies which are dominant in the @xmath0 band are evolving without significant change of number density from @xmath25 to the present ( totani et al .
a number evolution of @xmath22 for elliptical galaxies is already inconsistent with the data , as shown by dotted lines in fig
. 1 . on the other hand ,
a study of sdf k counts by a hierarchical merging ( hm ) model based on cdm - based structure formation is presented in nagashima et al .
( 2002 ) , where the selection effects are carefuly taken into account in a similar way .
the hm also fits to the sdf counts in low density cosmological models , but the ple and hm models are giving different predictions in the redshift distribution , which will be discussed later .
an interesting discovery by sdf is the existence of unusually red objects in nir colors , with @xmath2634 .
the brightest four objects of them are presented in maihara et al .
( 2000 ) , and a more careful analysis has been performed to estimate the number fraction of such objects as a function of @xmath0 magnitude ( totani et al .
they found that the number fraction of such objects sharply rises with increasing magnitude from @xmath27 , reaching a few percent at the faintest magnitudes ( see fig .
it should be noted that such nir color is even redder than the known population of the extremely red objects ( eros ) , which are difined by red optical - nir colors such as @xmath28 ; typical eros have @xmath29 of at most 2 .
furthermore , such red nir color can not be explained by passively evolving elliptical galaxies at any reasonable redshift without extinction by dust ( totani et al .
2001b ) , while recent studies of eros indicate that the majority of them are passively evolving ellipticals ( e.g. , daddi et al .
2000 ) . then
two interpretations remain for these hyper extremely red objects ( heros ) .
one is ultra - high @xmath1 objects at @xmath30 , the red color being due to the lyman - break between @xmath10 and @xmath0 bands . since heros are too faint to measure redshifts , it is difficult to verify observationally .
however , theoretically it is very unlikely ; the rest - frame uv luminosity indicates star formation rate of more than 100@xmath31/yr , but there should be very few objects which are massive enough to allow such high sfr at @xmath30 , according to the widely believed cdm - based structure formation theory .
the other interpretation is that they are dusty starbursts , which often show very red colors .
in fact , totani et al . (
2001b ) has shown that the colors and counts of heros are well reproduced by a simple model if present - day elliptical galaxies have formed by starbursts with a reasonable amount of dust ( i.e. , inferred from model metallicity ) at @xmath5 .
this redshift is similar to those estimated for the faintest submillimeter sources in recent years .
interestingly , the counts of heros are roughly the same with the faintest scuba sources , and submm flux expected by the dusty - starburst model of heros is in fact close to the scuba sensitivity limit . a detailed modeling of fir - submm counts by totani & takeuchi ( 2002 ) has indeed shown that scuba counts are nicely explained by the dusty starbursts of forming elliptical galaxies which are also responsible for heros .
these results suggest an interesting possibility that heros and scuba sources are the same population , which will evolve into present - day elliptical galaxies or bulges .
it is very important to examine the correlation between these two .
ultimate confirmation of this hypothesis would be brought in the ngst and alma era .
[ fig : hero ] both the ple and hm models fit to the sdf @xmath0 galaxy counts
. it may be because the galaxy counts do not have enough power to discriminate these two , or may be because galaxies dominant in the @xmath0 band in the hm model have only small or negligible number evolution .
this degeneracy can be broken by redshift distribution ; especially a @xmath0-selected sample has a strong power for the discrimination , since @xmath0-band light traces the stellar mass which has formed so far , rather than star formation rate at that time .
though spectroscopic redshifts are unavailable for the faintest sdf galaxies , a reasonably reliable test is possible by the photometric redshift technique , as is done by kashikawa et al .
the result is shown in fig .
3 , in a form of cumulative @xmath1-distribution separated by @xmath11 magnitude intervals . in bright
magnitude ranges of @xmath32 , the observed distribution is somewhat between the two models of ple and hm .
although in the range of @xmath33 the hm model looks better than the ple , statistical fluctuation is large because of the small number of galaxies .
the kolmogorov - smirnov ( ks ) test gives a chance probability of 2 and 1% of getting the observed distribution from the ple model .
considering model uncertainties which is not taken into accoun in the ks test , it is impossible to reject this model . on the other hand ,
the deficit of high-@xmath1 galaxies in the hm model at @xmath34 is statistically much more significant ( chance probabilities of @xmath35 and @xmath36 at @xmath37 and @xmath38 , respectively ) .
it should also be noted that the hm model used here gives @xmath1-distribution more weighted to high redshifts compared with another hm model of kauffman & charlot ( 1998 ) ( see the top panel ) .
we emphasize that this is a `` blind '' test , i.e. , the models shown here are those fitting best only to the sdf counts as described in totani et al .
( 2001c ) and nagashima et al .
( 2002 ) without further tuning of parameters to the new photo-@xmath1 data of kashikawa et al .
( 2002 ) .
these results seem to be controversial when compared with several previous papers doing a similar test ( but in shallower magnitudes ) .
fontana et al .
( 1999 ) and rudnick et al . (
2001 ) claimed that the photometric redshift distribution at @xmath39 is consistent with hm models , but not with ple models .
firth et al .
( 2002 ) reached a similar conclusion using a sample of @xmath40 .
it should be noted that ple models used by these groups are either dust - free or assuming only constant extinction at the level of galactic extinction .
such ple models are inconsistent with the observed @xmath1-distribution since they predict too many high-@xmath1 galaxies which are visible because of strong starbursts assumed in the formation of elliptical galaxies ( see top panel of fig . 3 for the dust - free ple prediction by kauffmann & charlot 1998 ) . however , dust - free model is obviously unrealistic , especially for initial starbursts expected for elliptical galaxies .
we know that starbursting populations of galaxies quite often show strong extinction and reddening .
a chemical evolution model of elliptical galaxies also suggests that the amount of metal produced in the initial starburst phase is huge , and hence strong extinction seems quite plausible ( totani & yoshii 2000 ) .
the ple model used here ( dashed line in fig .
3 ) is taking into account the extinction by a reasonable amount of dust inferred from chemical evolution .
in addition , observational selection effects discussed in 2 may also have affected previous results .
( the theoretical predictions by our ple and hm models appropriately included all known selection effects under the sdf condition ) .
cimatti et al .
( 2002 ) compared a spectroscopic redshift distribution at @xmath41 obtained by the k20 survey with the latest ple and hm models .
in fact , they found that , if dust extinction is taken into account ( another option is using the scalo imf rather than the salpeter ) , the difference between ple and hm models becomes much smaller than previously claimed .
they found that such ple models are in reasonable agreement with the data , but on the other hand , hm models predict too many low redshift galaxies .
from these results , we conclude that the deepest @xmath0-selected sample of the sdf is in overall agreement with the simple picture of ple for early type or elliptical galaxies , but the present version of hm models has a problem in the redshift distribution of the faintest galaxies . considering the overall success of the cdm structure formation theory against various tests @xmath42 based on galaxy luminosity and star formation activities ( e.g. , clustering properties or abundance of galaxy clusters ) , it is reasonable to think that the problem identified for hm models is related with the treatment of star formation activity .
it must incorporate a population of massive and dusty starbursts at high redshift ( @xmath43 , which presumably evolved into present - day elliptical galaxies or bulges without significant number evolution .
the author would like to thank all the collaborators of the sdf project , and the subaru telescope staffs who made this project possible . he has been financially supported in part by the jsps fellowship for research abroad .
matsumoto , t. 2000 , in the proceedings of the extragalactic infrared background and its cosmological implications , international astronomical union .
symposium no .
manchester , england , august 2000 . | the subaru deep field provides the currently deepest @xmath0-selected sample of high-@xmath1 galaxies ( @xmath2 at 5@xmath3 ) .
the sdf counts , colors , and size distributions in the near - infrared bands are carefully compared with pure - luminosity - evolution ( ple ) as well as cdm - based hierarchical merging ( hm ) models .
the very flat faint - end slope of the sdf @xmath0 count indicates that the bulk ( more than 90% ) of cosmic background radiation ( cbr ) in this band is resolved , even if we take into account every known source of incompleteness .
the integrated flux from the counts is only about a third of reported flux of the diffuse cbr in the same band , suggesting that a new distinct source of this missing light may be required .
we discovered unusually red objects with colors of @xmath4 34 , which are even redder than the known population of eros , and difficult to explain by passively evolving elliptical galaxies .
a plausible interpretation , which is the only viable one among those we examined , is that these are dusty starbursts at high-@xmath1 ( @xmath5 ) , whose number density is comparable with that of present - day ellipticals or spheroidal galaxies , as well as with that of faint submillimeter sources .
the photometric redshift distribution obtained by @xmath6 photometries is also compared with the data , and the hm model is found to predict too few high-@xmath1 objects at @xmath7 and @xmath8 ; the ple model with reasonable amount of absorption by dust looks more consistent with the data .
this result is apparently in contradiciton with some previous ones for shallower observations , and we discuss the origin of this .
these results raise a question for the hm models : how to form massive objects with starbursts at such high redshifts , which presumably evolve into present - day elliptical galaxies or bulges ? # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
this is a lattice talk , but at first it wo nt seem like it ; it will be about atomic systems , and effective field theories . but studying the effective description of the atomic system will lead to a question suited to lattice techniques .
i am summarizing 3 papers @xcite , all with peter arnold ; the middle one contains all the lattice stuff .
many systems exhibit bose condensation ; for instance , liquid helium , cooper pairs in superconductors , and arguably nucleon cooper pairs in large nuclei .
these examples condense because of interactions ; none exhibit bose condensation as originally proposed by bose and einstein , where condensation occurs exclusively because the temperature is so low that the only place to store the bosons is in the ground state .
the interest in atomic bose condensation mainly arises because this is what _ does _ happen here ; mutual interactions are not that important , the dominant reason for condensation is that there is not enough heat to store the density of atoms present in the excited states .
in fact , interactions _ must _ be weak in an atomic bose condensate ; otherwise it would condense into a solid .
the beauty of small interactions is it means we can calculate ; the coupling is weak enough that perturbatively based `` field theorist '' techniques , like effective field theories and loopwise expansions , can be used .
the question i will address in this talk is , how do interactions modify the relation between density and the temperature of the bose condensation transition ? for simplicity , and to follow the literature , i consider a homogeneous system .
( this also makes the question well posed . )
it means that the problem addressed is a little artificial and has less contact with experiment than we would like ; but nevertheless since there is already a surprisingly big literature on the problem as posed , it is worth addressing .
the right effective description of a trapped atomic gas of bosonic atoms , all in the same spin state , is a second quantized schrdinger equation : @xmath0 the atomic number density is @xmath1 , with a subtraction implied so the density vanishes in vacuum .
call this theory i. technically , theory i is already an effective theory , containing no information about atomic structure ; it is only valid on length scales larger than the atomic size . since we
are concerned with a dilute system we will be interested in scales large compared with the scattering length , in which case the nonlocal interaction term should be replaced with a local term plus higher dimension corrections .
furthermore we are interested in the behavior at finite temperature , and will only ask about static properties ( namely the number density ) , so we can go to the matsubara formalism , with euclidean periodic time .
i work in the grand canonical ensemble , so i introduce a chemical potential @xmath2 , and because the system is homogeneous , @xmath3 can be dropped .
i will also set @xmath4 from now on .
the appropriate effective theory is therefore @xmath5 plus high dimension operators .
the number density is still @xmath1 , with its zero temperature value subtracted .
this effective theory requires regularization , and i use @xmath6 dimensional regularization .
@xmath7 is the scattering length ; we could derive it by matching to theory i , but we rather define it to be the coefficient shown
. there are also high dimension operators , containing either more fields or derivatives ; but at the order of interest they can all be dropped .
i will call this theory ii .
theory ii contains three interesting length scales @xmath8 : 1 .
@xmath9 , where it breaks down ; 2 .
@xmath10@xmath11 , the thermal scale , where nonzero matsubara frequencies decouple and the behavior becomes effectively 3d ; 3 .
@xmath12 , the scale were the behavior becomes strongly coupled .
the physics of bose condensation is associated with the third scale , where the nonzero matsubara modes have decoupled and the physics is effectively 3d .
therefore we can write a still simpler effective theory which contains the physics of bose condensation , which i call theory iii : @xmath13 with @xmath14 a two component real field ( the two components correspond to the real and imaginary parts of @xmath15 ) .
again there are in principle high dimension terms , but we wo nt need them .
the possibility of such matching was realized by baym et .
al . @xcite .
the relation between @xmath2 , @xmath7 , and @xmath16 in theory ii and @xmath17 , @xmath18 , and @xmath19 in theory iii is determined by matching between the theories .
the coupling is weak at @xmath20 , so the matching may be done perturbatively . at tree level ,
@xmath21 , @xmath22 , and @xmath23 .
the operator @xmath24 mixes with the identity .
a one loop matching calculation of the mixing gives the free theory value for the particle number , @xmath25 to find the @xmath26 corrections to the number density , we need to find this mixing to two loops , @xmath17 to one loop , and the other quantities at tree level . we can do even better , and determine the @xmath27 corrections , by going to one higher loopwise order in every variable .
this is done in @xcite .
we can not go further without including high dimension operators , so we will stop at this order .
besides the matching calculation , we also need the critical values of two quantities in theory iii . at linear order in @xmath7
we need the value of @xmath19 , evaluated at the critical value @xmath28 ( the critical @xmath2 ) . at second order in @xmath7
, we also need to know the critical value @xmath28 .
@xmath28 is also needed to find the total number of atoms in a wide but finite trap @xcite .
the values of @xmath19 and @xmath28 are nonperturbative .
theory iii is the 3-d o(2 ) model , and its infrared behavior falls in the x - y universality class .
the infrared behavior is strongly coupled .
furthermore , the quantities of interest , @xmath19 and @xmath28 , are not universal .
they can not , to our knowledge , be reliably determined using known analytic techniques ( perturbation theory , @xmath29 expansion , high temperature expansion , etc . ) .
we turn instead to a lattice determination .
we will refer to the lattice regularization of theory iii as theory iv .
@xmath14 , a two component real field , is discretized on a 3d cubic lattice of spacing @xmath30 , with action @xmath31 relating lattice parameters and theory iii parameters , including @xmath32 , requires another matching calculation , which is identical in structure to the matching performed between theories ii and iii ; it can be done by lattice perturbation theory . in @xcite
we perform this matching to the same , and in some cases higher , loop order , as we used in matching theories ii and iii .
the matching is an expansion in @xmath33 , which means we want @xmath30 to be relatively small .
however we also want to be deep in the critical scaling regime , requiring a large physical volume . to improve the matching to theory iii we use an improved lattice laplacian , containing anti - ferromagnetic next - nearest couplings ( analogous to using the symanzik rather than wilson action in pure glue qcd ) .
this frustrates the cluster algorithm , but the multi - grid algorithm @xcite is available and proves quite efficient for monte - carlo use .
we determine the critical value @xmath28 by the method of binder @xcite , which accelerates the large volume convergence ; extrapolation of @xmath28 to infinite volume is well behaved .
our matching procedure leaves linear in @xmath30 errors in @xmath34 , so this extrapolation is less trivial : for @xmath19 the large volume extrapolation is tougher , and requires a careful accounting of the expected critical behavior .
however , the matching has eliminated more of the lattice spacing correction , see fig .
[ fig2 ] . after a double extrapolation
, we find @xmath37 @xcite .
putting these results together with the results of the matching calculation between theory ii and theory iii , we determine the correction to @xmath38 , at fixed number density @xmath39 , to second order in the scattering length . defining @xmath40 , @xmath41 where @xmath42 is determined by inverting eq .
( [ eq : n_is ] ) .
p. arnold and g. d. moore , phys .
lett . * 87 * , 120401 ( 2001 ) , cond - mat/0103228 .
p. arnold and g. d. moore , phys .
* e * , in production , cond - mat/0103227 .
p. arnold , g. d. moore and b. tomasik , cond - mat/0107124 .
g. baym , j. blaizot , m. holzmann , f. laloe and d. vautherin , phys .
83 * , 1703 ( 1999 ) , cond - mat/9905430 .
p. arnold and b. tomasik , cond - mat/0105147 . | i show how interaction corrections to the bose condensation temperature of an atomic gas can be computed using a combination of perturbative effective field theory and lattice techniques . |
in @xcite , the author has proved the following result which provides both a grss type inequality for orthogonal families of vectors in real or complex inner products as well as , for @xmath0 a counterpart of bessel s inequality .
[ t1.1]let @xmath1 be a family of orthornormal vectors in @xmath2 i.e. , @xmath3 if @xmath4 and @xmath5 @xmath6 @xmath7 a finite part of @xmath8 @xmath9 @xmath10 , and @xmath11
if either@xmath12or , equivalently,@xmath13hold , then we have the inequality@xmath14 ^{\frac{1}{2 } } \notag \\ & \ \ \ \ \ \ \ \ \ \ \ \ \times \left [ \func{re}\left\langle \sum_{i\in f}\gamma _ { i}e_{i}-y , y-\sum_{i\in f}\gamma _ { i}e_{i}\right\rangle \right ] ^{% \frac{1}{2 } } \notag \\ & \leq \frac{1}{4}\left ( \sum_{i\in f}\left\vert \phi _ { i}-\phi _ { i}\right\vert ^{2}\right ) ^{\frac{1}{2}}\cdot \left ( \sum_{i\in f}\left\vert \gamma _ { i}-\gamma _ { i}\right\vert ^{2}\right ) ^{\frac{1}{2}}. \notag\end{aligned}\]]the constant @xmath15 is best possible in the sense that it can not be replaced by a smaller constant . in the follow up paper @xcite , and by the use of a different technique , the author has pointed out the following result as well : [ t2]let @xmath1 , @xmath16 @xmath17 and @xmath18 be as in theorem [ t1.1 ] . if either ( [ 1.1 ] ) or ( [ 1.2 ] ) holds , then we have the inequality@xmath19the constant @xmath15 is best possible in the sense that it can not be replaced by a smaller constant .
it has also been shown that the bounds provided by the second inequality in ( [ 1.3 ] ) and the second inequality in ( [ 1.4 ] ) can not be compared in general .
the following counterpart of bessel s inequality holds . [ t2.1]let @xmath1 be a family of orthornormal vectors in @xmath2 @xmath7 a finite part of @xmath8 and @xmath20 @xmath21 real or complex numbers such that @xmath22 if @xmath23 is such that either 1 .
@xmath24or , equivalently , 2 .
@xmath25 holds , then one has the inequality @xmath26the constant @xmath15 is best possible in the sense that it can not be replaced by a smaller constant .
firstly , we observe that for @xmath27 the following are equivalent@xmath28and@xmath29now , for @xmath30 @xmath31 we have@xmath32 ^{\frac{1}{2 } } \\ & = \left ( \sum_{i\in f}\left\vert \phi _ { i}-\phi _ { i}\right\vert ^{2}\left\vert e_{i}\right\vert ^{2}\right ) ^{\frac{1}{2}}=\left ( \sum_{i\in f}\left\vert \phi _ { i}-\phi _ { i}\right\vert ^{2}\right ) ^{\frac{1}{2}},\end{aligned}\]]giving , for @xmath33 the desired equivalence .
now , observe that@xmath34 -\left\vert x\right\vert ^{2}-\sum_{i\in f}\func{re}\left
( \phi _ { i}% \overline{\phi _ { i}}\right ) , \end{gathered}\]]giving , from ( i ) , that@xmath35 . \label{2.4}\ ] ] on the other hand , by the elementary inequality@xmath36we deduce@xmath37 ^{% \frac{1}{2}}}+\left [ \sum_{i\in f}\func{re}\left ( \phi _ { i}\overline{\phi _ { i}}\right ) \right ] ^{\frac{1}{2}}. \label{2.5}\]]dividing ( [ 2.4 ] ) by @xmath38 ^{\frac{1}{2}}>0 $ ] and using ( [ 2.5 ] ) , we obtain@xmath39 } { \left [ \sum_{i\in f}\func{re}% \left ( \phi _ { i}\overline{\phi _ { i}}\right ) \right ] ^{\frac{1}{2 } } } , \label{2.6}\]]which is also an interesting inequality in itself . using the cauchy - buniakowsky - schwartz inequality for real numbers , we get@xmath40 &
\leq \sum_{i\in f}\left\vert \phi _ { i}\overline{\left\langle x , e_{i}\right\rangle } + \overline{\phi _ { i}}\left\langle x , e_{i}\right\rangle \right\vert \label{2.7 } \\ & \leq \sum_{i\in f}\left ( \left\vert \phi _ { i}\right\vert + \left\vert \phi _ { i}\right\vert \right ) \left\vert \left\langle x , e_{i}\right\rangle \right\vert \notag \\ & \leq \left [ \sum_{i\in f}\left ( \left\vert \phi _ { i}\right\vert + \left\vert \phi _ { i}\right\vert \right ) ^{2}\right ] ^{\frac{1}{2}}\left [ \sum_{i\in f}\left\vert \left\langle x , e_{i}\right\rangle \right\vert ^{2}% \right ] ^{\frac{1}{2}}. \notag\end{aligned}\]]making use of ( [ 2.6 ] ) and ( [ 2.7 ] ) , we deduce the desired result ( 2.1 ) . to prove the sharpness of the constant @xmath41 let us assume that ( [ 2.1 ] ) holds with a constant @xmath42 i.e. , @xmath43provided @xmath44 @xmath45 satisfies ( i ) .
choose @xmath46 @xmath47 @xmath48 @xmath49 @xmath50 with @xmath51 then , by ( [ 2.8 ] ) we get@xmath52provided@xmath53if @xmath54 then obviously ( [ 2.10 ] ) holds , and by ( [ 2.9 ] ) we get@xmath55giving @xmath56 for @xmath57 now , if in this inequality we choose @xmath58 @xmath59 @xmath60 then we get @xmath61 for @xmath62 from where we deduce @xmath63 [ r2.2]by the use of ( [ 2.6 ] ) , the second inequality in ( [ 2.7 ] ) and the hlder inequality , we may state the following counterparts of bessel s inequality as well:@xmath64 ^{\frac{1}{2}}% } \label{2.11 } \\ \times \left\ { \begin{array}{l } \max\limits_{i\in f}\left\ { \left\vert \phi _ { i}\right\vert + \left\vert \phi _ { i}\right\vert \right\ } \sum\limits_{i\in f}\left\vert \left\langle x , e_{i}\right\rangle \right\vert \\ \\ \left [ \sum\limits_{i\in f}\left ( \left\vert \phi _ { i}\right\vert + \left\vert \phi _ { i}\right\vert \right ) ^{p}\right ] ^{\frac{1}{p}}\left ( \sum\limits_{i\in f}\left\vert \left\langle x ,
e_{i}\right\rangle \right\vert ^{q}\right ) ^{\frac{1}{q}},\text { } \\ \hfill \text{for\ }
p>1,\ \frac{1}{p}+\frac{1}{q}=1 \\ \\ \max_{i\in f}\left\vert \left\langle x , e_{i}\right\rangle \right\vert \sum\limits_{i\in f}\left [ \left\vert \phi _ { i}\right\vert + \left\vert \phi _ { i}\right\vert \right ] .%
\end{array}% \right .
\end{gathered}\ ] ] the following corollary holds .
[ c2.3]with the assumption of theorem [ t2.1 ] and if either ( i ) or ( ii ) holds , then@xmath65where@xmath66 \right\ } } { \sum_{i\in f}\func{re}\left ( \phi _ { i}\overline{\phi _ { i}}\right ) } \right ] ^{\frac{1}{2}}. \label{2.12.1}\]]the constant @xmath15 is best possible
. the inequality ( [ 2.12 ] ) follows by ( [ 2.1 ] ) on subtracting the same quantity @xmath67 from both sides . to prove the sharpness of the constant @xmath41 assume that ( [ 2.12 ] ) holds with @xmath42 i.e. , @xmath68provided the condition ( i ) holds .
choose @xmath46 @xmath47 @xmath48 @xmath69 @xmath70 @xmath71 in ( [ 2.13 ] ) to get@xmath72provided@xmath73if @xmath74 @xmath75 @xmath76 then we have@xmath77and by ( [ 2.14 ] ) we get@xmath78now , if we let @xmath79 @xmath80 @xmath81 then obviously@xmath82which shows that ( [ 2.14.1 ] ) is fulfilled , and thus by ( [ 2.15 ] ) we obtain@xmath83for any @xmath84 this implies@xmath85for any @xmath84 finally , let @xmath86 @xmath87 , @xmath88 .
then from ( [ 2.16 ] ) we get @xmath89 for any @xmath90 which produces @xmath91 [ r2.4]if @xmath1 is an orthornormal family in the real inner product @xmath92 and @xmath93 , @xmath94 ( @xmath7 is a finite part of @xmath95 ) and @xmath23 are such that @xmath96 for @xmath94 with @xmath97 and@xmath98then we have the inequality@xmath99 ^{2}\leq \frac{1}{4}\cdot \frac{\sum_{i\in f}\left ( m_{i}-m_{i}\right ) ^{2}}{\sum_{i\in f}m_{i}m_{i}}\cdot \sum_{i\in f}% \left [ \left\langle x , e_{i}\right\rangle \right ] ^{2}. \label{2.17}\]]the constant @xmath15 is best possible
. the following counterpart of the schwarz s inequality in inner product spaces holds .
[ c2.5]let @xmath100 and @xmath101 @xmath102 with the property that @xmath103 if either@xmath104or , equivalently,@xmath105holds , then we have the inequalities@xmath106 } { \sqrt{\delta \overline{\delta } } } \label{2.20 } \\ & \leq \frac{1}{2}\cdot \frac{\left\vert \delta \right\vert + \left\vert \delta \right\vert } { \sqrt{\delta \overline{\delta } } } \left\vert \left\langle x , y\right\rangle \right\vert , \notag\end{aligned}\]]@xmath107@xmath108and @xmath109the constants @xmath110 and @xmath15 are best possible .
the inequality ( [ 2.20 ] ) follows from ( [ 2.6 ] ) on choosing @xmath111 @xmath112
@xmath113 @xmath114 @xmath115 the inequality ( [ 2.21 ] ) is equivalent with ( [ 2.20 ] ) .
the inequality ( [ 2.22 ] ) follows from ( 2.1 ) for @xmath116 and the same choices as above .
finally , ( [ 2.23 ] ) is obviously equivalent with ( [ 2.22 ] ) .
the following result holds .
[ t3.1]let @xmath1 be a family of orthornormal vectors in @xmath2 @xmath7 a finite part of @xmath95 , @xmath117 @xmath118 and @xmath11 if either@xmath119or , equivalently,@xmath120hold , then we have the inequality@xmath121where @xmath122 is defined in ( 2.12.1 ) .
the constant @xmath15 is best possible . using schwartz s inequality in the inner product space @xmath123 one has@xmath124and since a simple calculation shows that @xmath125and @xmath126for any @xmath127 then by ( [ 3.4 ] ) and by the counterpart of bessel s inequality in corollary [ c2.3 ] , we have@xmath128taking the square root in ( [ 3.5 ] ) , we deduce ( [ 3.3 ] ) .
the fact that @xmath15 is the best possible constant follows by corollary [ c2.3 ] and we omit the details .
the following corollary for real inner product spaces holds .
[ c3.2]let @xmath1 be a family of orthornormal vectors in @xmath2 @xmath7 a finite part of @xmath95 , @xmath129 @xmath130 and @xmath100 such that @xmath131 @xmath132 and@xmath133then we have the inequality@xmath134the constant @xmath135 is best possible . in the case where the family @xmath1 reduces to a single vector
, we may deduce from theorem [ t3.1 ] the following particular case first obtained in @xcite .
[ c3.3]let @xmath136 @xmath48 @xmath137 with @xmath138 @xmath139 and @xmath100 such that either @xmath140or , equivalently,@xmath141holds , then@xmath142where@xmath143 } { \func{re}% \left ( \phi \overline{\phi } \right ) } \right ] ^{\frac{1}{2}}.\]]the constant @xmath15 is best possible .
[ r3.4]if @xmath144 is real , @xmath136 @xmath145 and @xmath146 are such that @xmath147 @xmath148 and@xmath149then@xmath150the constant @xmath15 is best possible . if @xmath151 @xmath152 then the following equivalent form of ( [ 3.12 ] ) also holds@xmath153
the following companion of the grss inequality also holds .
[ t4.1]let @xmath1 be a family of orthornormal vectors in @xmath2 @xmath7 a finite part of @xmath95 , @xmath154 , @xmath100 and @xmath155 such that either@xmath156or , equivalently,@xmath157holds .
then we have the inequality@xmath158 \label{4.3 } \\ \leq \frac{1}{16}\cdot \frac{1}{\lambda \left ( 1-\lambda \right ) } \sum_{i\in f}m^{2}\left ( \mathbf{\phi } , \mathbf{\phi } , f\right ) \sum_{i\in f}\left\vert \left\langle \lambda x+\left ( 1-\lambda \right ) y , e_{i}\right\rangle \right\vert ^{2}.\end{gathered}\]]the constant @xmath135 is the best possible constant in ( [ 4.3 ] ) in the sense that it can not be replaced by a smaller constant . using
the known inequality@xmath159we may state that for any @xmath160 and @xmath161 @xmath162since@xmath163for any @xmath127 then , by ( [ 4.4 ] ) , we get@xmath164 \label{4.5 } \\ & = \func{re}\left [ \left\langle x-\sum_{i\in f}\left\langle x , e_{i}\right\rangle e_{i},y-\sum_{i\in f}\left\langle y , e_{i}\right\rangle e_{i}\right\rangle \right ] \notag \\ & \leq \frac{1}{4\lambda \left ( 1-\lambda \right ) } \left\vert \lambda \left ( x-\sum_{i\in f}\left\langle x , e_{i}\right\rangle e_{i}\right ) + \left ( 1-\lambda \right ) \left ( y-\sum_{i\in f}\left\langle y , e_{i}\right\rangle e_{i}\right ) \right\vert ^{2 } \notag \\ & = \frac{1}{4\lambda \left ( 1-\lambda \right ) } \left\vert \lambda x+\left ( 1-\lambda \right ) y-\sum_{i\in f}\left\langle \lambda x+\left ( 1-\lambda \right ) y , e_{i}\right\rangle e_{i}\right\vert ^{2 } \notag \\ & = \frac{1}{4\lambda \left ( 1-\lambda \right ) } \left [ \left\vert \lambda x+\left ( 1-\lambda \right ) y\right\vert ^{2}-\sum_{i\in f}\left\vert \left\langle \lambda x+\left ( 1-\lambda \right ) y , e_{i}\right\rangle \right\vert ^{2}\right ] .
\notag\end{aligned}\]]if we apply the counterpart of bessel s inequality from corollary [ c2.3 ] for @xmath165 we may state that@xmath166now , by making use of ( [ 4.5 ] ) and ( [ 4.6 ] ) , we deduce ( [ 4.3 ] ) .
the fact that @xmath135 is the best possible constant in ( [ 4.3 ] ) follows by the fact that if in ( [ 4.1 ] ) we choose @xmath0 then it becomes ( i ) of theorem [ t2.1 ] , implying for @xmath167 ( [ 2.12 ] ) , for which , we have shown that @xmath15 was the best constant .
[ r4.2]if in theorem [ t4.1 ] , we choose @xmath168 then we get@xmath169 \leq \frac{1}{% 4}m^{2}\left ( \mathbf{\phi } , \mathbf{\phi } , f\right ) \sum_{i\in
f}\left\vert \left\langle \frac{x+y}{2},e_{i}\right\rangle \right\vert ^{2 } , \label{4.7}\]]provided@xmath170or , equivalently,@xmath171
let @xmath172 be a measure space consisting of a set @xmath173 a @xmath174algebra of parts @xmath175 and a countably additive and positive measure @xmath176 on @xmath175 with values in @xmath177 let @xmath178 be a @xmath179measurable function on @xmath180 denote by @xmath181 the hilbert space of all real or complex valued functions defined on @xmath182 and @xmath183integrable on @xmath173 i.e.,@xmath184 consider the family @xmath185 of functions in @xmath186 with the properties that@xmath187where @xmath188 is @xmath189 if @xmath4 and @xmath190 if @xmath191 @xmath192 is an orthornormal family in @xmath193 [ p5.1]let @xmath185 be an orthornormal family of functions in @xmath194 @xmath7 a finite subset of @xmath8 @xmath195 @xmath196 such that @xmath197 and @xmath198 so that either@xmath199 d\mu \left ( s\right ) \geq 0 \label{5.3}\]]or , equivalently,@xmath200then we have the inequality@xmath201 ^{\frac{1}{2 } } } \label{5.5}\]]@xmath202 ^{\frac{1}{p}}\left ( \dsum\limits_{i\in f}\left\vert \dint_{\omega } \rho \left ( s\right ) f\left ( s\right ) \overline{f_{i}}\left ( s\right ) d\mu \left ( s\right ) \right\vert ^{q}\right ) ^{\frac{1}{q}},\text { } \\
\hfill \text{\ for \ } p>1,\ \frac{1}{p}+\frac{1}{q}=1 \\ \\ \max\limits_{i\in f}\left\vert \dint_{\omega } \rho \left ( s\right ) f\left ( s\right ) \overline{f_{i}}\left ( s\right ) d\mu \left ( s\right ) \right\vert \dsum\limits_{i\in
f}\left [ \left\vert \phi _ { i}\right\vert + \left\vert \phi _ { i}\right\vert \right ] .%
\end{array}% \right.\]]in particular , we have@xmath203the constant @xmath15 is best possible in both inequalities .
[ p5.2]assume that @xmath204 and @xmath7 satisfy the assumptions of proposition [ p5.1 ] .
then we have the following counterpart of bessel s inequality:@xmath205where , as above,@xmath206 \right\ } } { \func{re}\left ( \phi _ { i}\overline{\phi _ { i}}% \right ) } \right ] ^{\frac{1}{2}}. \label{5.8}\]]the constant @xmath15 is the best possible .
[ p5.3]let @xmath185 and @xmath7 be as in proposition [ p5.1 ] . if @xmath207 @xmath208 and @xmath209 so that either@xmath210 d\mu \left ( s\right ) & \geq 0 , \label{5.9 } \\
\int_{\omega } \rho \left ( s\right ) \func{re}\left [ \left ( \sum_{i\in f}\gamma _ { i}f_{i}\left ( s\right ) -g\left ( s\right )
\right ) \left ( \overline{% g}\left ( s\right ) -\sum_{i\in f}\overline{\gamma _ { i}}\text { } \overline{f_{i}% } \left ( s\right ) \right ) \right ] d\mu \left ( s\right ) & \geq 0 , \notag\end{aligned}\]]or ,
equivalently,@xmath211then we have the inequality@xmath212where @xmath122 is defined in ( 5.8 ) .
[ c5.4]let @xmath185 be an orthornormal family of functions in the real hilbert space @xmath213 @xmath7 a finite part of @xmath8 @xmath96 @xmath21 with @xmath214 and @xmath215 so that@xmath216then we have the inequality@xmath217 ^{2 } \label{5.13 } \\ & \leq \frac{1}{4}\cdot \frac{\sum_{i\in f}\left ( m_{i}-m_{i}\right ) ^{2}}{% \sum_{i\in f}m_{i}m_{i}}\cdot \sum_{i\in f}\left [ \int_{\omega } \rho \left ( s\right ) f\left ( s\right ) f_{i}\left ( s\right ) d\mu \left ( s\right ) \right ] ^{2}.
\notag\end{aligned}\]]the constant @xmath15 is best possible .
[ c5.5]let @xmath185 and @xmath7 be as above . if @xmath218 @xmath208 with @xmath219 and @xmath220 such that@xmath221and@xmath222then we have the inequality@xmath223 ^{\frac{1}{2}}. \notag\end{gathered}\ ] ] 9 s.s .
dragomir , a counterpart of bessel s inequality in inner product spaces and some grss type related results , _ rgmia res .
_ , * 6*(2003 ) , supplement , article 10 . `
[ on line : http://rgmia.vu.edu.au/v6(e).html ] ` | some new counterparts of bessel s inequality for orthornormal families in real or complex inner product spaces are pointed out .
applications for some grss type inequalities are also empahsized . |
the formation and evolution of massive galaxies is one of the most studied and debated topics in extragalactic astronomy today . in the local universe the most massive galaxies primarily populate a well defined relation in the color - magnitude space known as the `` red - sequence '' ( e.g. , * ? ? ?
* ; * ? ? ?
these galaxies host mostly old , passively evolving stellar populations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) .
the major processes behind mass assembly and structure formation of massive galaxies , however , are controversial . in particular , the role of major mergers ( as advocated in e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) in the mass build up and in the quenching of the star formation has recently been challenged by new evidence pointing towards a mass - induced truncation of the star - formation ( e.g. , * ? ? ?
* ; * ? ? ?
evidence supporting the merger picture is mostly restricted to @xmath8 from the tightness of the galaxy color - magnitude relation ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , the evolution of the mass function of red - sequence galaxies ( e.g. , * ? ?
* ; * ? ? ?
* ; but see cimatti et al.2007 , scarlata et al.2007 ) and the small scatter in the mass - to - light ratios of these systems ( e.g. , * ? ? ?
the super - solar @xmath9-element abundances found in local massive galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) can be reproduced by both merger- and mass - induced quenching as long as the process took place several gyr ago and on short time - scales ( @xmath10 , * ? ? ?
* ; * ? ? ?
recently _ minor _ mergers have also been considered as a viable mechanism for the size growth of massive galaxies , gaining more popularity after the discovery of the strong size evolution of early - type galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the discovery of the compact nature of massive quenched galaxies at @xmath11 implies they must become @xmath12-@xmath13 larger with time to match the sizes of quenched galaxies in the local universe ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
other authors , however , have suggested that the evolution of the mass - size relation is primarily driven by the appearance of new large galaxies at later time rather than by the growth of individual objects ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the epoch between @xmath14 is crucial for the assembly of massive galaxies , as most of their size and number density evolution seem to have taken place during this time ( e.g. , * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
critical to study this redshift range are data covering the rest - frame optical , and in fact , the past few years have seen the completion of deep ground- and space - based photometric ir surveys .
although large samples of both star - forming ( sf ) and quenched massive galaxies at @xmath14 can be assembled using color - selection techniques ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) , results based on them are still hampered by the lack of spectroscopic redshifts , which translates into strong systematic effects on stellar population parameters ( including , but not limited to , stellar masses and luminosity - weighted ages ) .
moreover , the mutual contamination between color - selected sf and quenched samples can only be established statistically as , typically , no further information is available for individual sources ( e.g. , @xmath15 contamination for @xmath16 selection , cameron et al.2011 ) .
detailed spectroscopic studies on individual sources are still extremely rare , and limited to the brightest , non representative sources @xcite .
additional complications appear when quiescent galaxies at @xmath17 are considered . without emission lines ,
the redshift determination must be constrained with continuum - emission features , like the 4000 and balmer breaks , redshifted to the near - ir regime .
this requires deep ir spectroscopic observations , which ( until recently ) were limited to single object spectroscopy .
these studies have shown that the bulk of the star formation in massive quenched galaxies took place between @xmath18 with formation time - scales below 1gyr @xcite .
ground - based near - ir spectroscopy , however , is limited to the @xmath19 , @xmath20 , and @xmath21 atmospheric windows , and suffers from high background emission and time - variable absorption features .
spectroscopic observations from space , then , represent the best way to acquire large samples of quiescent galaxies at @xmath22 with homogeneous spectroscopic data sets . here
we present a stellar population study of a sample of quenched galaxies at @xmath23 discovered in the wide field camera 3 ( wfc3 ) infrared spectroscopic parallel survey ( wisp , * ? ? ?
the hst near - ir grism spectroscopy covers the wavelength range @xmath24 m @xmath25 m , and allows us to study quenched high-@xmath26 galaxies on an individual basis .
the paper is structured as follows : in 2 we present the observations and analysis of photometry and spectra ; in 3 , we describe the color and magnitude selection of our sample . in
4 we characterize our galaxy sample , including the calculation of spectro - photometric redshifts , stellar population parameters and estimates of their systematic and random uncertainties . in 5
we discuss our main results , and summarize them in 6 .
five appendices describing details in the data analysis are included at the end of the paper . in this paper , we assume a flat cosmology with @xmath27 , @xmath28 , and @xmath29 .
photometric magnitudes are always expressed in the ab system @xcite .
the sample presented in this work was extracted from the first 27 wisp fields observed with the wfc3 on board of the hst . specifically , these fields were observed with both wfc3 grisms , and with both wfc3/ir and wfc3/uvis cameras .
the wfc3/ir provides a plate scale of 0.13@xmath30 , over a total field of view ( fov ) of 123@xmath31136@xmath32 .
the uvis channel of the wfc3 has a plate scale of 0.04@xmath30 , over a total fov of 162@xmath33162@xmath34 . for 7 of our fields
the uvis photometry was obtained from @xmath35 binned data ( see table[tab : schedule ] ) .
all the observations were carried out in pure parallel mode . in each field , imaging was obtained using the f475x , f600lp , f110w and f160w filters , with typical exposure times of 400 , 400 , 1000 , and 500 seconds , respectively ( see table[tab : schedule ] for details ) .
the two deeper fields par96 and par136 were observed in the uvis bands f606w and f814w instead . the wfc3/uvis ccds suffer a degradation of their charge transfer efficiency with time , affecting f475x ( f606w ) and f600lp ( f814w ) fluxes .
a flux correction was implemented and it is fully described in appendix [ app : cte ] .
dispersed images were obtained using the wfc3/ir camera and the g102 and g141 near - ir grisms .
the blue ( g102 ) grism provides a resolving power @xmath36 for a point source , and covers the 0.8 - 1.15@xmath37 m wavelength range .
the red ( g141 ) grism provides a resolving power @xmath38 for a point source , and covers the 1.07 - 1.7@xmath37 m range .
the wavelength overlap between the two grisms ensures an accurate flux calibration of the spectra .
details of reduction and calibration of the wfc3 data are presented in @xcite .
l c c c c c c c c c par95 & 01 10 04.38 & @xmath3902 24 54.9 & 909 & 5715 & 384 & 2209 & 400 & 400 & 2100 + par91 & 01 10 05.79 & @xmath3902 25 03.3 & 909 & 5715 & 384 & 2209 & 400 & 400 & 2100 + par97 & 01 10 06.30 & @xmath3902 23 44.7 & 859 & 5515 & 406 & 2109 & 400 & 400 & 2100 + par84 & 01 10 07.60 & @xmath3902 25 11.6 & 1162 & 7521 & 559 & 2809 & 400 & 400 & 2100 + par79 & 01 10 08.96 & @xmath3902 25 16.2 & 1187 & 7521 & 534 & 2809 & 400 & 400 & 2100 + par81 & 01 10 09.12 & @xmath3902 22 17.1 & 1187 & 7521 & 534 & 2809 & 400 & 400 & 2100 + par80 & 01 10 09.27 & @xmath3902 22 17.7 & 1187 & 7521 & 534 & 2809 & 400 & 400 & 2100 + par89 & 01 10 09.96 & @xmath3902 22 20.0 & 1187 & 7512 & 534 & 2809 & 400 & 400 & 2100 + par83 & 01 10 10.71 & @xmath3902 24 09.7 & 859 & 5515 & 406 & 2109 & 400 & 400 & 2100 + par96 & 02 09 24.40 & @xmath3904 43 41.6 & 4295 & 28081 & 1765 & 11430 & 3000@xmath40 & 3000@xmath40 & 2100 + par74 & 09 10 48.14 & + 10 17 20.3 & 1065 & 5918 & 431 & 2306 & 400 & 400 & 2100 + par87 & 09 46 46.39 & + 47 14 58.2 & 912 & 4915 & 406 & 1906 & 400 & 400 & 1500 + par114 @xmath41 & 10 40 58.09 & + 06 07 31.0 & 1137 & 7221 & 456 & 2909 & 600 & 600 &
+ par131 @xmath41 & 10 48 22.94 & + 13 03 50.5 & 2171 & 13039 & 884 & 5215 & 600 & 600 & + par115 @xmath41 & 11 18 55.08 & + 02 17 09.6 & 912 & 5215 & 381 & 2106 & 600 & 600 &
+ par135 @xmath41 & 11 22 24.01 & + 57 50 58.9 & 862 & 4712 & 406 & 1906 & 600 & 600 & 1500 + par136 @xmath41 & 12 26 28.84 & + 05 23 02.9 & 3036 & 18857 & 1137 & 7318 & 2000@xmath40 & 2000@xmath40 & + par76 & 13 27 22.17 & + 44 30 39.3 & 887 & 5515 & 406 & 2006 & 400 & 400 & 1500 + par120 @xmath41 & 13 56 51.50 & + 17 02 33.9 & 837 & 4512 & 381 & 1806 & 600 & 600 & 1500 + par73 & 14 05 12.86 & + 46 59 19.9 & 1034 & 6118 & 456 & 2509 & 400 & 400 & 1500 + par64 & 14 37 29.04 & @xmath3901 49 49.5 & 1112 & 5918 & 456 & 2306 & 400 & 400 & 2100 + par66 & 14 37 29.22 & @xmath3901 49 54.5 & 1237 & 7421 & 559 & 2909 & 400 & 400 & 2100 + par67 & 15 24 07.75 & + 09 54 53.9 & 959 & 5715 & 406 & 2209 & 400 & 400 & 1500 + par69 & 15 24 09.75 & + 09 54 50.0 & 1087 & 5721 & 431 & 2309 & 400 & 400 & 1500 + par94 & 22 05 26.66 & @xmath3900 17 48.5 & 1624 & 9024 & 534 & 3309 & 400 & 400 & 2100 + par68 & 23 33 33.04 & + 39 21 20.5 & 1215 & 7721 & 534 & 3009 & 400 & 400 & 1500 + par147 @xmath41 & 23 58 19.72 & @xmath3910 14 56.36 & 962 & 5418 & 406 & 2106 & 600 & 600 & + + [ tab : obs_table ] for 22 of the 27 fields we also obtained _ spitzer _ irac observations at 3.6@xmath37 m ( see table [ tab : schedule ] for details ) . the same pipeline used to produce the `` spitzer enhanced imaging products '' was also run on our data to measure the 3.6@xmath37 m fluxes ( capak et al . , in preparation ) . in the analysis of the galaxy sample
we simultaneously make use of the broad band fluxes and near - ir spectra .
thus , special attention should be placed to make sure that the aperture used for the extraction of the spectra and the aperture used to compute the broad band fluxes are the same .
in order not to introduce artificial correlations among adjacent pixels we worked on images with the original pixel scales . as a consequence , typical softwares used to compute object fluxes in matched apertures in different images could not be used . in order to create the multi - band catalog of aperture - matched fluxes we proceeded as follow .
we first used the sextractor software @xcite to detect objects in the deepest @xmath42-band images and to compute the parameters of the kron elliptical apertures ( semi - axes and position angle ) .
the kron apertures , as computed by sextractor , are intended to give the most precise estimate of total magnitudes for galaxies .
details about the routine can be found in the sextractor manual .
for each galaxy , we then computed the total flux in all bands , using the @xmath42-band - defined aperture and a custom idl code .
the elliptical apertures were then appropriately scaled to account for different pixel scales between the uvis and ir detectors .
the local background for each elliptical aperture was calculated within a square annulus , using the same prescription as in sextractor . within each annulus ,
the background per pixel was computed as a 3-@xmath43 clipped mean .
then , the total background - subtracted flux was computed within the aperture by adding all the aperture pixels .
flux uncertainties were computed using monte carlo simulations .
we created 1,000 images for each galaxy , by randomizing the flux in each pixel of the elliptical aperture according to a normal distribution with a width provided by the @xmath44 error map .
we measure the total elliptical flux as explained above on each of the simulated images .
we then computed the @xmath44 errors by fitting a gaussian to the distribution of the 1,000 measured fluxes .
we only retain in the catalog sources with @xmath42-flux larger than @xmath45 ( @xmath42 is our deepest near - ir band , see table [ tab : schedule ] ) .
for undetected sources in the other bands , we show the @xmath45 flux limit in the figures . because of the significantly lower spatial resolution of the _ spitzer _ images , irac 3.6@xmath37
m fluxes were computed independently from the hst fluxes , using the `` spitzer enhanced imaging products '' pipeline , with some modifications ( details in capak et al .
, in prep . ) .
briefly , the pipeline utilizes the spitzer mopex software package to create a mosaic for each irac image , before identifying and extracting sources using sextractor .
fixed aperture fluxes were computed for each object , using an aperture of 28 diameter .
this is a smaller aperture compered to the typical 38 and 58 diameter apertures used in the `` spitzer enhanced imaging products '' catalog .
however , our exposures are significantly deeper than the standard images analyzed with the pipeline , so a smaller aperture is more appropriate to reduce contamination from nearby neighbors .
the aperture fluxes were corrected to total fluxes assuming the objects are point sources .
this is a fair assumption as the fwhm of the _ spitzer _ point - spread function ( psf from now on ) of 15 is larger than the fwhm in near - ir of the largest resolved sources used in this paper ( 143 ) .
the _ spitzer _ and wfc3 catalogs were cross - matched using target coordinates , by searching within a radius of 15 .
we flagged all sources for which multiple objects fell within the psf of the irac data .
this procedure results in a 5-broad - band photometric catalog for our 27 fields of 15,302 sources brighter than 27.5 mag in @xmath42-band . of those ,
6,991 have irac 3.6@xmath37 m detections , 26% of which are blended in this band . due to the slitless nature of the spectroscopic observations , spectra from different sources may overlap .
if not properly removed , the flux contribution from nearby sources may substantially change the total flux and shape of the extracted spectra , preventing an accurate fit of the spectral energy distribution ( sed ; e.g. , * ? ? ?
in order to perform an optimal cleaning of the 2d data ( reduced and callibrated with axe software , * ? ? ?
* see atek et al .
2010 for details ) the 1d spectra were extracted using a custom written idl code .
the detail of our extractions and a comparison with the axe 1d extractions are presented in appendix[app : extraction ] . shortly , while axe assumes a single gaussian function to describe the spatial light profile of each source , we use a double gaussian function .
our approach results in more reliable flux levels compared to extractions performed by the axe code .
images shows and overall good agreement between them .
the comparison was based on 102 galaxies for which the spectral @xmath46 magnitudes were calculated after contamination correction.,width=321 ] in order to test the quality of our spectral cleaning and extraction , we compare the fluxes from the spectra with our nir photometric data . in figure
[ fig : f160wg141comp ] we show a comparison between the magnitudes computed from the extracted g141 spectra and the magnitudes computed from the @xmath46 images , for 102 galaxies .
the magnitudes from the spectra were computed by convolving them with the throughput of the @xmath46 filter , which is fully covered by the g141 spectral range .
the median magnitude difference between photometry and grism data is @xmath47mag with a @xmath44 scatter of @xmath48mag .
the 4000 break fully falls within the wavelength range covered by the grism spectra for objects in the @xmath49 redshift range .
this break can be used not only as redshift indicator , but also to study stellar population properties in galaxies with sufficient s / n spectra ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we limit the analysis of the spectra only to objects with @xmath42@xmath39@xmath46 @xmath50 , and magnitude brighter than @xmath46@xmath51 . as shown in figure[fig : color_z ] , the color cut preselect sources with spectral breaks broadly covered by our spectra , for which the redshifts can be measured ; while the magnitude cut ensures we have spectra with sufficient s / n . among all sources with @xmath46 @xmath52 we removed the stars using a size - magnitude diagram as explained in appendix[app : rm_stars ] .
after visual inspection of all the sources flagged as stars ( 315 ) , we found that 10 were misclassified galaxies , and were re - included in the sample .
@xmath39@xmath46 color versus redshift diagram based on @xcite models shows that with our color selection criterion ( @xmath42@xmath39@xmath46@xmath53 0.6 , dashed black line ) we mostly recover simple stellar population models ( quenched , red triangles ) between @xmath54 .
dust - free continuous star formation models ( blue diamonds ) are rejected at all redshifts , while some dusty ( @xmath55 ) continuous star formation models ( green crosses ) at @xmath56 pass our color cut .
for all models , solar metallicity and a redshift of formation of 4 have been assumed.,width=340 ] in the left panel of figure[fig : colselection ] we show the @xmath42@xmath39@xmath46 versus @xmath57@xmath39@xmath42 color - color diagram for 1,352 galaxies brighter than @xmath46@xmath58 .
the right panel is the corresponding plot for the two deep wisp fields , showing @xmath42@xmath39@xmath46 versus @xmath59@xmath39@xmath42 instead ( 163 galaxies ) .
sources brighter than @xmath46@xmath58 constitute @xmath60 of our photometric catalogue . out of 283 galaxies with @xmath42@xmath39@xmath46 @xmath61 ,
84 ( 5 ) are upper limits in the @xmath57 ( @xmath59 ) band and we show the @xmath62 limits in the color - color plots .
different model tracks showing the color evolution with redshift are overplotted ( see figure caption and legend ) .
note that at the highest redshift considered ( @xmath63 ) , the truncated 200myr burst model ( orange line ) is already passive , but significantly bluer in @xmath57@xmath39@xmath42 than the instantaneous burst at the same redshift ( red line ) .
figures[fig : color_z ] and [ fig : colselection ] show that dusty sf galaxies are also selected by our @xmath42@xmath39@xmath46 color criterion .
however , these galaxies can be identified later through the sed fitting of their combined photometry@xmath64spectra ( see section[sec : results ] ) .
we visually inspected both the spectra and the direct images of all the objects satisfying the @xmath42@xmath39@xmath46 color cut to remove galaxies for which the spectrum showed instrumental issues ( e.g. , for sources at the edge of the chip for which the spectrum is truncated ) , for which the spectrum was contaminated at a level that even our procedure could not recover ( e.g. , close to a bright star ) , and for which the wfc3/_spitzer _ photometry show blended sources .
we excluded 149 galaxies in this step .
note that about half of them were excluded because the full wavelength range of their spectra was not totally sample within the ccd detector ( i.e. , truncation ) .
we also excluded 32 galaxies for which the 2d dispersed stamps did not show any signal .
we compared these 32 galaxies with the parent sample , and found that they are at the faint end of the considered magnitude range ( with average @xmath46-band of 22.7 ) , and tend to have slightly larger radii than the remaining sources of the same magnitude .
this may introduce a bias against less concentrated objects in our faintest magnitude bin .
our _ master _ sample includes 102 sources with @xmath42@xmath39@xmath46@xmath65 and @xmath46@xmath52 . in 23 out of 102 galaxies we detected emission lines .
it is important to note that , because in slitless mode the spectral resolution depends on the size of the object in the dispersion direction , our capability to detect emission line galaxies depends not only on the line flux and equivalent - width , but also on the size of the galaxy itself .
a detailed line - completeness analysis for the full wisp survey is presented in colbert et al.(2013 , submitted ) .
the emission line recovery rate drops below 40% for objects
larger than 06 , the median size of our sample galaxies is 045 , corresponding to a completeness of about 70% . as discussed in colbert et al .
, the completeness never reaches 100% because of spectral contamination issues .
our sample of 102 galaxies does not suffer from this problem , so we expect the completeness rate in our case to be even larger than 70% .
previous to derive spectro - photometric redshifts and stellar population properties ( spps ) from our analysis , we study the effects of emission lines in retrieving these continuum - based parameters . our low spectral resolution might `` dilute '' emission lines in the continuum , affecting the retrieved redshifts and spps .
the results of this study are presented in appendix[app : emilines ] .
briefly , we find that in those cases were emission lines are diluted in the continuum , the flux contribution from the line is below the @xmath66 uncertainties in the grism data .
therefore , in addition of masking the detected emission lines in the seds , there is no need to make specific modeling of emission lines for our sed fitting process and simulations .
we compute spectro - photometric redshift for the 102 galaxies in our sample by fitting stellar population models ( see below ) to the combination of the photometric and spectroscopic data .
we developed our own idl code for this task because of the @xmath67 different spectral resolution in the two grisms , @xmath68 different spectral resolution in each individual source , and @xmath69 simultaneous fit to photometry and spectra .
we computed spectro - photometric redshifts by fitting 51 empirical templates from @xcite and @xcite to the @xmath70(@xmath71 ) , @xmath57(@xmath59 ) , g102 and g141 spectra , and irac 3.6@xmath37 m data , when available . before the @xmath72 minimization
, we matched the spectral resolution of the templates to that of the data ( on an object - by - object basis ) .
to do so , we re - binned both the data and the templates to the resolution element , computed from the fwhm of the object s spatial profile in the dispersion direction .
given the typical size of our sources , the spectral resolution elements are on average @xmath73 ( @xmath74 ) in g102 ( g141 ) .
we considered the @xmath75 range with constant redshift steps of 0.05 , and applied intergalactic medium absorption following @xcite . for sources with emission lines the purely spectroscopic redshift based on those features
was used instead ( @xmath76 of our master sample ) .
we determined redshifts for single - line galaxies assuming they are h@xmath9 emission .
see @xcite and colbert et al.(2013 , submitted ) for further details on emission line redshift estimations .
we estimated random errors on the redshifts by using monte carlo simulations .
for each galaxy , we varied both the spectra and the photometric points within the @xmath77 uncertainties , assuming gaussian distribution .
we created 500 realizations for each galaxy , and derived the random error on the redshift as the 1@xmath43 width of the resulting distribution .
( dashed line).,width=340 ] the spectro - photometric redshift distribution for the full master sample of 102 galaxies is shown in figure [ fig : z ] .
the distribution shows that 77% of the selected galaxies are at @xmath78 . [ cols="<,^,>,^,>",options="header " , ] we compute luminosity - weighted age , stellar mass , star formation rate ( sfr ) and star - formation history ( sfh ) of the master sample galaxies , using the same custom idl code applied in the previous section , keeping the redshift fixed and using the @xcite library of stellar population synthesis models .
we consider seven sfhs ( continuous , exponentially declining with
_ e_-folding times @xmath79gyr , and exponentially increasing with @xmath80gyr ) ; 70 log - binned ages between 10myr and 12gyr ; salpeter initial mass function ( imf , * ? ? ? * ) and solar metallicity .
we use a salpeter imf based on recent results on local massive galaxies . a variety of observations , including stellar kinematics ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , stellar populations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and gravitational lensing ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , suggest ellipticals and spiral bulges show `` heavier '' imfs ( salpeter - like , with larger fractions of low - mass stars ) than galaxy disks ( chabrier - like imf , * ? ? ?
thus , the assumption of a salpeter imf seems to be more representative of the massive and quenched galaxy population we address in the present study .
we note that if a chabrier - like imf was used , our inferred masses and sfrs would be smaller by a factor @xmath81 .
accordingly , our models and the mass - completeness limit we inferred from them ( see section [ sec : masscompl ] ) would be smaller in a similar amount without compromising our main results and conclusions .
we also note that our _ specific _ sfrs ( ssfrs ) are robust to imf changes .
this is of particular importance as we base our quenched / sf classification on this parameter .
systematic changes of the imf with galaxy mass are unlikely in the mass range considered in this work , and are briefly discussed at the end of appendix [ app : systerr ] .
solar metallicity is in agreement with recent spectroscopic results by @xcite on quenched galaxies at @xmath82 .
we considered a range of extinctions ( @xmath83 ) , and we used the @xcite extinction law .
the age of the stellar population is constrained to be smaller than the age of the universe in the adopted cosmology .
analogously to the redshift random error determination , random uncertainties on the spps were derived using monte carlo simulations .
we used the results of the stellar population modeling to select quenched galaxies as commonly done in the literature , based on the ssfr : _ quenched _ galaxies have @xmath84 , and constitute 71% of our master sample .
the statistical uncertainties associated with the measurements of redshifts and stellar population parameters are assessed in sections [ sec : redshift ] and [ sec : sedfit ] . here
, we further quantify the _
systematic _ uncertainties introduced by the degeneracies in stellar population models .
it is well known that , if the photometric / spectroscopic data do not provide adequate constraints , studying spps of galaxies through sed fitting may produce strongly degenerate results by different combinations of age , sfr , sfh and extinction , ( e.g. , * ? ? ? * and references therein ) .
for this reason we performed a set of simulations to assess how well we can recover the spps with the available data .
the simulations are described in detail in appendix[app : systerr ] .
briefly , we used @xcite models with known stellar population properties and redshifts to simulate our spectroscopic and photometric data .
these `` model data '' were treated in the same way as the real data : we applied both the color and magnitude cuts as for our galaxies , and the selected model data were fitted with our customized idl code to recover their spps . in this section ,
we only summarize the main results of these tests and their implications for our galaxy sample . not surprisingly , the main conclusion of our study was that , in general , the spps were recovered more accurately for galaxies at @xmath85 , where the 4000@xmath86 break was fully covered by the grism spectra .
more degeneracies appeared when the break was only coarsely covered by the uvis photometric points . at redshift @xmath87 ,
our simulations show we can recover the age measurements to within @xmath88 of the age value for our complete age range .
we can also distinguish between short ( @xmath89myr ) and long ( @xmath90gyr ) sfhs , although we can not separate sfhs with @xmath91myr and @xmath92myr .
most importantly , we find that our distinction between quenched and sf galaxies ( defined by a rough limit at @xmath93 ) is very robust . at lower redshifts , we found we can not reliably recover spectrophotometric redshifts below 1
. therefore we decided to exclude from our galaxy sample all sources below this redshift .
at @xmath94 we recovered most of the spp values as at @xmath95 , although with larger scatter . at these lower redshifts
, however , we found that the age and extinction show strong systematic offsets with respect to the input parameter values .
this was mainly produced by a strong anti - correlation between both galaxy properties . however , because the distinction between quenched and sf galaxies is defined as function of the ssfr only , even at @xmath96 we can robustly distinguish between both galaxy types .
the stellar mass is the most robust and best constrained parameter at all redshift .
overall , the addition of irac @xmath97 data to the sed fits slightly reduces the scatter on the ages and stellar - mass - systematic uncertainties .
the use of irac data provides the strongest constraints in the determination of sfr , were the scatter of our models is clearly reduced .
the recovered redshifts also suffer a very mild improvement in their systematic uncertainties when irac data is used , while the other spps are mostly unchanged . from the discussion above
, we decided to remove from the following analysis all galaxies with @xmath98 .
also , we will not to use in our analysis the age of galaxies at @xmath99 and the extinction of all galaxies independently of their redshift . in table[tab : systerr ] we present a summary of the results from the systematic error analysis for redshift and spps , including results with and without irac data in the sed fits .
formally , many of the systematic offsets found are consistent with zero within the uncertainties . however , some cases have larger uncertainties than others . therefore in those cases
we present our results as zero - systematic - error @xmath100 their 68%-percentile .
the estimate of the mass completeness is not straight forward in a magnitude _ and _ color selected sample like ours .
together , these criteria imply not only a segregation in mass but also in current star formation activity and sfh .
first we explore the @xmath46-band completeness of our photometric catalog . in the left panel of figure[fig : masscompl ] we show the @xmath46-band histogram for our entire photometry ( section[sec : photcat ] ) .
we proceed by run 1,000 simulations of our @xmath46-band fov and re - extract with sextractor our simulated galaxies .
each of the 1,000 galaxies per simulation was modeled as a random combination of observational properties which ranges were set based on the data .
we allowed a range of @xmath46-band magnitudes between 26 and 15 .
then , light profiles were modeled with a 2d s@xmath101rsic function with @xmath102 , centered at random positions , with ellipticities between 0 and 1 , and random projected orientations .
the effective radii were allowed to vary between 0.2 and 06 ( typical range for our galaxies ) .
we add random noise to every single model run such as an average @xmath46=22 mag galaxy has an integrated s / n=200 , like in the data .
after extracting galaxies from each simulation we study the completeness as a function of @xmath46-band magnitude , in bins of 0.2mag .
we found we are 100% complete down to magnitudes of 24.2 . , reaching 50% completeness at 25 mag .
we also study completeness as function of other parameters .
based on the effective radius we are 100% complete up to 0.35 arcseconds , reaching 70% completeness at 043 .
in terms of the s@xmath101rsic index we are 70% complete down to @xmath103 . finally , we are 75% complete for ellipticities up to 0.4 . based on these results , our magnitude cut at @xmath46@xmath51 for our galaxy sample ( section [ sec : selectetg ] )
is conservative enough to allowed us study a 100% complete sample in @xmath46-band with also very high completeness in other observables . in the right panel of figure
[ fig : masscompl ] we show the galaxy stellar mass as a function of redshift for our galaxies ( after cleaning the sample of cases with possible strong systematics in their spps , see previous section ) . due to our @xmath46@xmath51 magnitude cut , the lower mass envelope changes with redshift , and depends on the galaxy sfh and age . in order not to introduce any bias in the analysis , we conservatively estimate the mass completeness limit to be the mass of the model with the highest m / l ratio for @xmath46@xmath51 at each redshift , as indicated by the solid red curve in the figure@xmath51 , for an exponentially declining sfh , with @xmath104 myr , and formation redshift @xmath105 . in
what follows , we adopt a mass completeness limit of @xmath106 m@xmath107 ( dashed black line in right panel of figure [ fig : masscompl ] ) , which is the minimum mass measurable at @xmath108 , with our conservative assumptions .
we note that @xmath109 of our galaxies were above this mass cut . of them , 40% are at or below redshift 1.5 .
the final sample of quenched galaxies includes sources at @xmath110 . at @xmath111
our mass limit implies that we find only a few of the rarest most massive galaxies , while at @xmath112 , our simulations show that we can not reliably retrieve photometric redshifts with our data ( see section[sec : systerr ] and appendix[app : systerr ] ) .
after the redshift and ssfr cuts , our sample includes 41 quenched sources with @xmath113 , @xmath114 , @xmath115 , and redshifts in the range @xmath116 .
four of the quenched galaxies ( @xmath117 ) have detected emission lines .
we do not remove these galaxies from the sample of quenched objects for the following reasons : first , if star - formation is at the origin of the detected emission lines , the line fluxes implies sfrs of about 0.4@xmath118 ( using a kennicutt 1998 conversion ) , which , given the masses of these objects , would still make them quenched according to our ssfr definition .
second , the emission lines might be produced by ionization due to an active galactic nucleus ( agn ) , rather than star formation .
various studies connect agn activity with the quenching of star formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , so we do not want to bias our result against agn activity .
because we lack information on the origin of the gas ionization , we leave the 4 emission line galaxies in the sample , and identify them in the analysis when needed . as it will be shown in the following sections , the presence of quiescent galaxies with emission lines does not affect our main results and conclusions . in figure[fig
: spectra ] ( available in the electronic format of the paper ) we show the @xmath42-band postage stamp of each galaxy of our final sample . in the same figure we present the g102 and g141 spectra together with the wfc3/uvis and irac 3.6@xmath37 m photometric points used in the sed fits . the best sed model fit and some spps
are also included .
this 41-galaxy sample will be used in the rest of the paper .
in the following section we discuss the properties of the final sample of 41 quenched galaxies .
( in red ) .
dashed line corresponds to the median of the distribution .
we compare our median age for quenched galaxies with ages from early - type galaxies from literature : in turquoise , @xcite mean age range ( predictions from different models ) of 10 etgs at @xmath108 ; in purple , @xcite age range of 15 low-@xmath119-emission galaxies at @xmath120 ; in green the mean age of 18 @xmath121 passively evolving galaxies from @xcite.,width=340 ] quenched galaxies emphasizes the differences between on - red - sequence sources ( top , 19 sources ) and off - red - sequence galaxies ( bottom , 13 sources ) .
stacked spectra and uvis data points for individual galaxies are shown in orange while best fit @xcite models to the spectra are shown in black .
the models are exponentially declining sfhs with @xmath122 , solar metallicity , salpeter imf and ages of 3.0 and 1.4gyr for on- and off - red - sequence galaxies , respectively . in the top of each panel
we show the number of galaxies used in the coadded spectrum as function of wavelength .
flux densities normalized to median flux between 5500 and @xmath123.,title="fig:",width=340 ] quenched galaxies emphasizes the differences between on - red - sequence sources ( top , 19 sources ) and off - red - sequence galaxies ( bottom , 13 sources ) . stacked spectra and uvis data points for individual galaxies
are shown in orange while best fit @xcite models to the spectra are shown in black .
the models are exponentially declining sfhs with @xmath122 , solar metallicity , salpeter imf and ages of 3.0 and 1.4gyr for on- and off - red - sequence galaxies , respectively . in the top of each panel
we show the number of galaxies used in the coadded spectrum as function of wavelength .
flux densities normalized to median flux between 5500 and @xmath123.,title="fig:",width=340 ] in the top panel of figure [ fig : histoetg ] we show the distribution of redshifts for our quenched galaxy sample .
redshifts in our sample spread between 1.0 and @xmath124 , with a median of 1.5 .
the sharp drop of the distribution for @xmath125 is a direct consequence of our selection against these galaxies . as described in appendix [ app : systerr ] the redshift determination at @xmath125
is dominated by strong systematics , obviously affecting the derived spps .
we present the stellar mass distribution in the bottom panel of figure[fig : histoetg ] .
the stellar masses of quenched galaxies spread from @xmath126 ( our mass cut ) to 11.70 with a median 11.15 .
the sfhs of our quenched galaxies are consistent with short star - formation bursts , with all the galaxies having exponentially declining sfh with @xmath127 myr . in figure[fig :
histageav ] we present the luminosity weighted age distribution for galaxies at @xmath128 ( where this parameter could be constrained , see section[sec : systerr ] and appendix[app : systerr ] ) . at these redshifts
our sample consists of 32 quenched galaxies .
we observe a wide distribution of ages between 1 and 4gyr with a median of 2gyr .
we compare these results with data of quenched galaxies at similar redshifts from literature .
the mean ages ( @xmath129 ) found by @xcite , @xcite and ( * ? ? ?
* @xmath130-@xmath131 ) are systematically above our median age , even though our age histogram shows a peak at 3 - 4gyr consistent with these works
. this may be due to different selections between our sample and those from literature .
in particular , being selected as extremely red objects ( eros , @xmath132 ) , the @xcite sample is probably biased towards the oldest galaxies at these redshifts .
the @xcite sample is based on wfc3 grism data as in our study ( though they only have the g141 grism ) .
however , their sample is mass - selected while we use mass _ and _ color to selected our quenched galaxies .
moreover , the reported ages in van dokkum et al .
are for galaxies significantly more massive than our sample ( @xmath133 ) .
finally , in order to maximize observing efficiency , the @xcite sample was constrained to galaxies inhabiting regions of high galaxy overdensity , where the star formation and quenching take place more quickly .
therefore , onodera et al .
sample would include mostly very old galaxies at the redshifts they were observed . compared to @xmath16 @xmath78 galaxy selection ( and alike ) , which separate sf and quenched galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ) , our combination of spectral fitting and @xmath42@xmath39@xmath46 color cut allows for blue ( younger ) quenched galaxies to be included in the sample ( see section below ) . in summary ,
our wide age distribution cover the different age regimes previously reported in literature .
although our median age ( @xmath134 ) is below previous reports , we have shown this does not imply an inconsistency between the different results but most likely it reflects different selection bias intrinsic to each galaxy sample . in figure[fig
: umr_mass ] we show rest - frame _ u_@xmath39_r _ color versus stellar mass for our sample of 41 quenched galaxies ( in red , asterisks for galaxies with emission lines ) together with local data from sdss - dr7 ( grey region , * ? ? ?
sdss colors for our galaxies were computed using the best - fit sed model and filter curves obtained from the sdss webpage .
when galaxies within our stellar mass range ( @xmath115 ) are considered , then the vast majority of sdss galaxies belong to the red - sequence .
the color distribution for these objects shows a narrow peak at ( @xmath135)@xmath136 . this is visible in the right panel of figure [ fig : umr_mass ] , where we show the ( _ u_@xmath39_r_)@xmath137 distribution of the local galaxy sample with @xmath115 ( grey histogram ) , together with the color distribution for our quenched galaxy sample .
our massive galaxies show comparatively a much broader color distribution than massive galaxies in the local universe , with ( @xmath135)@xmath137 colors as blue as 1.7 ( consistent with recent results , see e.g. , * ? ? ?
* ; * ? ? ?
we note that , because of the color preselection applied to our galaxy sample we are excluding even bluer , higher sfr galaxies .
therefore the number of blue massive galaxies we obtain can be considered as a lower limit , implying an even large spread in ( _ u_@xmath39_r_)@xmath137 colors at these redshifts compared to the local massive galaxy sample . in figure[fig :
umr_mass ] we find that our quenched - galaxy color distribution peaks at ( _ u_@xmath39_r_)@xmath138 , approximately 0.2 magnitudes bluer compared to the peak of the color distribution of local red - sequence galaxies .
this color difference matches the passive evolution of a short burst of star formation expected between @xmath108 ( the median redshift of our galaxy sample ) and @xmath139 . about 32% of our sample , however , has colors substantially bluer ( @xmath140 magnitudes ) than the local red - sequence , overlapping with the colors of local blue - cloud galaxies at masses below m@xmath141 m@xmath107 . in figure
[ fig : umr_mass ] we show the best - fit red - sequence to the @xmath142 sdss data passively evolved to @xmath108 ( 0.2mag offset , dashed line ) and with an additional offset so galaxies fall within the scatter of the red - sequence ( 0.22mag , continuous line ) .
the sdss red - sequence was computed by fitting a linear relation to the sdss log(m@xmath143 ) and ( _ u_@xmath39_r_)@xmath137 , while the width was taken as the 2@xmath43 dispersion with respect to this fit .
we use this line to split our sample in quenched galaxies _ on _ and _ off _ the @xmath108 red - sequence .
approximately @xmath144 of the sample is _ off _ the red - sequence . at the beginning of this section
we describe how our @xmath42@xmath39@xmath46 selection method produces a quenched galaxy sample with lower median age than previous works in literature .
we also describe how the selection effects of each of the cited works systematically bias their samples towards older ages . driven by this result we further investigate on the ages of our two sub - samples : quenched galaxies on and off the red - sequence . limiting the sample to quenched galaxies at @xmath128 ( where the age - sensitive 4000 / balmer break is covered by the spectroscopy ) ,
we find a significant age difference between the stellar populations of galaxies on and off the red - sequence .
the red - sequence quenched galaxies have a median age of 3.1gyr , while quenched galaxies off the red - sequence have a median age of only 1.5gyr .
we further investigate this age difference by stacking the spectra of the quenched galaxies on and off the red - sequence in order to increase our s / n .
figure[fig : coad ] shows the coadded spectra of quenched red - sequence ( top , 19 sources , orange ) and off - red - sequence galaxies ( bottom , 13 galaxies , orange ) , together with their best fit @xcite stellar population models ( black curves ) .
the stack data were fitted using our full model library , where only metallicity was fixed to solar .
the best fit models are exponentially declining sfhs with @xmath122 , solar metallicity , salpeter imf and ages of 3.0 and 1.4gyr for on- and off - red - sequence galaxies , respectively .
as we see these models provide a good representation of the stacked data .
also they confirming our results from the two galaxy sub - samples based on individual - galaxy analysis .
these results are consistent with recent findings of @xcite in coadded spectra , even though their ages for blue ( @xmath145 ) and red quenched galaxies ( @xmath146 ) are younger than ours . the @xcite models we considered in section [ discussion : yando ] assume a sfh with a single massive star- formation events that passively declines with time . in this scenario , a possible interpretation of the distinction between the measured ages of galaxies on and off the red - sequence arises from the progress of the sample towards the red - sequence observed at different epochs of evolution for individual galaxies .
even though these models fit well the coadded spectra , these are not the only possible sfhs for these galaxies . the presence of small secondary star - bursts ( sb ) after quenching of the main stellar component have been observed at low and high redshifts ( e.g. , * ? ? ?
* and references therein ) . in this section ,
we consider the possibility that our quenched galaxies lie _ off _ the red - sequence due a relatively recent minor burst of star formation which rejuvenates the galaxies seds , making them appear relatively younger due to the production of luminous , high - mass o / b / a/ stars . to test this scenario we constructed a new set of model sfhs , which are defined by a primary star - formation event ( fixed at very high-@xmath26 ) followed by a smaller , secondary sb .
both , the main stellar component and the sb are modeled with exponentially declining sfrs , with solar metallicity , salpeter imf and no extinction .
the main component produces a total stellar mass of @xmath147 m@xmath107 with @xmath122 .
the age of this main component before the sb ranges between 0.5 - 4.25gyr in steps of 0.25gyr ( 16 model variants ) .
the secondary sb has intensities of 1 , 5 , 10 and 50% the total galaxy mass ( 4 model variants ) with @xmath148 .
we observe the resulting seds at the time of the sb peak and after 50 , 100 , 250 and 500myr of that event ( 5 model variants ) .
the combinations of these possible parameters make a library of 320 models to be fitted to the coadded data . in figure
[ fig : coadsb ] we present the best model fits for our coadded spectra of quenched galaxies on and off the red - sequence ( orange ) .
( orange ) .
_ top panel : _ coadded spectra and individual - galaxy photometry of quenched galaxies on the red - sequence together with the best model fit with two stellar components ( main+sb , in blue ) . _
middle panel : _ our two - stellar - component models can not reproduce a " rejuvenation sfh ( old main component + young sb ) consistent with the data .
we show the coadded data of quenched galaxies off the red - sequence together with the best model fit with two stellar components ( both components young , in blue ) .
_ bottom panel : _ off the red - sequence galaxy data together with three models : in purple , best model fit among those observed at the sb peak . in green ,
best model among those with ages of observation @xmath149 .
in red , best model found for galaxies on the red - sequence ( top panel ) , with the proper scaling factor to better fit the off red - sequence data.,title="fig:",width=340 ] ( orange ) . _
top panel : _ coadded spectra and individual - galaxy photometry of quenched galaxies on the red - sequence together with the best model fit with two stellar components ( main+sb , in blue ) .
_ middle panel : _ our two - stellar - component models can not reproduce a " rejuvenation sfh ( old main component + young sb ) consistent with the data .
we show the coadded data of quenched galaxies off the red - sequence together with the best model fit with two stellar components ( both components young , in blue ) . _
bottom panel : _ off the red - sequence galaxy data together with three models : in purple , best model fit among those observed at the sb peak . in green , best model among those with ages of observation @xmath149 . in red ,
best model found for galaxies on the red - sequence ( top panel ) , with the proper scaling factor to better fit the off red - sequence data.,title="fig:",width=340 ] ( orange ) . _
top panel : _ coadded spectra and individual - galaxy photometry of quenched galaxies on the red - sequence together with the best model fit with two stellar components ( main+sb , in blue ) .
_ middle panel : _ our two - stellar - component models can not reproduce a " rejuvenation sfh ( old main component + young sb ) consistent with the data .
we show the coadded data of quenched galaxies off the red - sequence together with the best model fit with two stellar components ( both components young , in blue ) . _
bottom panel : _ off the red - sequence galaxy data together with three models : in purple , best model fit among those observed at the sb peak . in green , best model among those with ages of observation @xmath149 .
in red , best model found for galaxies on the red - sequence ( top panel ) , with the proper scaling factor to better fit the off red - sequence data.,title="fig:",width=340 ] just as a point of reference , in the top panel of figure [ fig : coadsb ] we see ( in blue ) the best model fit for quenched galaxies _ on _ the red - sequence that includes a secondary sb .
this model has an age of 4.5gyr .
the main stellar component is 4.0 gyr old before the sb , which burst mass is 1% .
the coarse shape of the coadded spectra is fitted by this model , showing however , a larger @xmath72 value than the single - burst best model ( top panel figure [ fig : coad ] ) . the predicted 4.5gyr is older than the 3gyr from the single - burst model .
this is not surprising as the galaxy must passively evolve after the sb to recover the originally red sed .
in the central panel of figure [ fig : coadsb ] we present the best model fit to the coadded spectra of quenched galaxies _ off _ the red - sequence .
this model has an age of 1.5gyr .
the main stellar - component age before the sb is 1.0 gyr , with a burst mass of 1% .
the predicted age of 1.5gyr is consistent with the single - burst model prediction for these galaxies .
we highlight that the best model for the quenched off red - sequence galaxy data is not a combination of an old , passive stellar population with a younger sb .
the fitting process demands both stellar populations to be similarly young , resembling a single - component young burst , i.e. , the `` rejuvenation '' scenario is not supported by this analysis . given the above result , we proceed by being more selective on choosing models that resemble a `` rejuvenation '' sfh , and test if they fit the data of off red - sequence galaxies . in the bottom panel of figure [ fig : coadsb ]
we show three examples .
in purple we plot the best model fit to the coadded spectrum for all models observed at the peak of the sb .
the main stellar component is , again , young ( 1.5gyr ) .
more important , even the weakest sb we are testing ( 1% ) can not reproduce the data and clearly under - predicts the flux in the red extreme of the coadded spectrum .
also , the blue flux is over - predicted as it can be seen by comparing the best fit models to the individual - galaxy uvis data points ( not included in fit ) .
only one upper - limit among 26 data points is consistent with this blue sed model .
in addition , we also considered only those models with old ( @xmath150 ) primary stellar component .
the resulting best - fit model is shown in green in figure [ fig : coadsb ] . similarly to the previous library , also this best - fit model over - predicts the flux at these blue wavelengths , and it is consistent with only four out of 26 observed uvis data points .
finally , we overplot in red the best fit model found for galaxies _ on _ the red - sequence , with the proper minimization - scaling factor to fit the _ off _ red - sequence data .
this model is reasonably compatible with the individual - galaxy uvis data ( as before , not included in fit ) , but it clearly over - predicts the flux in the red extreme of the coadded spectrum . in summary , after testing a library of models that include a secondary sb , we have found that they can not reproduce the `` rejuvenation '' scenario for most of our _ off _ red - sequence galaxies . the better fits are produced with overall young stellar populations ( for both , main component and sb ) , resembling a single - young stellar component .
the use of @xmath149 main stellar components systematically over / under predict the data flux at different wavelengths .
we remind the reader , however , we have not consider extinction in the model library used in this section . as we have seen before , extinction is a poorly constraint parameter that also introduces strong degeneracies with other spps .
therefore , the results shown in this section must be taken with caution , and restricted to the model library we use here .
we also investigate the fraction of quenched galaxies which have already settled on the red - sequence ( @xmath151 ) and those which have not , at the average redshift of our sample ( @xmath152 ) . using the red - sequence computed above
, we found that 28/41 galaxies fall within 2@xmath43 of the red - sequence .
these galaxies comprise 68% of our whole sample .
since all blue sf galaxies are excluded from our selection , we infer that less than @xmath153 of all galaxies at @xmath108 with @xmath115 fall on the red - sequence .
we stress that this is an upper limit .
while our @xmath42@xmath39@xmath46 color pre - selection allows us to select all of the red galaxies at @xmath154 , it systematically excludes blue galaxies at the same redshifts .
we can improve the upper limit on @xmath151 by including emission - line - selected sources in the same apparent magnitude , redshift and mass ranges of our sample .
for this , we use the @xcite sample of emission - line galaxies from 17 wisp fields ( all fields in common with our sample ) .
these galaxies are selected only by the presence of the h@xmath9 emission line .
dominguez et al .
limit their search to @xmath155 , thus probing a smaller volume compared to our study . as a result
the value of @xmath151 provided below has still to be considered as an upper limit .
we applied the same @xmath46-band magnitude , mass and redshift selection criteria used in our sample galaxies , to identify sources within our parameter ranges .
this selection results in 15 galaxies which are not in common with our galaxy sample .
stellar masses from dominguez et al . were scaled by a factor of 1.7 to account for the different imfs used ( chabrier in their case , salpeter in ours ) . by scaling the number of emission - line sources to our number of fields ,
we derive an expected number of 24 blue massive emission - line galaxies .
this implies that , at @xmath152 , @xmath156 .
recently , @xcite predicted the evolution of the red fraction as a function of galaxy mass and redshift , based uniquely on empirically - motivated relations .
peng et al .
predict that at @xmath108 , at the median mass of our galaxy sample , between 70 to 90% of the galaxies should have colors consistent with being on the red - sequence .
these fractions were derived from their predictions for galaxies in low - density environments .
we also note that , even though the typical mass of our sample is well within what peng et al .
identify as the `` mass quenching '' regime , ( i.e. , quenching depends only on intrinsic properties of the galaxy and _ not _ by the local environment it inhabits ) our color - mass relation in figure[fig : umr_mass ] does not show such a mass dependency . in our data ,
whether a quenched galaxy resides on or off the red - sequence appears to be independent of the galaxy s mass . on the contrary
, a mass - dependent quenching mechanism would demand a rising left - to - right gradient in the color - mass diagram , such as the most massive galaxies mostly populate the red - sequence while less massive systems should be bluer .
this is not observed in our galaxy sample , despite covering one order of magnitude in mass .
therefore , it seems reasonable to conclude that both internal ( e.g. , mass quenching ) and external ( e.g. , environment ) mechanisms are at play in quenching the star formation in some of these massive galaxies .
we speculate that the young massive galaxies with no on - going star formation observed at @xmath108 may be in transition between a phase of vigorous star - formation at @xmath157 and the @xmath152 red - sequence .
the large masses and short sf - timescales that characterize these galaxies imply very high past sfrs ( of the order of @xmath158m@xmath107 yr@xmath159 ) .
the survey volume covered in the @xmath160 redshift range is a few @xmath161mpc@xmath162 , implying a volume density for our quenched galaxies of @xmath163mpc@xmath164 . even considering that we had to remove from our sample about half of the galaxies because of spectral contamination
, this volume density is still comfortably below the observed space densities of some of the likely progenitors of these sources , i.e. , ultra luminous ir galaxies at @xmath165 ( 6@xmath166mpc@xmath164 * ? ? ?
+ + finally , we would like to highlight the results presented in this paper do not change strongly because of a misclassification of emission - line galaxies as `` quenched '' . if we remove the 4 quenched galaxies with emission from the sample we find the following results .
first , for quenched galaxies on and off the red - sequence the median ages ( @xmath167 and @xmath168 ) and single - component ages from coadded spectra ( @xmath169 and @xmath168 ) are virtually unchanged with respect to our complete sample estimations .
second , we find that 72% ( 68% previously ) of the sample galaxies are on the red - sequence . also the @xmath156 does not change with respect the upper limit found previously . in consequence
, we can be confident that neither the results not the global picture presented in this work are affected by sf galaxies misclassified as quenched systems .
we have used the wisp survey to identify a sample of massive ( @xmath115 ) galaxies at redshift @xmath170 .
the sample was selected to have @xmath42@xmath39@xmath46@xmath65 and @xmath46@xmath171 , in 27 independent fields , overcoming the effect of cosmic variance , which typically plagues massive galaxy samples at these redshifts .
our color selection implies that the final sample is biased against star - forming galaxies .
we derived stellar population parameters by fitting stellar population models to the combined broad band photometry ( hst / wfc3-uvis , hst / wfc3-ir , _
spitzer_/irac ) and ir spectra ( @xmath172m@xmath173 m ) .
we have shown that the availability of rest - frame optical spectra covering the 4000 /balmer breaks dramatically improves the reliability and accuracy of the derived stellar population parameters .
our results are based on a color- _ and _ mass - selected sample 41 quenched galaxies . in agreement with other studies , we find that , at @xmath174 , the mass range above @xmath126 is populated by galaxies with a wide range of stellar population properties .
we find that quenched galaxies are well fitted with exponentially decreasing sfhs , and short star - formation timescales ( @xmath3 ) .
they also show a wide distribution in stellar ages , between 1 - 4 gyr .
we find that quenched galaxies are far from being an homogeneous population . in the @xmath4-versus - mass space ,
quenched galaxies have a large spread in rest - frame color at a given mass .
most quenched galaxies populate the @xmath175 `` red - sequence '' , although 32% of them have substantially bluer colors .
we find that quenched galaxies _ on _ the red - sequence have older median ages ( 3.1gyr ) than the quenched galaxies _ off _ the red - sequence ( 1.5gyr ) .
the average ages of the two subsamples ( on and off the red - sequence ) are confirmed by the analysis of their stacked spectra .
furthermore , we also demonstrated that a `` rejuvenated '' sed can not reproduce the observed stacked spectra .
we derive the upper limit on the fraction of galaxies on the red - sequence at @xmath152 to be @xmath156 , in disagreement with empirical model predictions from peng et al .
( 2010 , @xmath17670 - 90% ) .
this mismatch can partially be due to the peng et al. assumption of an instantaneous quenching mechanism .
however , the homogeneous spread in mass of our quenched galaxies on and off the red - sequence suggests that more than one mechanism is responsible of the quenching at these stellar masses ( i.e. , internal , galaxy - mass dependant vs. external , environmental triggers ) .
we speculate that the young massive galaxies with no on - going star formation are in a transition phase between vigorous star formation at @xmath6 and the @xmath2 red - sequence . according to their estimated ages , the time required for quenched galaxies _ off _ the red - sequence to join their counterparts _ on _ the @xmath152 red - sequence is of the order of @xmath7 .
the open question remains concerning what
_ mechanisms _ halts the star formation . having been quenched more recently , the galaxies _ off _ the red - sequence with no on - going star formation will be the ideal laboratory to further investigate this process .
once the wisp survey is completed , we will be able to study this and other galaxy populations in greater detail , with a much larger number statistics .
the wisp survey is supported by grants hst - go-12283 and hst - go-12568 awarded by the space telescope science institute .
clm acknowledges support from nsf grant # ast-1109288 .
wfc3/uvis ccds experience a degradation of their charge transfer efficiency ( cte ) over time , introduced by their exposure to energetic radiation . the cte degradation and fractional associated losses worsen for low sky background , faint fluxes , and distance from the readout amplifiers .
these effects are discussed in @xcite for observations obtained between october 2009 and october 2011 .
correction for cte losses is currently not implemented in the reduction packages for the wfc3-uvis data , so we corrected the measured fluxes as follows .
@xcite provides linear fits of the cte losses as a function of background electron counts , position on the detector , and source flux .
we interpolated the linear fits to compute , for each galaxy , a unique correction as a function of the source counts and position on the ccd . for sources fainter than 500 e@xmath177
the correction was obtained by extrapolating the solution from higher counts ; while no correction was applied for sources brighter than 16,000 electrons . as the source detection for our photometric catalogue
was performed in the deeper ( and brighter ) @xmath178 images , many of the wfc3/uvis fluxes are below the extrapolation limit of 500 e@xmath177 ( @xmath179 of the catalogue sources ) .
typically , these sources have median magnitudes between 25.0 - 25.5 in the different uvis bands used in this study .
most of these sources ( @xmath180 ) , however , are also below our @xmath181 detection limit , so upper limits for their magnitudes ( independent of the cte correction ) are used instead ( see section[sec : photcat ] for details ) .
therefore , in practice most of the sources with @xmath181 are above the extrapolation limit for the cte correction . in figure
[ fig : cte_corr ] , we show an example of the cte correction in one of our fields versus the original source magnitude ( left ) and versus y - axis distance with respect the amplifiers ( right ) . most of our final galaxy sample ( see section [ sec : selectetg ] ) have corrections below 0.1mag , typically corresponding to @xmath182 in their total fluxes . their total fluxes ) ; _ right : _ in the cte correction versus distance to the amplifiers [ pixel ] plot we see that the correction is usually larger as the distance to the amplifier increases.,title="fig:",width=321 ] their total fluxes ) ; _ right : _ in the cte correction versus distance to the amplifiers [ pixel ] plot we see that the correction is usually larger as the distance to the amplifier increases.,title="fig:",width=321 ]
one of axe s features is the two - dimensional ( 2d ) modeling of the source s spatial profile for each galaxy . in what follows
we refer to the target galaxy as the primary , and to other galaxies which may contaminate the spectrum as secondaries . in figure [ fig:2dspeclean ]
we show a typical example .
the second panel of figure [ fig:2dspeclean ] shows the 2d dispersed stamp extracted by axe .
the stamp is centered on the spectrum of the primary ( circle in the top panel ) , and the spectra of two secondary objects are visible below it . within axe ,
source profiles for each galaxy are fitted at each wavelength assuming a _ single _ gaussian profile .
although the single gaussian is a sufficient approximation for faint , poorly resolved sources , this assumption breaks down for brighter resolved galaxies ( such as those in the example shown in figure [ fig:2dspeclean ] ) . resolved , brighter sources typically present extended wings , which are poorly modeled by a single gaussian profile . after testing different functional forms to describe the extended wings , we converged on describing each galaxy s profile with a combination of two gaussians sharing the same center , but with different amplitudes and width .
we fit a profile of the form : @xmath183 where @xmath184 and @xmath185 are the amplitudes of the two gaussian components , @xmath186 and @xmath187 are the width , and @xmath188 is the coordinate of the center in the crossdispersion direction .
this function provides the best compromise between functional flexibility and number of parameters to adjust . as it is clear from figure [ fig:2dspeclean ] , the contamination from secondary sources change with wavelength ( e.g. , due to relative position of sources in the sky , spectral features like absorption breaks and emission lines , variation of the grism transmission ) . for this reason
, we fit a combination of @xmath189 functions ( one for the primary and one for ach secondary ) to the spatial profile extracted from the 2d dispersed stamps after a three pixel binning in the wavelength direction .
note we are only binning the data to perform the fit .
the spectral resolution for the rest of the analysis was not degraded the fit is performed by minimizing the @xmath72 between the models and the observed profile at each wavelength position . the first guess for the parameters ( @xmath190 )
are estimated from the direct f110w images -for the g102 spectra- and f160w images -for the g141 spectra- .
contrary to their appearance in the dispersed images , galaxies are typically well separated in the sky , allowing us to obtain accurate relative positions and widths for all components entering the fit . in figure [ fig:2dspeclean ]
we show an example of this procedure .
the second , third and fourth pannels show an original ( axe output ) 2d dispersed stamp , a version clean from secondary contaminants using the axe model and a version cleaned using our procedure , respectively . in the top panel of figure [ fig : specleanprofile ]
we show the mean spatial profile of figure [ fig:2dspeclean ] example target ( black curve ) .
we show our best fit component to the primary ( blue curve ) and the secondary ( green curve ) profiles , and also include the best fit axe estimation of the primary profile ( dashed orange curve ) . in the inset panel , we show axe s and our profiles with a common ( arbitrary ) normalization . because in the optimal spectral extraction the profile is used to weight each pixel in the final spectrum , the axe profile assigns relatively more weight to the lower @xmath191 wings with respect to the peak flux .
this effect , together with a poorer axe removal of contaminants , may translate into an overestimate of the total flux in the 1d extracted spectrum , and in a lower @xmath191 . in figure
[ fig : g102comp ] we compare the integrated flux in the wavelength range 0.81.1@xmath37 m computed in axe s and our extraction . in the comparison , we consider only sources brighter than @xmath42@xmath52 .
although there is a broad agreement between the 2 measurements , the residuals ( shown in the bottom panel ) show that axe s fluxes are systematically overestimated from about 10% for bright objects up to @xmath192% for some faint sources .
the flux overestimate is due to different effects for bright and faint sources .
in fact , what dominates in the bright sources is the poor fit of the single gaussian profile , while at the fainter level the dominant contribution is due to the poor removal of nearby sources . ) .
axe systematically overestimates the fluxes between 10 - 50% from bright to faint galaxies .
blue symbols correspond to a @xmath42 @xmath19323.0 selected sample from wisp fields .
black symbols correspond to the 41 galaxies used in this study ( see sec.[results : samplespp ] ) .
red continuous line is the 1:1 relation .
red dashed lines are the @xmath194 deviations from the 1:1 trend.,width=321 ]
the contamination from foreground stars was removed using a magnitude - size diagram . the @xmath178 measurements and total - light radius estimations from source extractor software ( @xcite ) were used to separate galaxies from stars . in figure[fig : starrm ] we see how stars form a well defined trend .
all these sources with magnitudes @xmath195 were flagged as stars ( similar magnitude limit than our final galaxy sample ) . some residual contamination from stars
could be expected among the reminning sources ( galaxies ) and vice versa .
we pay special attention to this while selecting our final galaxy sample ( see section[sec : selectetg ] ) .
a total of 234 stars were flagged in this way .
another 71 stars ( out of the demarcated area in figure[fig : starrm ] ) were identified individualy while checking individual galaxy spectra and photometry .
therefore , a total of 305 stars were flag and removed from our 27 fields .
diagram shows that contamination from stars ( sources within the dashed red area ) can be removed as they define a distictive sequence with respect to galaxies .
black symbols are all @xmath42 detections in photometry .
the dashed red lines define a star - selection area for sources brighter than @xmath42@xmath196.,width=321 ]
we have studied the effects of diluted emission lines in the continuum in order to establish if this has an important effect on the sed fitting and therefore in the retrieved spps . in all the sed fits presented in this paper , regions with detected emission lines
have been masked to include only flux from the continuum .
potentially problematic cases , however , come from undetected emission lines ( e.g. , low s / n , large fwhm ) that might artificially increase the continuum flux used in the sed fits .
we explored these cases by modeling passive seds with emission lines . with these simulations we attempted to model the worst possible scenarios for the detection of emission lines .
the models were made as follows : * we use the @xcite model of a typical quenched galaxy ( @xmath122 , @xmath197 , sfr=0 , solar metallicity , @xmath198 ) and include emission lines represented as gaussians .
we model the lines [ oiii]@xmath199 and h@xmath9 as those are the most common features detected in our data set . in a first step ,
a variety of ews ( between 35 -our detection limit- and 500 ) and intrinsic fwhms ( between 40 -the pixel size on g141- and 300 ) were explored for these lines . *
the spectra were redshifted to @xmath26=1.3 so the [ oiii ] and h@xmath9 lines are both in g141 , the poorer resolution grism .
then we degraded the spectra to g141 resolution .
we considered the larger angular size of a galaxy in our sample along the dispersion direction of the slitless spectroscopy ( 525 in g141 ) to mimic the worst case resolution where emission lines can be easily missed . * in these simulations we use a noise range characteristic for our galaxy sample .
the s / n per pixel typically ranges between 20 and 80 with a median of 40 ( corresponding to a s / n per of 0.8 in g141 ) * we produce a library of models with different combinations of s / n , line ew and fwhms . those cases where the emission lines are lost in the continuum were flagged .
the task was performed by both , using the automatic line finder of colbert et al .
( 2013 , submitted ) , and by eye . *
this resulted in a set of parameter ranges between which the emission lines are not detected ( being diluted in the continuum ) : ew : 35 - 50 ( above 50 the lines are always identified ) , fwhm : 50 - 270 s / n : 30 - 80 ( below 30 the lines can not be distinguish from the noise ) . in figure[fig :
emiexample ] , we show examples of 6 combinations of s / n , ew and fwhm that cover the parameter space where the lines are not detected ( 6 for h@xmath9 and 6 for [ oiii ] ) . in black ,
the pure continuum spectra while in color , spectra+emission line . .
bottom panels , [ oiii ] . in black ,
continuum spectra ; in colors , continuum+emission line.,width=377 ] for the 6 combinations of parameters we run 100 mc simulations with different random noise .
then we calculate the mean flux within 3-@xmath43 of the emission line in both , the pure continuum and continuum+emission spectra .
finally , we estimate the percentage increase in flux in the emission - line cases with respect their pure continuum spectra . in figure[fig : emiflxpercent ] we show those percentages versus s / n ( h@xmath9 in blue , [ oiii ] in red ) .
large symbols are the means while the small @xmath200 symbols are the 100 mc for each case . for all s / n ( h@xmath9 in blue , [ oiii ] in red ) .
small x symbols are 100 mc simulations for each case while large symbols are the mean values.,width=321 ] as we see the differences in flux are @xmath201 , which is below the @xmath202 flux uncertainty in the grism spectroscopy .
therefore , in those cases were emission lines are diluted in the continuum , the flux contribution from the line is lower than our 1-@xmath43 errors in the grism data .
this implies that as far as we mask the detected emission lines for our sed fits , there is no need to make specific modeling of emission lines for our sed fitting process and simulations .
in this appendix we present a study on the systematic effects of sed fitting in our data .
our aim is to constrain for which specific data sets ( e.g. , with uvis detections or upper limits , with or without irac ) and spp - ranges our predictions are reliable , also providing systematic errors for the different spps .
our study is based on simulating our data set at different redshifts with stellar population synthesis models to later recover the ( known ) redshifts and spps of these models using our customize idl @xmath72-minimization code ( section[sec : redshift ] and [ sec : sedfit ] ) .
we simulate our photometric and spectroscopic data by using @xcite models .
the complete single - burst model library considered to retrieve spps of our galaxy sample was used here ( see section[sec : sedfit ] for the different spp - ranges considered ) . at this point
we want to mention that assuming the sfhs of our sample galaxies have a single - burst component is a simplification .
real galaxies may ( and probably do ) have more complex sfhs , like those that include multiple star - bursts during their lives .
a proper exploration of these scenarios for the estimation of systematic errors , however , would substantially increase the number of models to be tested . among others , multiple combinations of different numbers of star - burst , occurring at different times , with different intensities and @xmath203-folding time - scales would be needed . in literature ,
such number of simulations are not usually ( if ever being ) attempted for the kind of analysis presented in this appendix , mainly because the immense amount of computational time is prohibitive .
therefore , we decided to limit our simulations to the original single - burst model library . as a consequence
we remind the reader our systematic error estimates should be consider as lower - limits .
the model seds were re - scaled to agree with typical data broad - band fluxes and corresponding stellar masses .
we multiplied all the seds for a given factor such that a typical galaxy in our sample ( exponentially declining sfh with @xmath122 , 2gyr old , no star - formation and no extinction ) would have a stellar mass of @xmath204 .
variations in the stellar mass between 10.6 and 11.8dex ( the range of our data ) do not produce any significant change in the results of our simulations .
therefore most of the results presented here correspond to an input model mass of @xmath205 ( a mild spread in input mass , typically of @xmath206 , occurs given the different combinations of model spps ) .
redshifts between 0.6 and 2.4 in steps of 0.2 were tested and the model seds were redshifted accordingly . using a wide redshift range for the models allowed us to explore the effects of redshift misidentification in the error budget and derived quantities . for @xmath207
none of the models pass our @xmath42@xmath39@xmath46@xmath65 galaxy selection criterion . at @xmath208
the contribution to the mass function integral is very small @xcite .
then we degraded the models to g102 and g141 spectral resolutions to obtain our simulated grism spectroscopy .
random noise was also added to these models in order to match our data .
a signal - to - noise per pixel @xmath209 was adequate to represent most of our galaxy spectra .
then the model spectra were binned in wavelength to reproduce the slitless - resolution dependency on the angular size of the source in the dispersion direction . for the angular size
we use the median luminosity - profile - fwhm of our sample galaxies in @xmath42-band , corresponding to @xmath210 in the wfc3 detector ( @xmath211 and @xmath212 in g102 and g141 , respectively ) .
in addition to simulating the grism spectra , we use the model seds to simulate broad - band photometry ( f475x , f600lp , f110w and f160w filters ) .
for the two wfc3/uvis bands we considered a typical data @xmath44 uncertainty of 0.15 mag .
their fluxes were randomized using a gaussian distribution with this width .
then we compared the model uvis magnitudes with typical @xmath45 upper limits from our galaxy master sample ( 25.4 and 24.9 mag in f475x and f600lp , respectively ) .
if a model uvis magnitude was brighter than the limit , it was considered as a detection with its corresponding uncertainty . if equal or below the limit , the uvis magnitude was assigned the @xmath45 upper limit for the band and considered as such for the @xmath72-minimization process .
we measured and randomized the irac @xmath213 m photometry in an analogous way with a characteristic error of 0.3 mag .
once we had our sed library redshifted and simulating real data conditions , all the models passed through the same selection criteria in color ( @xmath42@xmath39@xmath46@xmath65 ) and magnitude ( @xmath46@xmath214 ) than our galaxies .
then we used our customized idl code to recover redshifts and spps of those models that passed the selection . in
what follows , the original redshifts and spps from which the model seds were made are referred as `` input parameters '' ( or sub - index `` in '' , e.g. , @xmath215 , age@xmath216 ) while the redshifts and spps retrieved by our idl code are referred as `` recovered parameters '' ( or sub - index `` out '' , e.g. , @xmath217 , age@xmath218 ) . in order to make a reliable representation of our data set we need more than retrieving the galaxy spps and modeling the specific features of the photometric and spectroscopic data .
when observations cover a given redshift range , different selection effects are present like the different cosmic volumes and regions of the luminosity ( mass ) function sampled at different redshifts .
we took into account these effects in our systematic error calculations by assign a relative weight to each simulated sed as function of redshift ( @xmath215 ) .
the selection effect was introduced as a normalized multiplicative factor for a retrieved redshift or spp ( e.g. , @xmath217 , age@xmath218 ) .
the weights come from the redshift distribution of galaxies from candels @xcite after selecting sources with the same color - magnitude criteria used in our data .
a normalized histogram of the redshift distribution is shown in the left panel of figure[fig : z_out ] , together with a smoothed version used in our calculations .
( black ) . in red , the smoother function used for our calculations .
the weights are from the redshift distribution of candels @xcite after selecting galaxies with the same color - magnitude criteria used in our data .
systematic error in @xmath217 versus @xmath217 . weighted mean systematic offsets for nine bins ( 0.1 wide in @xmath217 ) are shown as black dots .
error bars represent the 68% percentiles . in the bottom
we show the color code for @xmath215 and in parentesis the number of models that pass our color - magnitude selection criteria .
the positions of individual data points have been slightly randomized for presentation purposes.,title="fig:",width=321 ] ( black ) . in red ,
the smoother function used for our calculations .
the weights are from the redshift distribution of candels @xcite after selecting galaxies with the same color - magnitude criteria used in our data .
systematic error in @xmath217 versus @xmath217 . weighted mean systematic offsets for nine bins ( 0.1 wide in @xmath217 ) are shown as black dots .
error bars represent the 68% percentiles . in the bottom
we show the color code for @xmath215 and in parentesis the number of models that pass our color - magnitude selection criteria .
the positions of individual data points have been slightly randomized for presentation purposes.,title="fig:",width=321 ] in the right panel of figure[fig : z_out ] we show the systematic errors in @xmath217 versus @xmath217 with median systematic offsets for nine redshift bins ( black dots ) and their 68% percentiles . in this plot
we clearly see that for @xmath219 the redshifts are completely misrecovered having in most cases values @xmath220 .
the fact we were not sampling the 4000 break with the spectroscopy at low-@xmath26 has a relevant systematic effect for @xmath221 . as a result of this redshift - recovering tests we decided to remove from our galaxy sample
all sources with recovered spectrophotometric redshifts @xmath222 as they are not reliable . for redshifts
@xmath223 the systematics are consistent with zero , within the uncertainties .
we present those uncertainties in table [ tab : systerr ] . as retrieving spectrophotometric redshifts
is the first step in characterizing our galaxy sample , any uncertainty in the systematic error of this parameter will have repercussions in the recovered spps .
for this reason the systematics for most of the spps shown in table [ tab : systerr ] are also presented at different redshift ranges . in figure
[ fig : inout_all ] we show comparisons between input and output age , sfh , extinction ( a@xmath224 ) , stellar mass and sfr for @xmath225 models .
the different colors in symbols and histograms represent different @xmath215 as described in the legend .
the continuous red lines indicate no difference between input and output values .
models and stellar mass @xmath226 . in panel
( a ) we show the percentage difference between input and output ages versus @xmath227 .
black dots with error bars represent weighted mean ( % ) systematic offsets and 68%-percentiles in bins of 1gyr in @xmath227 . the red dots and error bars only consider models with @xmath228 . in panel
( b ) we show @xmath229 vs @xmath230 . in ( c )
we present the difference between input and output extinctions versus @xmath231 .
dots with error bars represent weighted mean systematic offsets and 68%-percentiles in bins of 0.2mag . in panel
( d ) we show histograms for recovered stellar mass ( @xmath232 ) at different redshifts .
vertical dashed lines represent the median of each distribution .
the black dot with error bars represents the weighted mean ( % ) systematic offset in stellar mass . in ( e )
we present the difference between input and output sfrs versus @xmath233 .
dots with error bars represent weighted mean systematic offsets and 68%-percentiles in different bins .
the last dot with an arrow represent the systematic offset for @xmath234 .
the different colors in x symbols and histograms represent different @xmath215 as described in the leyend ( number of models per @xmath215 in parenthesis).,width=680 ] in panel ( a ) of figure [ fig : inout_all ] we show the percentage difference between input and output ages ( @xmath235 ) versus @xmath227 .
the median systematic offsets ( black dots ) clearly show large uncertainties for ages @xmath7 .
we notice , however , that the larger contribution to these offsets comes from low-@xmath215 were the 4000 break is not sample with the spectroscopy . on the other hand
, we found an important degeneracy between age and @xmath236 .
this degeneracy also contributes to the systematics in @xmath227-@xmath235 described before , and its strength strongly depends on how well can we constraint these spps at different redshifts .
the @xmath235 versus @xmath227 systematics become more evident at lower redshifts where we lack the spectroscopic constraint of the 4000 break . but even at larger redshifts we also observe some degree of degeneracy between young - underestimated ages and high - overestimated extinctions . taking all this in consideration we decided to proceed as follows : we only considered ages at @xmath228 as constrained enough in order to be discussed in this paper .
the extinctions and lower - redshift age estimations are too degenerate to make meaningful statements based on them . as for the redshifts , the systematics in @xmath227 are consistent with zero within the uncertainties .
we show these results in table [ tab : systerr ] . in panel ( d ) of figure[fig :
inout_all ] we show histograms for the difference ( in % ) between recovered and input stellar masses . in this particular case
we show models with @xmath237 as this is a representative mass for our galaxy sample and there are no strong variantions in the results within our galaxy mass range .
the stellar mass is the most robust spp retrieved with our data . between @xmath238
there are virtualy no systematic offsets within @xmath239-@xmath240 ( @xmath241 ) . at @xmath242 systematic offsets
become more relevant , varying with stellar mass from 30 , 20 and 10% at @xmath243 , 11.0 and 10.6 , respectively ( on average ) .
these systematics are shown in table [ tab : systerr ] .
for the remaining two spps ( sfr and sfh ) different combinations of data sets could determined if the recovered values are ( or not ) reliable . in panel ( b ) of figure[fig : inout_all ] we show output versus input sfhs .
from this figure we highlight two points .
first , for our entire redshift range we can distinguish between short ( @xmath89myr ) and long ( @xmath90gyr ) sfhs , although we can not separate sfhs with @xmath91myr and @xmath92myr .
second , short sfhs ( @xmath89myr ) are recovered independently of the data set used to derive redshifts and spps . for more extended sfhs ( @xmath244 ) , however , we found that we are able to recover similarly extended sfhs only for data sets with 2 uvis detections independently of the redshift .
overall , these simulations imply we were able to broadly discriminate between @xmath245 above and below @xmath246 . in panel ( e ) of figure[fig :
inout_all ] we present the difference between output and input sfrs versus @xmath233 . over the entire redshift range we can successfully recover @xmath247 with uncertainties @xmath248 . for @xmath249 , however , the uncertainties in the systematics increase steadily reaching @xmath250 at @xmath251 .
the different retrieved redshifts make no much of a difference on this trend , but only increase the scatter on the median systematic errors at high @xmath217 .
for this reason , in table [ tab : systerr ] we present the systematics in sfr independently of redshift .
the strong systematic over - prediction of the sfr for @xmath251 , however , does not have a significant impact in our paper .
first , we base the science on a quenched galaxy sample which always has the low - sfrs we can successfully retrieve .
second , we do not make claims based on the precise values of the sfrs .
we only use it through the _ specific _ sfr to identify quenched galaxies ( @xmath252 ) . as no further claims
are made based on ssfrs , the relevant test is to determine if an intrinsically quenched / sf galaxy can be recovered as such with our procedure , independently of the exact value of their ssfr .
stellar population properties like ssfr and sfh are closely linked .
for example , at the time of observation a galaxy with an extended sfh is more likely to show a higher ssfr than a short - burst galaxy .
also , our ability to constraint these properties is tightly related to the constraints we can provide with the data . of particular importance is the number of uvis detections available to constraint these spps .
detections in the rest - frame uv are more likely to occur in sf galaxies . for these reasons we decided to explore the success in recovering ssfrs and sfhs for different uvis data sets in our entire redshift range . in this exercise
we did not explore individual values of ssfr and sfh but _ ranges _ within which these parameters are reliable .
for the ssfr two regimes were defined as above / below @xmath253 ( our quenched / sf selection criterion ) . as
already mentioned in previous paragraphs , sfhs are defined as short - sfhs if @xmath254 , and extended if @xmath244 .
our results show that the reliability of the recovered sfhs and ssfrs depends on the constraints from the uvis data .
at all redshifts we reliably recovered the quenched and short - sfh models ( the relevant sources for this paper ) .
this was independent on the level of constraint from uvis data ( two upper limits , one and two detections ) . on the contrary , and
just to give the complete picture to the reader , sf and extended - sfh models were harder to constraint . only those models with 2 uvis detections provided reliably recovered sfhs and ssfrs .
+ in figure [ fig : inout_all_noirac ] we present the systematic uncertainties of all our spps and redshifts for sed fits on which no irac @xmath97 data was used ( referred as no - irac from now on ) .
a summary of the systematic errors for no - irac models is shown in the right column of table[tab : systerr ] .
we observe that @xmath255 of the model seds experience convergence issues during the @xmath72-minimization .
this percentage is fairly constant for our entire range of @xmath215 models . in a closer analysis of these cases we found that models with @xmath256 are frequent among the nonconverging cases , particularly at @xmath257
this does not imply , however , that a majority of high-@xmath258 models do not converge . during the selection of our galaxy sample ,
the cases of nonconvergence for no - irac galaxies were very few . in any case ,
given the number of no - irac galaxies in our final sample , a @xmath259 of failed convergence would imply that at the most we are loosing 1 galaxy for this reason .
overall , we conclude from this analysis that the irac @xmath97 data makes it eaier to constraint sources with high extinction . as a test , to avoid misinterpreting the slightly smaller number statistics of no - irac simulations with respect the irac counterpart , we remove from the later the non - convergent models of the former .
we found that the systematic errors and dispersions of the simulations with irac barely change after removing those models .
therefore we can confidently interpret differences among irac / no - irac models as the consequence of lacking the @xmath97 data .
for all the spps no major changes were observed in the median systematic errors by using or not irac data . for @xmath217 no changes in the systematics
was observed with respect the simulations with irac .
the systematics in stellar ages remain fairly stable , with a 10% increment in the systematic uncertainties for @xmath260 .
only a reduction from 75 to 65% appears for the younger ages at @xmath261 .
however , as the uncertainties for that age regime are very large , we decided to keep 75% of error in the systematics for the no - irac models for those age and redshift ranges .
the sfhs are still well recovered within the broad categories defined above ( exponentially declined with @xmath262 and @xmath263 ) .
the no - irac model runs only suggest a larger fraction of misidentified @xmath264 as @xmath265 .
also , without irac , less models are misclassified as exponentially - increasing sfhs .
the lack of irac data has limited effect on the median systematics of stellar mass .
it mostly widens the distributions shown on panel d of figure [ fig : inout_all_noirac ] were the stronger effects are observed at high-@xmath215 .
this implies that the stellar mass of a galaxy is probably better constraint by using irac , particularly at high retrieved redshift .
however , this does not translate into a larger _ average _ offset in the systematics on no - irac models : only at @xmath266 a mild increment of 5% in the systematic error is measured .
the @xmath267 suffer from minor variations in the uncertainty on the systematics for no - irac models .
as it might be expected at this point , the lack of irac data does not prevent us to constraint the ssfr as we did for the models with irac photometry .
+ but for sed fits without irac @xmath268 data .
analogously to figure [ fig : z_out ] ( right panel ) , in panel f we show the weighted mean systematic uncertainties in @xmath217 for sed fits without irac data .
the @xmath215-color code is the same as in figures [ fig : z_out ] and [ fig : inout_all].,width=680 ] in summary , our study of the systematic uncertainties in the sed fitting procedure allowed us to identify those specific data sets ( combinations of photometry and spectroscopy ) which did not provide enough constraint to reliably recover certain spps .
apart of a@xmath224 and ages from sources at @xmath26<1.3 , quenched - galaxy spps are fairly well constrained independently of the data set used . concerning the actual systematic uncertainties we retrieved , there is no significant bias in the derived stellar masses , ( s)sfr , sfh and ages ( the later at @xmath26>1.3 ) for galaxies with and without irac photometry .
our ability to constraint spps with our current data set is in part constrained by our uncertainty in redshift determination , for which the 4000 - break is the key ingredient .
a summary of the systematic errors for each discussed spp is presented in table[tab : systerr ] .
+ + some final remarks : we should mention that systematic uncertainties would increase / decrease in case of an imf variation within our galaxy sample .
the variation in the imf is a very debated topic where most of the recent developments are restricted to local galaxy data .
a recent paper of @xcite on local early - type galaxies suggests the existence of a range of imfs from heavier - salpeter to chabrier cases . in this work , however , and for the specific mass - range of our galaxy sample ( assuming @xmath115 galaxies have velocity dispersions @xmath269 ) , there is no correlation between imf and stellar mass .
galaxies in this mass - range spread from heavier - salpeter to chabrier like imfs ( see also * ? ? ?
@xcite present a theoretical study on the systematic effects of sed fitting in stellar mass estimations . for different stellar - population - synthesis models ( a bruzual & charlot version and maraston et al.2005 ) and imfs ( * ? ?
* and a non - universal imf ) these authors find no imf - dependency on mass estimations in our mass - range . for high masses the scatter in the systematics is large independently of model / imf used with dispersions ranging from 20 - 300% the value of the original input mass .
all these results may imply that imf variations are a galaxy - to - galaxy phenomenon , information we certainly lack .
therefore we chose to use a unique salpeter imf for our calculations as it seems to be representative of the galaxy population we are studying .
it also allowed us to directly compare our results with similar stellar population studies in literature .
abazajian , k. n. , adelman - mccarthy , j. k. , ag@xmath270eros , m. a. , et al .
2009 , apjs , 182 , 543 atek , h. , malkan , m. , mccarthy , p. , et al .
2010 , apj , 723 , 104 auger , m. w. , treu , t. , gavazzi , r. , et al .
2010 , apjl , 721 , l163 baldry , i. k. , glazebrook , k. , brinkmann , j. 2004 , apj , 600 , 681 balogh , m. l , morris , s. l. , yee , h. k. c. , carlberg , r. g. , & ellingson , e. 1999 , apj , 527 , 54 barro , g. , faber , s. m. , perez - gonzalez , p. g. , et al .
2013 , 765 , 104 bauer , a. e. , conselice , c. j. , p@xmath271rez - gonz@xmath272lez , p. g. , et al .
2011 , mnras , 417 , 289 bekki , k. 1998 , apj , 502l , 133 bell , e. f. , wolf , ch . ,
meisenheimer , k. , et al .
2004 , apj , 608 , 752 bell , e. f. , van der wel , a. , papovich , c. et al .
2012 , apj , 753 , 167 bertin , e. & arnouts , s. 1996 , a&a supl . , 117 , 393 brewer , b. j. , dutton , a. a. , treu , tommaso , et al .
2012 , mnras , 422 , 3574 bruzual , g. & charlot , s. 1993 , apj , 405 , 538 bundy , k. et al .
2009 , apj , 651 , 120 calzetti , d. , armus , l. , bohlin , r. c. , et al .
2000 , apj , 533 , 682 cameron , e. , carollo , c. m. , oesch , p. a. et al .
2011 , apj , 743 , 146 cappellari , m. , mcdermid , r. m. , alatalo , k. , et al .
2012 , nature , 484 , 485 carollo , c. m. , bschorr , t. j. , renzini , a. , et al .
2013 , submitted to apj ( arxiv1302.5115c ) cassata , p. , giavalisco , m. , guo , y. , et al .
2011 , apj , 743 , 96 cassata , p. , giavalisco , m. , williams , c. c. , et al .
2013 , submitted to apj ( arxiv:1303.2689 ) chabrier , g. 2003 , pasp , 115 , 763 chapman , s. c. , blain , a. w. , smail , i. , ivison , r. j. 2005 , apj , 622 , 772 cimatti , a. , cassata , p. , pozzetti , l. et al .
2008 , a&a , 482 , 21 cimatti , a. , nipoti , c. , & cassata , p. 2012 , mnras , 422l , 62 coleman , g.d . , wu , c .- c . , & weedman , d. w. 1980 , apjs , 43 , 393 conroy , c. , & van dokkum , p. 2012 , apj accepted ( arxiv:1205.6473 ) conselice , c. j. , bluck , a. f. l. , buitrago , f. et al .
2011 , mnras , 413 , 80 cowie , l. l. , songaila , a. , hu , e. m. , & cohen , j. g. 1996 , aj , 112 , 839 daddi , e. , cimatti , a. , renzini , a. , et al .
2004 , apj , 617 , 746 daddi , e. et al .
2005 , apj , 626 , 680 daddi , e. , dickinson , m. , morrison , g. , et al .
2007 , apj , 670 , 156 dekel , a. , sari , r. , ceverino , d. 2009 , apj , 703 , 785 de ravel , l. et al .
2009 , a&a , 498 , 379 di matteo , t. , springel , v. , & hernquist , l. 2005 , apj , 433 , 604 dom@xmath273nguez s@xmath272nchez , h. , pozzi , f. , gruppioni , c. , et al .
2011 , mnras , 417 , 900 dom@xmath273nguez , a. , siana b. , henry , a. l. , scarlata , c. , bedregal , a. g. , et al .
2012 , accepted in apj ( arxiv:1206.1867 ) drory , n. , feulner , g. , bender , r. et al .
2001 , mnras , 325 , 550 dutton , a. a. , treu , t. , brewer , b. j. , et al .
2012 , mnras accepted ( arxiv:1206.4310d ) erb , d. k. , steidel , ch .
c. , shapley , a. e. , et al .
2006 , apj , 644 , 813 faber , s. m. , willmer , c. n. a. , wolf , c. , et al .
2007 , apj , 665 , 265 ferreras , i. , pasquali , a. , malhotra , s. , et al .
2009 , apj , 706 , 158 ferreras , i. , pasquali , a. , khochfar , s. et al .
2012 , aj , 144 , 47 fontana , a. et al .
2006 , a&a , 459 , 745 fontana , a. , santini , p. , grazian , a. , et al .
2009 , a&a , 501 , 15 kauffmann , g. , et al .
2003 , mnras , 341 , 54 kaviraj , s. , rowlands , k. , alpaslan , m. , et al .
2013 , mnras in press , ( arxiv:1307.8127 ) kelson , d. d. , illingworth , g. d. , van dokkum , p. g. , & franx , m. 2000 , apj , 531 , 184 kennicutt , r. c. jr . 1998 , ara&a , 36 , 189 kennicutt , r. c. jr . 1983 , apj , 272 , 54 kinney , a. l. , calzetti , d. , bohlin , r. c. , et al .
1996 , apj , 467 , 38 kong , x. , daddi , e. , arimoto , n. , et al .
2006 , apj , 638 , 72 kriek , m. , van dokkum , p. g. , franx , m. , et al .
2006 , apj , 645 , 44 kriek , m. , van dokkum , p. g. , franx , m. , et al .
2006 , apj , 649 , 71 kriek , m. , van der wel , a. , van dokkum , p. g. et al .
2008 , apj , 682 , 896 kriek , m. , van dokkum , p. g. , franx , m. , illingworth , g. d. , & magee , d. k. 2009 , apj , 705 , 71 kriek , m. , van dokkum , p. g. , whitaker , k. e. , et al .
2011 , apj , 743 , 168 k@xmath270mmel , m. , walsh , j. r. , pirzkal , n. , kuntschner , h. , & pasquali , a. 2009 , pasp , 121 , 59 kuntschner , h. , lucey , j. r. , smith , r. j. , hudson , m. j. , davies , r. l. 2001 , mnras , 323 , 615 labb@xmath271 , i. , huang , j. , franx , m. , et al .
2005 , apj , 624 , 81 lin , l. , dickinson , m. , jian , h .- y . , et al .
2012 , apj , 756 , 71 longhetti , m. , saracco , p. , severgnini , p. et al .
2005 , mnras , 361 , 897 l@xmath274pez - sanjuan , c. , ballcels , m. , p@xmath271rez - gonz@xmath272lez , p. g. , barro g. , et al .
2010 , apj , 710 , 1170 madau , p. 1995 , apj , 441 , 18 maraston , c. 2005 , mnras , 362 , 799 marchesini , d. , van dokkum , p. g. , f@xmath275rster schreiber , n. m. et al .
2009 , apj , 701 , 1765 m@xmath272rmol - queralt@xmath274 , e. , trujillo , i. , p@xmath271rez - gonz@xmath272lez , p.g . ,
varela , j. , & barro , g. 2012 , mnras , 422 , 2187 mccracken , h. j. , capak , p. , salvato , m. , et al .
2010 , apj , 708 , 202 mclure , r. j. , pearce , h. j. , dunlop , j. s. , et al .
2013 , mnras , 428 , 1088 mehlert , d. , thomas , d. , saglia , r. p. , bender , r. , wegner g. 2003 , a&a , 407 , 423 mihos , j. ch . , & hernquist , l. 1994 , apj , 425l , 13 mitchell , p. d. , lacey , c. g. , baugh , c. m. , cole , s. 2013 , submitted to mnras ( arxiv:1303.7228 ) muzzin , a. , marchesini , d. , van dokkum , p. g. , labb@xmath271 , i. , kriek , m. , franx , m. 2009 , apj , 701 , 1839 nagy , s. r. , law , d. r. , shapley , a. e. , steidel , ch . c. 2011 , apj , 735 , 19 nelan , j. e. et al .
2005 , apj , 632 , 137 newman , a. b. , ellis , r. s. , treu , t. , bundy , k. 2010 , apj , 717l , 103 newman , a. b. , ellis , r. s. , bundy , k. , & treu , t. 2012 , apj , 746 , 162 noeske , k. g. , baggett , s. , bushouse , h. , et al .
2011 , wfc3 uvis cte and charge injection : june 2011 update for cycle 19 observers oke , j. b. & gunn , j. e. 1983 , apj , 266 , 713 oser , l. , ostriker , j. p. , naab , t. , johansson , p. h. , & burkert , a. 2010 , apj , 725 , 2312 onodera , m. , renzini , a. , carollo , m. et al .
2012 , apj , submitted ( arxiv:1206.1540 ) oser , l. , naab , t. , ostriker , j. p. , & johansson , p. h. 2012 , apj , 744 , 63 pannella , m. , carilli , c. l. , daddi , e. , et al .
2009 , apj , 698 , l116 peng , y. , lilly , s. j. , kovac , k. , et al .
2010 , apj , 721 , 193 peng , y. , lilly , s. j. , renzini , a. & carollo , m. 2012 , apj , 757 , 4 poggianti , b. m. , calvi , r. , bindoni , d. , et al .
2012 , proceedings of the iau s295 : `` the intriguing life of massive galaxies '' , editors d. thomas , a. pasquali & i. ferreras ( arxiv1212.3207p ) raichoor , a. , mei , s. , nakata , f. , et al .
2011 , apj , 732 , 12 reddy , n. a. , steidel , ch .
c. , fadda , d. , et al .
2006 , apj , 644 , 792 renzini , a. 2006 , ara&a , 44 , 141 renzini , a. 2009 , mnras , 398l , 58 rettura , a. , mei , s. , & stanford , s. a. 2011 , apj , 732 , 94 rodighiero , g. , cimatti , a. , gruppioni , c. , et al .
2010 , a&a , 518 , l25 rodighiero , g. , daddi , e. , baronchelli , i. , et al .
2011 , apj , 739l , 40 ryan , r. e. , mccarthy , p. j. , cohen , s. h. , et al .
2012 , apj , 749 , 53 salpeter , e. e. 1955 , apj , 121 , 161 s@xmath272nchez - bl@xmath272zquez , p. , gorgas , j. , cardiel , n. , gonz@xmath272lez , j. j. 2006 , a&a , 457 , 809 saracco , p. , longhetti , m. , & gargiulo , a. 2011 , mnras , 412 , 2707 scarlata , c. , carollo , c. m. , lilly , s. j. , et al .
2007 , apjs , 172 , 494 spiniello , c. , trager , s. c. , koopmans , l. v. e. , chen , y. 2012 , apjl , 753 , l32 spiniello , c. , trager , s. c. , koopmans , l. v. e. , conroy , ch .
2013 , submitted to apj ( arxiv:1305.2873 ) stanford , s. a. , eisenhardt , p. r. , dickinson , m. 1998 , apj , 492 , 461 thomas , d. , greggio , l. , bender , r. 1999 , mnras , 302 , 537 thomas , d. , maraston , c. , bender , r. , mendes de oliveira , c. 2005 , apj , 621 , 673 tortora , c. , romanowsky , a. j. , napolitano , n. r. 2012 , apj letters submitted trager , s. c. , worthey , g. , faber , s. m. , burstein , d. , gonz@xmath272lez , j. j. 1998 , apjs , 116 , 1 treu , t. , auger , m. w. , koopmans , l. v. e. , et al .
2010 , apj , 709 , 1195 trujillo , i. et al .
2006 , apj , 650 , 18 trujillo , i. , ferreras , i. , & de la rosa , i. g. 2011 , mnras , 415 , 3903 trump , j. r. , weiner , b. j. , scarlata , c. et al .
2011 , apj , 743 , 144 valentinuzzi , t. , poggianti , b. m. , saglia , r. p. , et al .
2010 , apj , 721l , 19 van dokkum , p. g. et al .
2008 , apj , 677 , l5 van dokkum , p. g. , & brammer , g. 2010 , apj , 718l , 73 van dokkum , p. g , whitaker , k. e. , brammer , g. , et al .
2010 , apj , 709 , 1018 van dokkum , p. g. , brammer , g. , fumagalli , m. et al .
2011 , apj , 743l , 15 whitaker , k. e. , kriek , m. , van dokkum , p. g. , et al . 2012a , apj , 745 , 179 whitaker , k. e. , van dokkum , p. g. , brammer , g. , & franx , m. 2012b , apj , 754 , 29 whitaker , k. e. , van dokkum , p. g. , brammer , g. , et al .
2013 , apj , 770 , 39 williams , r. j. , quadri , r. f. , & franx , m. 2011 , apj , 738l , 25 wuyts , s. , f@xmath275rster schreiber , n. m. , lutz , d. , et al .
2011 , apj , 738 , 106 | we combine hst g102 & g141 nir grism spectroscopy with hst / wfc3-uvis , hst / wfc3-ir and _ spitzer_/irac[@xmath0 photometry to assemble a sample of massive ( @xmath1 ) and quenched galaxies at @xmath2 .
our sample of 41 galaxies is the largest with g102+g141 nir spectroscopy for quenched sources at these redshifts .
in contrast to the local universe , @xmath2 quenched galaxies in the high - mass range have a wide range of stellar population properties .
we find their seds are well fitted with exponentially decreasing sfhs , and short star - formation time - scales ( @xmath3 ) . quenched
galaxies also show a wide distribution in ages , between 1 - 4gyr . in the @xmath4-versus - mass space
quenched galaxies have a large spread in rest - frame color at a given mass .
most quenched galaxies populate the @xmath2 red - sequence ( rs ) , but an important fraction of them ( 32% ) have substantially bluer colors .
although with a large spread , we find that the quenched galaxies _ on _ the rs have older median ages ( 3.1gyr ) than the quenched galaxies _ off _ the rs ( 1.5gyr ) . we also show that a rejuvenated sed can not reproduce the observed stacked spectra of ( the bluer ) quenched galaxies _ off _ the rs .
we derive the upper limit on the fraction of massive galaxies _ on _ the rs at @xmath2 to be @xmath5 .
we speculate that the young quenched galaxies _ off _ the rs are in a transition phase between vigorous star formation at @xmath6 and the @xmath2 rs . according to their estimated ages , the time required for quenched galaxies _ off _ the rs to join their counterparts _ on _ the @xmath2 rs is of the order of @xmath7 . |
precision measurements at high energy colliders in the recent past led to stringent tests of the standard model and important bounds on new physics .
measurements of jet rates and shape observables in @xmath1 annihilation are of particular importance , as they allow for a precise determination of the strong coupling constant @xmath2 . at hadron colliders ,
a cross section involving @xmath3 jets is proportional to @xmath4 at leading order , such that an accurate knowledge of @xmath2 will be important at the lhc , where many interesting processes contain jets in the final state .
a determination of @xmath2 from jet rates and shape observables is described e.g. in @xcite , where one can see that for lep measurements , the experimental uncertainty is smaller than the theoretical one , which is based on resummed next - to - leading order ( nlo ) calculations .
as the theoretical error is dominated by scale uncertainties , an nnlo calculation will considerably improve this situation .
a future international linear collider will allow for precision measurements at the per - mille level , which offer the possibility of a determination of @xmath2 with unprecedented accuracy , provided that theoretical predictions at nnlo are available .
the calculation of @xmath5jets at order @xmath6 requires the calculation of virtual two - loop corrections combined with a @xmath7 parton phase space , one - loop corrections combined with a @xmath8 parton phase space where one parton can become soft and/or collinear ( unresolved " ) , and the tree level matrix element squared for @xmath9 partons where up to two partons can become unresolved .
the unresolved particles lead to a complicated infrared singularity structure which manifests itself in @xmath10 poles upon phase space integration .
these singularities have to be subtracted and cancelled with the ones from the virtual contributions before a monte carlo program can be constructed . to achieve this task ,
two different approaches have been followed , one relying on the manual construction of an analytical subtraction scheme@xcite , the other one relying on sector decomposition@xcite .
the main features of the methods based on the explicit construction of a subtraction scheme are the following : the subtraction terms are integrated analytically over the unresolved phase space , such that the pole coefficients are obtained in analytic form .
this requires appropriate phase space factorisation and subtraction terms which are simple enough to be integrated analytically in @xmath11 dimensions .
this method naturally leads to a close to minimal number of subtraction terms , and allows insights into the infrared structure of qcd . in the sector decomposition approach
, the poles are isolated by an automated algebraic procedure acting in parameter space , and the pole coefficients are integrated numerically .
the advantages of this approach reside in the fact that the extraction of the infrared poles is done by the computer , and that the subtraction terms can be very complicated as they are integrated only numerically . on the other hand , the algorithm which isolates the poles increases the number of original functions , thus producing rather large expressions .
the application of sector decomposition to real radiation at nnlo first has been presented in @xcite , and the combination of the sector decomposition approach with a measurement function first has been proposed in @xcite .
a number of nnlo results based on this method have been obtained meanwhile @xcite .
its application to the double real radiation part of @xmath5jets at order @xmath6@xcite is particularly challenging due to the high number of massless particles in the final state , which leads to a very complicated infrared singularity structure .
the wide range of applicability of sector decomposition goes back to the fact that it acts in parameter space by a simple mechanism .
the parameters can be feynman parameters in the case of multi - loop integrals , or phase space integration variables , or a combination of both . in the following
, the working mechanism of sector decomposition will be outlined only briefly for the example of phase space integrals , details can be found in @xcite .
the phase space integral of a matrix element squared , combined with some measurement function @xmath12 defining a physical observable , typically contains overlapping " structures like @xmath13 where only the dependence on two of the integration variables is shown for pedagogical simplicity . in order to extract
the poles in @xmath10 , the singularities for @xmath14 need to be factorised . sector decomposition is a way to achieve the factorisation of this type of entangled singularities in an algorithmic way : first the integration region is split into sectors where the variables @xmath15 and @xmath16 are ordered by multiplying with unity in the form @xmath17 $ ] .
then the integration domain is remapped to the unit cube : after the substitutions @xmath18 in sector ( a ) and @xmath19 in sector ( b ) , one has @xmath20 where the singularities are now factorised . for more complicated functions , several iterations of this procedure may be necessary , but it is easily implemented into an automated subroutine . once all singularities are factored out , they can be subtracted using identities like @xmath21 where we recognise the form of plus distributions . the result
can subsequently be expanded in @xmath22 , such that a laurent series in @xmath23 is obtained , where the pole coefficients are sums of finite parameter integrals which can be evaluated numerically . for the numerical evaluation it has to be assured that no integrable singularities are crossed which spoil the numerical convergence .
in the double real radiation part of @xmath24 jets , singularities which are located in the interior of the integration region do indeed occur .
however , it is always possible to remap them to endpoint singularites by a convenient variable transformation .
after such a remapping , they are amenable to sector decomposition .
in order to calculate the double real radiation part of @xmath24jets at nnlo , sector decomposition is applied to extract the poles appearing in massless @xmath9 particle integrals . as a simple example , let us consider the 5-particle cut of the ladder graph shown in fig . 1 . ( 180,100 ) ( 25 , 40 ) ( 10 , 22)figure 1 .
the ladder graph sector decomposition leads to @xcite @xmath25\right.{\nonumber}\\ & & \left.+\frac{1}{{\epsilon}}\ , [ 5.5959 - 4.4978\ , \log{\left(\frac{q^2}{\mu^2}\right)}\right.\nonumber\\ & & + \left . 0.7498\,\log^2{\left(\frac{q^2}{\mu^2}\right ) } ] \,+\,\mbox{finite } \right\}\ ; , \label{t5}\end{aligned}\ ] ] where the numerical accuracy is better than 1% .
the correctness of the result can be checked by exploiting the fact that the sum over all cuts of a given ( uv renormalised ) diagram must be infrared finite .
this is shown diagrammatically in fig .
after uv renormalisation , we obtain the condition @xmath26 where @xmath27 denotes the diagram with @xmath28 cut lines .
the renormalisation constants @xmath29 ( in feynman gauge ) are given by@xcite @xmath30 the expressions in eq .
( [ cfin ] ) for @xmath31 combine to @xmath32\right.\nonumber\\ & & + \frac{1}{{\epsilon}}\,[5.61-\frac{9}{2}\log{\left(\frac{q^2}{\mu^2}\right ) } + \frac{3}{4}\log^2{\left(\frac{q^2}{\mu^2}\right)}]{\nonumber}\\ & & + \left . \,\mbox{finite}\right\}\;.\label{c234}\end{aligned}\ ] ] we can see that the poles in ( [ c234 ] ) are exactly cancelled by the 5-parton contribution ( [ t5 ] ) within the numerical precision . although the sector decomposition approach is considered to be a numerical method " , as the pole coefficients are only calculated numerically , the isolation of the poles is an algebraic procedure , leading to a set of finite functions for each pole coefficient as well as for the finite part .
these finite functions are written to a fortran program and evaluated numerically using monte carlo techniques . in order to obtain results which are differential in a certain physical observable
, any ( infrared safe ) measurement function can be included at the level of the final monte carlo program , which means that the subtractions and expansions in @xmath23 do _ not _ have to be redone each time a different observable is considered .
further , the measurement function does not have to be an analytic function , but can be a subroutine acting on the four - momenta of the final state particles , as it is typically the case for a jet clustering routine .
the reason why such a maximal flexibility is possible resides in the fact that the program described in @xcite has the architecture of a partonic event generator .
this requires to keep track of the mappings of the original phase space variables to the variables used in each sector , i.e. each endpoint of the iterated decompsition tree .
these mappings will be different for each sector ( cf . the arguments of the function @xmath12 in the first and second line of eq .
( [ remap ] ) ) , but keeping this information allows to express the energies and angles of the final state particles in terms of the sector variables and thus to reconstruct the fully differential information on the final state . in this procedure , function evaluations in sectors which do not pass the kinematical requirements imposed by the measurement function are unavoidable . having to deal with a large number of sectors
, this can lead to a serious drop in efficiency of the monte carlo program .
therefore , in order to construct a program which produces results within a reasonable time scale , it is crucial to keep the number of sectors low .
this can be achieved by ( a ) using optimised phase space parametrisations for each topology , ( b ) using information on physical limits in the decomposition algorithm . optimising the phase space parametrisations
means the following : the matrix element squared can be divided into a certain number of topologies , which are defined by a certain set of denominators ( which will be combinations of mandelstam invariants , see e.g. eq . ( [ ex ] ) ) . in a given phase space parametrisation , some of the invariants will naturally be in a factorised form , but such a form can not be achieved for _ all _ invariants simultaneously .
an optimised parametrisation is one where the number of non - factorising invariants in the denominator is kept minimal .
therefore one has to choose different parametrisations for each topology , resp . for each class of topologies
with the same factorisation properties . using information on physical limits exploits the fact that in most cases of physical relevance , the measurement function is such that it would prevent certain poles from arising at all if it were included in the @xmath23-expansion .
for example , poles associated with a 2-jet configuration , where 3 of the 5 final state particles become theoretically unresolved , will be killed by a 3-jet measurement function .
it would therefore be desirable to suppress the terms associated with such configurations already at the level of the @xmath23-expansion , in order to avoid an unnecessarily large number of terms associated with the isolation of these poles .
on the other hand , we would like to keep the flexibility to include any measurement function only at the stage of the final monte carlo program . focusing on the process @xmath24jets ,
this dilemma has been solved by including some preselection rules " in the @xmath23-expansion which reject configurations which will surely be 2-jet configurations . in the example shown here , this can be achieved by introducing a cut parameter @xmath33 which must be smaller than any possible experimental resolution parameter @xmath34 for the variable @xmath35 , as @xmath36 always corresponds to a 2-jet configuration . in this way , one can reduce the size of the expressions considerably without loosing the flexibility to specify different jet algorithms later . to illustrate the action of the jet function , 3 , 4 and
5jet rates using the jade algorithm@xcite are shown in fig .
[ fig3 ] , based on the toy matrix element built from the graphs shown in fig . [ fig2 ] . for the toy matrix element , height=200 ] as a more complex example , the double real radiation part of a non - planar topology ( see fig .
4 ) also has been calculated .
( 280,200 ) ( -10 , 126 ) ( -10 , 105)figure 4 .
( a ) non - planar topology ( 95 , 108 ) ( -10 , 92)(b ) 3 , 4 and 5jet rates at order @xmath37 ( 3 , 82 ) for the non - planar topology in this case , square - root terms in the denominator are unavoidable , which implies that the expressions produced by the sector decomposition will be larger due to the presence of more non - factorising denominators .
the virtual corrections to this topology have not yet been included .
note that the results shown in figs .
[ fig3 ] and 4 are unphysical , as they do not contain all contributions to form a gauge invariant quantity .
the purpose of these figures is merely to demonstrate the action of the jet function and thus the power of the method to produce differential results .
the cpu time is @xmath38 for a precision better than 1% .
note that all topologies can be calculated in parallel , such that the cpu time for the full double real radiation part will be of the order of the one needed for the most complicated topology . what remains to be done , besides the inclusion of the remaining topologies , is the combination with the one - loop plus single real and the two - loop virtual corrections . how to proceed efficiently by combining sector decomposition with analytical results is sketched in @xcite
the method based on sector decomposition to calculate @xmath0jets at nnlo has the advantage that the isolation of the infrared poles is done by an automated routine and that it does not require the analytical integration of subtraction terms .
however , the method produces large expressions , which is an issue for a process like @xmath0jets at nnlo where the matrix element to start with is already large .
it has been described how to limit the proliferation of terms in the course of sector decomposition , and differential results have been shown for a subpart of the process .
the inclusion of massive particles within this method is certainly more straightforward than in analytical approaches .
in fact , as the masses act as infrared regulators , the number of decompositions will be less and therefore the produced expressions should be smaller than in massless cases .
this work was supported by the swiss national science foundation ( snf ) under contract number 200020 - 109162 .
s. bethke , hep - ex/0606035 .
d. a. kosower , phys . rev . *
d67 * , 116003 ( 2003 ) .
s. weinzierl , jhep * 03 * , 062 ( 2003 ) and hep - ph/0606008 .
a. gehrmann - de ridder , t. gehrmann , and e. w. n. glover , nucl .
b691 * , 195 ( 2004 ) . w. b. kilgore , phys . rev . * d70 * , 031501 ( 2004 ) . s. frixione and m. grazzini , jhep * 06 * , 010 ( 2005 ) .
g. somogyi , z. trocsanyi , and v. del duca , jhep * 06 * , 024 ( 2005 ) .
a. gehrmann - de ridder , t. gehrmann , and e. w. n. glover , jhep * 09 * , 056 ( 2005 ) . k. hepp , commun . math
* 2 * , 301 ( 1966 ) .
m. roth and a. denner , nucl
b479 * , 495 ( 1996 ) .
t. binoth and g. heinrich , nucl . phys . *
b585 * , 741 ( 2000 ) .
g. heinrich , nucl .
proc . suppl . * 116 * , 368 ( 2003 ) .
a. gehrmann - de ridder , t. gehrmann , and g. heinrich , nucl . phys .
* b682 * , 265 ( 2004 ) . c. anastasiou , k. melnikov , and f. petriello , phys
. rev . * d69 * , 076010 ( 2004 ) .
t. binoth and g. heinrich , nucl .
b693 * , 134 ( 2004 ) . c. anastasiou , k. melnikov , and f. petriello , phys .
* 93 * , 032002 ( 2004 ) .
g. heinrich , nucl .
. suppl . *
135 * , 290 ( 2004 ) . c. anastasiou , k. melnikov , and f. petriello , nucl .
b724 * , 197 ( 2005 ) . c. anastasiou , k. melnikov , and f. petriello , hep - ph/0505069 .
k. melnikov and f. petriello , hep - ph/0603182 .
g. heinrich , hep - ph/0601062 .
jade collaboration , s. bethke et al .
, phys . lett . *
b213 * , 235 ( 1988 ) .
g. heinrich , nucl .
* 157 * ( 2006 ) 43 . | we report on the sector decomposition approach to the double real emission part of @xmath0jets at nnlo . |
theoretical metamagnetic models are systems that have both antiferromagnetic and ferromagnetic coupling interactions . at zero external field @xcite , they undergo a second - order phase transition at the nel temperature @xmath10 .
besides the usual nearest - neighbor staggered arrangements of the spins , the ordered phase can be , for instance , ferromagnetic planes ordered antiferromagnetically to each other , or even ferromagnetic chains ordered antiferromagnetically , where the latter phase is usually called a superantiferromagnetic phase . at non - zero field , applyied longitudinally to the direction of the magnetization , the nel transition temperature decreases as the magnitude of the external field increases and the line of second - order transition comes to an end at a finite temperature . beyond this point , the transition is first order and ends up , at zero temperature , at a finite field @xmath11 .
this phenomenon has been previously well described by landau @xcite and by griffiths @xcite , where one has the presence of a tricritical point ( the point joining first- and second - order transition lines ) . on the other hand , through the years , experimental realizations of magnetic materials exhibiting the above characteristics have been studied .
for example , chernyi and co - workers @xcite have considered the kinetics of a magnetization process in quasi - one - dimensional ising superantiferromagnet named as trimethylammonium cobalt chloride [ ( ch@xmath12)@xmath12nh]cocl@xmath132h@xmath14o , denoted by cotac , belonging to a wider series of organometallic compounds with general chemical formula [ ( ch@xmath12)@xmath12mx@xmath12]cocl@xmath132h@xmath14o with m = mn , co , ni , fe , and x = br or cl . however , one of the first studies known in the literature on the effects of a longitudinal magnetic field in superantiferromagnetic systems has been done on the ( c@xmath14h@xmath16nh@xmath12)@xmath14cucl@xmath17 compound @xcite .
it is worth mentioning that superfluid mixtures of the two helium isotopes @xmath18he and @xmath19he , and a class of anisotropic metamagnets such as fecl@xmath14 , febr@xmath14 and ni(no@xmath12)@xmath20h@xmath14o , although microscopically different at first sight , show in their thermodynamic behavior a striking similarity in that they exhibit a phase diagram in which a line of @xmath21type transition points ( second - order transition line ) ends up at a tricritical point ( see , for instance , reference @xcite ) .
based on these real experimentations , we will treat herein a quantum version of a metamagnet conveying not only the basic features of the relevant classical interactions above discussed , which leads to a tricritical behavior similar to the well known classical blume - capel @xcite and beg models @xcite , but also the inclusion of quantum fluctuations due to a transverse field , which will be in fact important in the low temperature regime .
moreover , there has been , in recent years , an extensive literature on competitive interaction models that present a superantiferromagnetic ( saf ) ordering , where , for instance , ferromagnetic chains are coupled in an antiferromagnetic way in two dimensions . according to these lines ,
the system to be treated herein corresponds to a two - dimensional spin-@xmath22 ising model with different exchange interactions along the two lattice directions , in the presence of transverse and longitudinal magnetic fields .
additional motivations to study the ising model is because it can be used to describe the critical behavior of a broader class of materials , including easy - axis magnets , binary alloys , simple liquids and their mixtures , polymer solutions , subnuclear matter , etc . @xcite . in the present case
, the corresponding model can be described by the following hamiltonian @xmath23 where @xmath24 is the @xmath25 pauli spin-@xmath22 operator component at site @xmath26 on a square lattice of @xmath27 sites , @xmath28 is the exchange coupling along the @xmath29 axis , the first and second sum are over nearest - neighbors along the @xmath8 and @xmath9 axis , respectively , @xmath0 is the longitudinal magnetic field , and @xmath30 is the transverse magnetic field .
the corresponding ordered state is composed by a superantiferromagnetic phase , and is characterized by a parallel spin orientation in the @xmath8 direction , and an antiparallel spin orientation in the @xmath9 direction , therefore exhibiting a kind of nel order between two sublattices of linear chains that can be denoted by @xmath31 and @xmath32 . the classical version of model ( [ 1 ] ) ,
i.e. @xmath33 , on anisotropic square lattices has been investigated by using a modified mean - field theory , in which the intrachain interaction is treated exactly and the interactions between chains are taken into account in a mean - field way ( linear chain approximation - lca ) @xcite .
several other approaches , such as usual mean - field approximation ( mfa ) @xcite , effective - field theory ( eft ) @xcite , mean - field renormalization group ( mfrg ) @xcite , effective - field renormalization group ( efrg ) @xcite , monte carlo simulations ( mc ) @xcite , and high - temperature series expansion ( se ) @xcite have also been applied to this classical model .
although these approaches agree with the overall phase diagram in two dimensions , lca @xcite and eft @xcite approaches produce a reentrant behavior in the first - order transition line .
the same reentrant behavior is obtained for the three - dimensional model by the eft with four spins @xcite . on the other hand , the quantum version of the model , given by the hamiltonian ( [ 1 ] ) ,
has not been ubiquitously treated in the literature as its classical counterpart .
nevertheless , phase diagrams and some thermodynamic properties have been obtained by using the eft @xcite , and a pair approximation for the free energy ( the latter case only for @xmath34 @xcite ) .
it has been noted that only second - order phase transitions are present in the computed phase diagrams . in the present paper , using the mfa based on a variational method for the free- energy , we investigate the phase diagram behavior of the spin-@xmath22 ising superantiferromagnet in the presence of both longitudinal and transverse external fields .
we would like to seek out not only the effect of the quantum fluctuations in the model but also the presence or not of the tricritical points and first - order transitions , since the eft has only given a second - order character .
the remaining of the paper is organized as follows . in the next section we outline the formalism and its application to the transverse ising superantiferromagnet in the presence of a longitudinal magnetic field ; in sec .
iii we discuss the results and present some final comments .
we will treat model ( [ 1 ] ) by the mean - field approximation ( mfa ) using a variational method based on peierls - bogoliubov inequality , which can be formally written as @xmath35 where @xmath36 and @xmath37 are free energies associated with two systems defined by the hamiltonians @xmath38 and @xmath39 , respectively , the thermal average @xmath40 should be taken in relation to the canonical distribution associated with the trial hamiltonian @xmath39 , with @xmath41 standing for the variational parameters .
the approximated free energy @xmath36 is then given by the minimum of @xmath42 with respect to @xmath41 , i.e. @xmath43 . the trial hamiltonian @xmath44 is chosen as free spins , distributed in two different sublattices @xmath31 and @xmath32 .
each sublattice consists of ferromagnetic linear chains coupled antiferromagnetically with two neighboring chains .
we then have @xmath45 where @xmath46 and @xmath47 are two variational parameters .
it is not difficult to compute the right hand side of eq .
( [ 2 ] ) and , after minimizing @xmath42 , the variational parameters @xmath46 and @xmath47 can be written as a function of the sublattice magnetizations @xmath48 and @xmath49 . the approximated mean - field helmholtz free energy per spin , @xmath50 ,
can thus be written as @xmath51with the corresponding sublattice magnetizations @xmath48 and @xmath49 given by @xmath52and @xmath53where @xmath54 , @xmath55 , @xmath56 and @xmath5 is the ratio between ferromagnetic and antiferromagnetic interactions . for a given value of the set of parameters @xmath57 , @xmath58 , @xmath59 and @xmath60 , eqs .
( [ 4 ] ) and ( [ 5 ] ) are numerically solved for @xmath48 and @xmath49 and , when several solutions are found , the stable phase will be the one that minimizes the free energy @xmath61 . in this context
, the equilibrium state corresponds always to the minimum value of @xmath61 with respect to @xmath48 and @xmath49 . for a metamagnetic system , it is more convenient to formulate the problem in terms of the total @xmath62 and the staggered @xmath63 magnetizations , which are defined as @xmath64 and @xmath65 . therefore , we can rewrite eqs .
( [ 3 ] ) , ( [ 4 ] ) and ( [ 5 ] ) as @xmath66^{2}+\delta^{2}}\right\rbrace \nonumber \\ & & + \left(1-\alpha\right)m^{2}+\left(1+\alpha\right)m_{s } , \label{6}\end{aligned}\]]@xmath67^{2}+\delta^{2 } } } \nonumber \\ & & \tanh\frac{1}{t}\sqrt{\left[h+2m\left(1-\alpha\right)-2\left(-1\right)^{p } \left(1 + \alpha\right)m_{s}\right]^{2}+\delta^{2 } } , \label{7}\end{aligned}\]]@xmath68^ { 2}+\delta^{2 } } } \nonumber \\ & & \tanh\frac{1}{t}\sqrt{\left[h+2m\left(1-\alpha\right)-2\left(-1\right)^{p } \left(1 + \alpha\right)m_{s}\right]^{2}+\delta^{2 } } , \label{8}\end{aligned}\]]from which one can obtain the frontiers separating the saf and p phases ( the paramagnetic phase p consists of all spins aligned with the external longitudinal field ) .
although the first - order transition line must be computed by numerically seeking the minimum of the free energy , the second - order transition line , as well as the location of the tricritical point , can be obtained through a landau expansion of the free energy , given by eq .
( [ 6 ] ) , in a power series of the order parameter @xmath69 . after a lengthy algebra
the final result can be written as @xmath70 from which we get the second - order transition lines when @xmath71 and @xmath72 , and the tricritical point when @xmath73 , @xmath74 and @xmath75 ( see , for instance , refs .
@xcite ) . in present case
, @xmath76 is not important for our purposes , so the @xmath77 , @xmath78 , and @xmath79 coefficients can be written as @xmath80-\frac{4\gamma_{1}^{2}}{\lambda}\left[\beta x + \frac{x}{\lambda}\right]\right\rbrace+2(1+\alpha ) , \label{a2}\ ] ] @xmath81\\ \nonumber & & + \frac{24}{\beta^{2}\lambda}\left(\frac{\beta\gamma_{1}\gamma_{2}}{\lambda}\right)^{4}\left[x^{3}-\frac{x}{(\gamma_{1}\gamma_{2})^{2}}-\beta x\right]+\frac{36}{\beta^{3}\gamma_{1}^{2}}\left(\frac{\beta\gamma_{1}\gamma_{2}}{\lambda}\right)^{4}\left[2\left(1-\frac{1}{\beta\lambda}\right)+\left(\frac{\gamma_{1}x}{\beta\lambda}\right)^{2}\right ] \\
\nonumber & & -\frac{60\beta\gamma_{1}^{4}\gamma_{2}^{4}}{\lambda^{6}}\left(1-\frac{x}{\lambda\gamma_{2}^{2}}\right ) , \label{a4 } \end{aligned}\]]@xmath82where @xmath83 , @xmath84 , @xmath85 and @xmath86 .
the functions @xmath87 , @xmath88 , are rather lengthy to be reproduced here .
depending on the value of the hamiltonian parameters , the model ( [ 1 ] ) has first- or second - order phase transitions from the superantiferromagnetic phase ( saf ) to the paramagnetic phase ( p ) . for @xmath33 ,
one has the well known mean - field results for the classical model , in which the critical temperature decreases from @xmath89 as the longitudinal field increases and , at @xmath90 ( ground - state ) , a first - order transition occurs at @xmath91 .
the tricritical point is located at @xmath92 and @xmath93 for the isotropic lattice @xmath94 .
it would then be quite interesting to see the global effects of the quantum fluctuations , which are driven by the transverse field , on the corresponding phase diagram . ) .
( a ) in the reduced temperature @xmath57 versus reduced longitudinal magnetic field @xmath58 plane , for several values of @xmath59 .
( b ) in the reduced temperature @xmath57 versus the reduced transverse field @xmath59 , for several values of @xmath58.,title="fig:",width=264,height=264 ] ) .
( a ) in the reduced temperature @xmath57 versus reduced longitudinal magnetic field @xmath58 plane , for several values of @xmath59 .
( b ) in the reduced temperature @xmath57 versus the reduced transverse field @xmath59 , for several values of @xmath58.,title="fig:",width=264,height=264 ] in figs .
[ mfa_diagrams_1](a ) and [ mfa_diagrams_1](b ) we have the phase diagrams in the @xmath95 ( for some selected values of @xmath59 ) and @xmath96 ( for some selected values of @xmath58 ) planes , in the isotropic lattice case @xmath94 . from fig .
[ mfa_diagrams_1](a ) we can see that as the transverse field increases , the transition temperature decreases .
this is a result of the quantum fluctuations destroying not only the superantiferromagnetic order among chains , but also the corresponding ferromagnetic order inside each chain .
the first - order line and the tricritical point survive up to @xmath97 ( this value , as well as the ones given below , can be numerically obtained with much more precision . however , for the sake of simplicity , we will only present them with two decimal digits ) . for @xmath98
the transition is always second order , and quantum phase transitions start to develop at @xmath90 , while for @xmath99 the system is always in the paramagnetic phase .
[ mfa_diagrams_1](b ) presents the corresponding phase diagram in the @xmath96 plane for several values of @xmath58 and again @xmath94 . for @xmath100
, we have the mean - field solution of the transverse ising model , where the critical temperature goes to zero at the known value @xmath101 .
as expected , by increasing the longitudinal field the transition temperature decreases . for @xmath102
the transition is always second order .
tricritical points and first - order transition lines , at low temperatures , appear for @xmath103 ( in this region , the critical quantum phase transition is suppressed ) . in the range @xmath104 , only first - order transition is observed and , for @xmath105 , the system is always in the paramagnetic phase for any value of the temperature . in figs .
[ mfa_diagrams_2](a ) and [ mfa_diagrams_2](b ) we have the phase diagrams for the anisotropic lattice @xmath106 .
the general behavior is the same , we have only changes in the values of the corresponding hamiltonian parameters . for instance , in the @xmath95 plane the tricritical point survives for @xmath107 , for @xmath108 only second - order transition lines are present and for @xmath109 the system is always in the paramagnetic phase . in the @xmath96 plane one has always second - order transitions for @xmath110 , tricritical points and first - order transition lines appear for @xmath111 , only first - order transitions occur for @xmath112 , and the paramagnetic phase is always stable for @xmath105 .
note here the smaller range of @xmath58 for having first - order and tricritical points in this plane phase diagram . for @xmath106.,title="fig:",width=264,height=264 ] for @xmath106.,title="fig:",width=264,height=264 ] the same trend is achieved for still smaller values of @xmath60 , where a crossover from the two - dimensional behavior to the one - dimensional behavior is observed when @xmath113 . for this quasi - one - dimensional model , however , the range of the parameters in order to observe multicritical phenomena gets narrower , and it is sometimes difficult to numerically access it .
moreover , as the present one - spin mean - field approach provides a final transition temperature even in one dimension , better procedures should be very welcome to treat the present model .
nevertheless , from the present results , one can clearly see that the results are indeed different from those obtained by employing eft @xcite and no reentrant behavior is observed in the first - order transition lines at low temperatures .
the same qualitative results are expected for the model in three - dimensions , where the chains are arranged in a staggered way .
in summary , we investigated the anisotropic two - dimensional nearest - neighbor ising model with competitive interations in an uniform longitudinal and traverse fields by using the mfa approach .
we obtained the phase diagrams in the @xmath114 and @xmath115 planes varying the value of @xmath60 , where the critical frontier separates the saf order with the paramagnetic disorder . at zero temperature , the critical field
is exactly obtained , so @xmath116 . for a @xmath117 ,
the model reproduces the classic result @xcite .
we see that there are the same trend is achieved for still smaller values of @xmath60 , where a crossover from the two - dimensional behavior to the one - dimensional behavior is observed when @xmath113 .
we showed also that for this approach appear in the phase diagrams lines of the first order as well as tri - critical points , and that the first order lines are derived from the construction of maxwell .
furthermore , the investigations of this three - dimensional model are expected to show many characteristic phenomena as , for example , the reentrant behavior .
this will be discussed in future works . | * abstract * magnetic phenomena of the superantiferromagnetic ising model in both uniform longitudinal ( @xmath0 ) and transverse ( @xmath1 ) magnetic fields are studied by employing a mean - field variational approach based on peierls - bogoliubov inequality for the free energy .
a single - spin cluster is used to get the approximate thermodynamic properties of the model .
the phase diagrams in the magnetic fields and temperature ( @xmath2 ) planes , namely , @xmath3 and @xmath4 , are analyzed on an anisotropic square lattice for some values of the ratio @xmath5 , where @xmath6 and @xmath7 are the exchange interactions along the @xmath8 and @xmath9 directions , respectively .
depending on the range of the hamiltonian parameters , one has only second - order transition lines , only first - order transition lines , or first- and second - order transition lines with the presence of tricritical points .
the corresponding phase diagrams show no reentrant behavior along the first - order transition lines at low temperatures .
these results are different from those obtained by using effective field theory with the same cluster size .
* pacs numbers * : 64.60.ak ; 64.60.fr ; 68.35.rh |
one of the essential results in heavy ion collisions was the observation that particle yields measured in a final state closely resemble a thermal equilibrium population @xcite .
the natural question was whether this statistical behavior is a unique feature of high energy nucleus - nucleus collisions or whether it is also applicable in elementary collisions like , e.g. , @xmath2 .
previous publications @xcite indicated that indeed hadron production in @xmath2 collisions can be well described within a thermal model provided that local quantum number conservation is properly implemented . in view of the most complete and extended data summarized e.g. by the particle data group ( pdg ) @xcite the above question has been recently addressed independently in @xcite and @xcite . in this contribution
we discuss hadron production in @xmath2 annihilations at @xmath3 gev based on the thermal model analysis of @xcite .
we also focus on different implementations of the statistical model and discuss the importance of quantum statistic effects and the mass cut in the hadron mass spectrum .
we quantify the production of heavy flavors and compare the model predictions with available data .
the usual form of the statistical model in the grand canonical ensemble formalism can not be used when the number of produced charged particles is small .
this is the case if either the temperature @xmath4 or the volume @xmath5 or both are small . as a rule of thumb one needs @xmath6 for a grand canonical description to be applicable @xcite . in @xmath0 annihilations where modelling the particle production within a thermal approach requires an exact formulation of conservation laws .
in such a system one needs to account for an exact conservation of five quantum numbers : the baryon number @xmath7 , strangeness @xmath8 , electric charge @xmath9 , charm @xmath10 and bottom @xmath11 .
the appropriate tool to deal in a statistical mechanics framework with a system of quantum numbers @xmath12 is the canonical partition function @xcite @xmath13 where @xmath14 and @xmath15 is a five component vector @xmath16 containing the quantum numbers of the particle species @xmath17 .
the quantity @xmath18 is an element of the symmetry group @xmath19 ^ 5 $ ] related with additive conservation laws . in this expression
, each @xmath20 corresponds to the conservation of the corresponding quantum number @xmath21 and @xmath22 is the single particle partition function for particle with mass @xmath23 , spin - isospin degeneracy factor @xmath24 , and a system with volume v and temperature t. the sum in eq .
( [ eq : partition1 ] ) runs over all particle species in the hadronic gas . the integral representation of the partition function in eq .
( [ eq : partition1 ] ) is not convenient for numerical analysis as the integrand is a strongly oscillating function .
thus , we first expand the logarithm @xmath25 and then , using the method described in @xcite , we express the partition function ( [ eq : partition1 ] ) in a series of bessel functions to obtain a result that is free from oscillations . furthermore , from eq .
( [ eq : partition1 ] ) we obtain the multiplicity @xmath26 for particle species j by introducing a fugacity parameter @xmath27 which multiplies the particle partition function @xmath22 and by differentiating @xmath28 the first term in the expansion of the logarithm in eq .
( [ eq : log ] ) corresponds to the boltzmann approximation which is suited only for @xmath29 .
such a condition is satisfied for baryons since their masses are larger than a typical temperature of the hadron resonance gas which never exceeds a critical value @xmath30 mev required for deconfinement .
however , for light bosons like pions or kaons the quantum statistics is of importance as the temperature is comparable to their masses .
( [ fig1]left ) shows relative deviations of pion and kaon multiplicities from their quantum statistics values with increasing numbers of terms @xmath31 in the expansion ( [ eq : log ] ) .
the calculations were performed for @xmath32 and @xmath33 @xmath34 .
it is clear that the boltzmann approximation is by far not sufficient to reproduce the quantum statistics results .
the pion yield under boltzmann approximation deviates by more than @xmath35 from the exact quantum statistics result . for kaons
this difference is only @xmath36 due to the larger mass . for pions ,
five terms are needed in the expansion ( [ eq : log ] ) to get quantum statistics value with deviations below @xmath37 . for heavier particles , for instance for protons ,
the quantum and boltzmann statistics differs by less than @xmath38 .
thus , in the model comparison with @xmath0 data we apply quantum statistics for light mesons and we use boltzmann statistics for all baryons .
collisions at @xmath39=91 gev .
the upper panel shows the fit in an uncorrelated jet scheme when including the feed - down contribution from heavy quarks .
the middle panel is after subtraction of this contribution .
the results in the lower panel are obtained in a correlated ( 5-flevour ) jet scheme fit where the flavor abundances are extra input parameters from data ( see text ) .
the best thermal model fit parameters are listed for each case.,scaledwidth=54.0% ]
the hadron multiplicity calculation within the statistical model basically proceeds in two steps .
first , a primary hadron yield @xmath40 , is calculated using ( [ eq : partition1 ] ) and ( [ eq : partition2 ] ) .
a crucial assumption of the model is that the final yields of all particles are fixed at a common temperature , the chemical decoupling point . as a second step ,
all resonances in the gas which are unstable against strong decays are allowed to decay into lighter stable hadrons , using appropriate branching ratios ( b ) and multiplicities ( m ) for the decay @xmath41 published by the pdg @xcite .
the abundances in the final state are thus determined by @xmath42 where the sum runs over all resonance species . from eq .
( [ res ] ) it is clear , that the final multiplicity of stable hadrons depends on the number of resonances used in the sum . in general
one should include contributions of all known resonances as listed by the pdg .
( [ fig1]-right ) shows the relative change of different particle yields when applying the mass cut @xmath43 and @xmath44 gev in eq .
( [ res ] ) . for mesons
this difference amounts to @xmath45 and is as large as @xmath46 for baryons .
thus , it is clear that restricting the mass spectrum only to the resonances with the mass @xmath47 gev might be not sufficient at the level af accuracy of data of a few percent as is the case in @xmath0 collisions .
however , our knowledge of decay properties of heavier resonances is by far not complete , which causes systematic uncertainties of the statistical model .
the importance of heavier resonances originating from the hagedorn mass spectrum in the analysis of particle production in heavy ion collisions has been recently analyzed in @xcite . in the following we account for resonance contributions up to the mass @xmath48 gev . in general , the resonance gas model formulated in the canonical ensemble
is described by only two basic thermal parameters : the temperature t and the volume v of the system . to explore a possible strangeness undersaturation
we introduce an additional parameter @xmath49 into the partition function to account for a possible deviation of strange particle yields from their chemical equilibrium values @xcite .
when applying the statistical model to particle production in @xmath0 annihilations we have to take into account that most hadronic events in such collisions are two - jet events , originating from quark - antiquark pairs of the five lightest flavors .
since we would like to address the issue of overall equilibration in these systems , we have to specify how the initial quantum numbers are distributed between the two - jets .
we will consider two scenarios : an _ uncorrelated _ and _ correlated _ jet scheme . in an _
uncorrelated _ jet scheme each jet is treated as a fireball with vanishing quantum numbers as fixed by the entrance channel .
it is clear at this point that hadrons from jets with heavy quarks ( c and b ) will be greatly underestimated by the model because of the large boltzmann suppression factors . in this approach
the issue of equilibration is effectively addressed only for hadrons with light quarks ( u , d , s ) .
it is important to recognize that the measured yields of these hadrons contain the contribution from the @xmath0 annihilation events into @xmath50 and @xmath51 .
heavy - quark production is indeed significant and is very precisely measured , in particular at the @xmath52 mass , where the measurements are very well described by the standard model .
hence , heavy - quark production is manifestly non - thermal in origin .
we therefore consider two cases : i ) we fit the data as measured and ii ) we subtract from the yields of hadrons carrying light quarks the contribution originating from charm and bottom decays based on available data for the charmed and bottom hadron production and their branching ratios . in a _ correlated _ jet scheme ( case ( iii ) in fig .
( [ fig2 ] ) ) the initial quantum numbers are distributed such that each jet carries quantum numbers of either @xmath53 , @xmath54 , @xmath55 , or vanishing quantum numbers in the case of @xmath56 and @xmath57 jets .
the fractions of the quark flavors in hadronic events @xcite are external input values , unrelated to the thermal model ( see also table ii in ref .
@xcite ) . for the fit procedure we use the complete set of
all measured hadron yields with the exception of those containing charm or bottom quarks .
a @xmath58 fit is performed by minimizing @xmath59 as a function of the three parameters t , v and @xmath49 , taking account of the experimental uncertainties @xmath60 .
the resulting best fit to the data for a correlated and uncorrelated jet scheme is shown in fig .
( [ fig2 ] ) .
the two cases of the fit , without and with the subtraction of the contribution from heavy quarks in an uncorrelated jet scheme are also shown in this figure .
we first note the overall behavior of the data , namely an approximately exponential decrease of the particle yield with an increasing particle mass .
such a behavior is expected in the hadron resonance gas model due to the boltzmann factors , thus indicating the presence of statistical features of hadron production in elementary collisions .
the quantitative description of the data with the statistical model is , however , rather poor and certainly no improvement is visible for the case of subtracting charm and bottom contributions . the low fit quality is reflected through the large @xmath58 values per degree of freedom .
in addition , discrepancies between individual data points and fit values larger than 5 standard deviations are not rare .
there are also problems in determination of the fit parameters : @xmath61 and @xmath49 . in @xmath58 contour plots , both in @xmath62- and @xmath63-plane ,
one notices strong anticorrelations between fit parameters @xcite .
in addition there is a series of local minima which makes it difficult to uniquely determine the model parameters .
such local minima are typical for poor fits and imply large uncertainties in the determinations of the fit parameters . in the canonical formulation of the statistical model with exact charge conservation the abundances of charged particles depend crucially on the overall charge in a system . in order to illustrate this
let us consider a model where only one charge , e.g. charm , is conserved exactly .
in such a case the multiplicity @xmath64 of particle @xmath65 with mass @xmath66 that carries charge @xmath67 in a system of the total charge @xmath10 , volume @xmath5 and temperature @xmath4 is obtained under the boltzmann approximation from @xcite : @xmath68 where @xmath69 and where @xmath70 and @xmath71 are bessel functions and the argument of @xmath70 quantifies the total number of charged particle pairs . from eq . ( [ eq : c ] )
one recognizes an essential difference in particle yields if the total charge @xmath72 or @xmath73 . indeed , assuming that @xmath74 and applying an asymptotic expansion of the bessel functions one finds e.g. that multiplicities of particles with charge @xmath75 is @xmath76 if the total charge of a system @xmath72 , and @xmath77 if the total charge of a system @xmath78 .
collisions at @xmath39=91 gev . also shown are the model parameters used in the calculations . ] for a charge neutral system the charged particles yield ( [ eq:1 ] ) are strongly suppressed due to canonical effects , since a particle has to be produced in a pair with an antiparticle in order to fulfill charge neutrality .
this is a well known `` canonical suppression effect '' which is crucial e.g. to quantify `` strangeness enhancement '' and production in hadron - hadron and heavy ion collisions at lower energies @xcite . for the total charge @xmath73 the canonical effect results in enhancement of particle yields as shown in eq .
( [ eq:2 ] ) .
the charge @xmath73 of a system is here redistributed between all particles that carry @xmath67 with weights given by the ratio of the thermal phase space of particle @xmath65 to the phase
space of all negatively or positively charged particles . here , even the boltzmann suppression is to a large extent cancelled out in the ratio ( see eq .
( [ eq:2 ] ) ) .
the above examples imply a qualitative difference in open charm and bottom production in @xmath0 collisions when using an uncorrelated and correlated jet scheme scenario . if @xmath79 in a jet then the thermal phase space of charm and bottom is strongly suppressed as in eq .
( [ eq:1 ] ) , leading to large discrepancies between model results and data .
one needs to check whether in the correlated jet scheme , where the open charm and bottom are statistically enhanced as in eq .
( [ eq:1 ] ) , the measured yields are comparable with model predictions . in quantitative analysis ,
one calculates the heavy quark yields with the partition function ( [ eq : partition1 ] ) rather than with the approximate eq .
( [ eq:2 ] ) .
( [ fig3 ] ) shows the statistical model results for charmed and bottom particles obtained in the correlated jet scheme .
one recognizes a good description of data by the canonical statistical model .
thus , if the fireballs related with each jet are charged then the model description of the distribution of these charges between different particles agrees with data .
the value of @xmath80 indicates a good quality of the model description of data .
we have to stress that , in the canonical model , yields of hidden bottom and charm meson are still strongly suppressed by the boltzmann factors , thus their multiplicities deviate from data . indeed , in fig .
( [ fig3 ] ) , the yield of @xmath81 is larger by several orders of magnitudes than the model results ( not shown in fig.([fig3 ] ) ) , indicating a non - thermal origin of this particle .
this is not the case for @xmath82 since in @xmath0 annihilations at @xmath1 gev as they are almost entirely originating from decays of bottom @xcite ..
we have analyzed the experimental data on hadron yields with light and heavy quarks in @xmath0 collisions at @xmath83 gev within the statistical model .
the conservation of five quantum numbers was included in the framework of the canonical partition function .
our results qualitatively confirmed the previous finding @xcite that statistical features are present in hadron production in @xmath0 annihilation .
the resulting temperature of @xmath84 mev lies in the bulk expected at the chemical freezeout in heavy ion collisions at high energy @xcite and agrees with first results obtained in @xmath0 systems @xcite .
our results on open charm and bottom hadrons , using the measured @xmath85 and @xmath86 fractions of jets , showing good agreement with data also confirmed earlier observation @xcite .
however , in view of a rather poor fit to measured yields of hadrons with light quarks and a clearly non - thermal origin of the hidden charm and bottom particles as well as essentially different characteristics of the collision fireball in @xmath0 and in heavy ion collisions @xcite , the general validity of the thermodynamical approach to the particle production in @xmath0 annihilation at lep energies is not obvious | we discuss the production of hadrons in @xmath0 collisions at @xmath1 gev .
we address the question wether the particle yields measured in the final states are consistent with the statistical model predictions . in the model formulation
we account for exact conservation of all relevant quantum numbers using the canonical description of the partition function . within our model
the validity of the thermodynamical approach to quantify particle production in @xmath0 annihilations is not obvious . |
to illustrate what is now known as parrondo s paradox @xcite , consider the following games : * game a : the fortune @xmath0 of the player after @xmath1 independent games is @xmath2 * game b : let @xmath3 be the 3-periodic function on @xmath4 such that @xmath5 and @xmath6 .
the fortune @xmath7 of the player after @xmath1 games is given by @xmath8 it is well - known that game a is fair if and only if @xmath9 .
in fact , if @xmath10 , the markov chain @xmath11 is transient and @xmath12 if @xmath13 , while @xmath14 if @xmath15 . when @xmath9 , @xmath11 is recurrent and @xmath16 for game b , the process @xmath17 can be seen as a particular case of a random walk in a random environment ; in fact , the space @xmath18 of environments has dimension @xmath19 with @xmath20 and @xmath21
. for more details on periodic and almost periodic environments , see , e.g. , ( * ? ? ?
* examples 1 - 2 ) .
@xcite studied the behavior of random walks in a random environment when the environments are i.i.d .
, which is not necessarily the case here . following instead ( * ? ? ?
* theorem 2.1 ) , one concludes that the process @xmath17 is recurrent if and only if @xmath22 , where @xmath23 as a result , @xmath24 .
otherwise , when @xmath25 , @xmath7 is transient and @xmath26 if @xmath27 , while @xmath28 if @xmath29 . for example , if @xmath30 and @xmath31 , then @xmath22 . if @xmath32 and @xmath33 , then @xmath34 .
suppose now that @xmath13 and @xmath35 .
then , according to the previous observations , if a player always plays game a or game b , her fortune will tend to @xmath36 with probability one .
however , if she plays game a twice , then game b twice and so on ( game c ) , or if she chooses the game at random with probability @xmath37 ( game d ) , her fortune will tend to @xmath38 .
this is parrondo s paradox and it is illustrated in figure [ fig : parrondo ] .
now consider game c , where the player alternates between game a and game b , i.e. , she plays game a once , then game b once , and so on .
what happens in this case ?
the answer will be given at the end of example [ ex : periodic2 ] .
games a and b are particular cases of random walks in a random environment , while games c and d are examples of regime switching markov chains in random environment .
the aim of this paper is to study their asymptotic behavior .
one of the first rigorous work in this field is @xcite , who studied some particular cases of random environment , namely the so - called periodic case , where each random walk is like in game b , while the player chooses at random between two games .
the author also consider some `` deterministic '' mixtures , namely the cases studied in @xcite . for other kind of random walks in random environment that are not covered by our setting , see e.g. , @xcite . in what follows , the player chooses the game to play according to a finite markov chain , and each game is a random walk in a random environment , extending the work of @xcite and @xcite .
more specifically , the model is described in the next section , together with characterizations of their asymptotic behavior . in section [ sec : rsrwre ] , the model , which is basically a regime switching random walks in random environments is described and one of the main results ( theorem [ thm : main ] ) is stated , namely that these models behave like simple random walks , in the sense that they are transient , meaning that they converge to either of @xmath39 with probability @xmath40 , for almost every environment , or they are recurrent , meaning that the limsup and liminf converge respectively to @xmath39 with probability @xmath40 , for almost every environment . in section
[ sec : hitting ] , the three possible cases are characterized in terms of the rank of the transition matrix and the dimensions of some random subspaces occurring in the famous oseledec s theorem .
these results generalize the best known particular cases : ( i ) when the games are chosen independently , the transition matrix has rank @xmath40 , and the results can be recovered as well from standard arguments applied to random walks in random environments @xcite ; ( ii ) the periodic choices studied by many authors , starting with @xcite , where the transition matrices have full rank . finally , the proofs of the results can be found in a series of appendices .
they are inspired by the results of @xcite who studied random walks in random environments with bounded increments but no regime switching .
the results of @xcite were later refined by @xcite .
we first describe the model and then study its asymptotic behavior .
first , let @xmath41 be a complete probability space with a measure preserving transformation @xmath42 , assumed to be @xmath43-measurable and ergodic , i.e. , the @xmath42-invariant sigma - field is trivial .
next , for any @xmath44 , and any @xmath45 , @xmath46 are @xmath47-valued variables , where @xmath48 , @xmath49 . hence , the processes @xmath50 are stationary and ergodic . next , for a given @xmath51 , let @xmath52 be the nearest neighbor random walk in a random environment defined by the process @xmath50 , i.e. , @xmath53 these random walks will be the fortunes of the player as she chooses each game .
her decision process is based on the irreducible markov chain @xmath54 on @xmath55 , with transition matrix @xmath56 .
for example , in game c , one can choose @xmath57 , @xmath58 , @xmath59 , where @xmath60 is the ( deterministic ) process determined by game a , @xmath61 is the stationary ergodic process determined by game b , and @xmath62 . for game
d , @xmath63 and @xmath64 . further note that for game c , then @xmath65 , @xmath60 is the ( deterministic ) process determined by game a , and @xmath66 is the stationary ergodic process determined by game b. finally , set @xmath67 , @xmath68 , i.e. , if @xmath69 , then @xmath70 . under these assumptions ,
given @xmath71 , @xmath72 is an homogeneous markov chain on @xmath73 with transition matrix @xmath74 where @xmath75 and @xmath76 .
note that for a given environment @xmath77 , @xcite called @xmath11 a random chain in a random environment , the sequence @xmath78 playing the role of the `` environment '' .
[ rem : bolthausen ] note that our setting is a particular case of a random walk in in a random environment on a strip introduced in @xcite .
however , their results can not be applied in general here since they assumed that the resulting markov chain has only one communication class ( condition c ) , i.e. , the markov chain is almost surely irreducible ( * ? ? ?
* remark 2 ) .
this is generally not the case here .
for example , in games c and c , there are two closed classes : @xmath79 and @xmath80 . for an explicit example where their main result does not hold , see appendix [ app : bolthausen ] .
one is interested in the asymptotic behavior of the process @xmath11 .
more precisely , one would like to find conditions under which the so - called parrondo s paradox holds , i.e. , for any @xmath81 , @xmath82 , while @xmath83 here , @xmath84 means that for environment @xmath77 , the process @xmath85 starts from @xmath86 at time @xmath87 ; similarly , @xmath88 means that for environment @xmath77 , the process @xmath89 starts from at @xmath90 at time @xmath87 . for @xmath91 ,
define the ( stationary ergodic ) process @xmath92 by @xmath93 , @xmath94 .
then the asymptotic behavior of @xmath89 is completely determined by the expectation of @xmath95 , as proven in ( * ? ? ?
* theorem 2.1 ) .
[ thm : alili ] let @xmath91 be given and suppose that @xmath96 is well defined , with values in @xmath97 $ ] . if @xmath98 , then for any @xmath99 , @xmath100 if @xmath101 , then for any @xmath99 , @xmath102 finally , if @xmath103 , then for any @xmath99 , @xmath104 in addition to the trivial case where @xmath50 is constant , one can also consider the i.i.d .
case , i.e. , where for a given @xmath105 , @xmath106 are i.i.d . here
are other examples of stationary ergodic sequences .
[ exe : periodic ] @xcite studied parrondo s paradox when for each @xmath91 @xmath50 is periodic of period @xmath107 , i.e. , @xmath108 . these are particular cases of stationary ergodic sequences for which each environment @xmath109 , @xmath110 has equal probability @xmath111 , so @xmath112 this example contains games a and b as particular cases .
[ ex : contfraction ] set @xmath113 and define @xmath114 , @xmath115 .
then @xmath42 is a measure - preserving map for the law with density @xmath116 , @xmath117 .
set @xmath118 and use the two - sided extension theorem , e.g. , ( * ? ? ?
* theorem 7.1.2 ) , to obtain a stationary sequence defined for any @xmath119 .
it is easy to check that @xmath120 .
therefore , according to theorem [ thm : alili ] , the associated random walk converges to @xmath36 with probability one , for almost every environment .
note that @xmath42 is not invertible . in the next section , we study the asymptotic behavior of the process @xmath11 . to reduce the notations ,
the random environment is fixed , unless otherwise specified . throughout the rest of the paper
, one writes @xmath121 to denote @xmath122 . for any @xmath123 ,
set @xmath124 .
the proof of the following lemma in given in appendix [ pf : mainlemma ] .
[ lem : main ] let @xmath123 be given .
then @xmath125 for every @xmath126 if and only if for every @xmath126 , @xmath127 the next result is proven in appendix [ app : pf_propexit ] .
[ prop : exit ] let @xmath128 be given .
then for any @xmath91 , @xmath129 conditioning on the first play of the game yields the following .
[ prop : time ] let @xmath130 be given and set @xmath131 , @xmath132 .
then for any @xmath133 , @xmath134 the following proposition is an interesting consequence of the previous result .
its proof is given in appendix [ pf : proptime3 ] [ prop : time3 ] let @xmath130 and @xmath77 be given . if @xmath135 for some @xmath136 then @xmath137 for any @xmath138 .
similarly , if @xmath139 for some @xmath140 then @xmath137 for any @xmath141 .
next , if @xmath142 for some @xmath143 , and some @xmath138 , then @xmath144 for every @xmath145 and any @xmath138 .
moreover , if @xmath142 for some @xmath143 , and some @xmath141 , then @xmath144 for every @xmath145 and any @xmath141 .
[ prop : time4 ] for any @xmath68 and any @xmath146 , @xmath147 moreover @xmath148 , @xmath149 , is a stationary ergodic sequence and @xmath150 similarly , @xmath151 moreover @xmath152 , @xmath149 , is a stationary ergodic sequence and @xmath153 for any @xmath154 , @xmath155 as a result , for any @xmath77 , @xmath156 , showing that the logarithm of the sequence is subadditive . then follows easily .
next , @xmath157 .
this proves that @xmath148 , @xmath149 , is a stationary ergodic sequence , so by using the subadditive ergodic theorem @xcite , there exists a constant @xmath158 so that for almost every @xmath49 , @xmath159 proving the first part .
next , if @xmath160 , then @xmath161 so again @xmath162 .
the rest of the proof follows from the previous one .
we are now in a position to state the main result , proven in appendix [ pf : mainthm ] .
[ thm : main ] set @xmath163 .
then @xmath164 and one of the following three mutually exclusive cases occur : 1 . if @xmath165 and @xmath166 , then for every @xmath167 , @xmath168 2 . if @xmath169 and @xmath170 , then for every @xmath167 , @xmath171 3 . if @xmath172 , then for every @xmath167 , @xmath173
first , we express the relationship between hitting probabilities . these will be needed for computation purposes .
then it will be shown that @xmath174 and @xmath175 are related to dimensions of some random spaces through the famous oseledec s theorem stated in appendix [ app : oseledec ] .
for any given @xmath176 recall that @xmath124 and set @xmath177 . for any choice of @xmath178 set @xmath179 and similarly @xmath180 .
it is easy to check that @xmath181 , for any @xmath178 .
also , when @xmath182 , then @xmath183 .
it then follows that @xmath184 first , let @xmath185 be the random diagonal matrix with entries @xmath186 .
then , using , @xmath187 , where @xmath188 , @xmath189 , and @xmath190 where @xmath191 so that @xmath192 for any @xmath94 .
note that for @xmath193 , @xmath194 for all @xmath143 if and only if @xmath195 .
further note that @xmath196 .
it then follows that @xmath197 for all @xmath198 .
this corresponds to choosing the regimes independently , so @xmath11 is a markov chain with transition matrix @xmath199 given by @xmath200 if @xmath201 , @xmath202 if @xmath203 , and @xmath204 , whenever @xmath205 , where @xmath206 , @xmath207 , @xmath99
. then it follows from that @xmath208 , for every @xmath182 .
set @xmath209 , and @xmath210 , @xmath94 .
next , it is easy to check that for every @xmath211 , @xmath212 finally , since @xmath11 is itself a random walk in a random environment , one can apply theorem [ thm : alili ] , to obtain the following corollary .
[ cor : ind ] if @xmath213 , then for any @xmath214 , @xmath215 if @xmath216 , then for any @xmath214 , @xmath217 if @xmath218 , then for any @xmath214 , @xmath219 for game d , one finds that @xmath220 is 3-periodic with values @xmath221 .
using corollary [ cor : ind ] , one obtains that @xmath222 , where @xmath223 given by formula . in this specific example , @xmath224 , so that for any @xmath225 , @xmath226 a.s .
figure [ fig : graphmud ] illustrates that @xmath227 quite often .
since @xmath56 is invertible , it follows that @xmath228 and @xmath229 , as defined in , are invertible . for @xmath230 ,
set @xmath231 , where @xmath232 .
since each @xmath233 is invertible and the sequence of @xmath234 matrices @xmath233 is stationary and ergodic , it follows from oseledec s theorem ( theorem [ thm : oseledec ] ) that with probability 1 , the random sets @xmath235 and @xmath236 are subspaces with deterministic dimensions @xmath237 and @xmath238 respectively , provided @xmath239 is integrable .
note that for any @xmath99 , @xmath240 .
also , with probability 1 , the random sets @xmath241 and @xmath242 are subspaces with deterministic dimensions @xmath243 and @xmath244 respectively , provided @xmath245 is integrable .
the proof of the next result is given in appendix [ pf : thmoseledec ] .
[ thm : oseledecfull ] suppose that @xmath56 is invertible and assume that @xmath239 and @xmath245 are integrable . 1 . if @xmath246 , then @xmath247 and @xmath248 , @xmath77 a.s . , for any @xmath143 and any @xmath141 .
2 . @xmath249 if and only if @xmath250 . in the latter case , @xmath251 for any @xmath143 .
3 . if @xmath252 , then @xmath253 , and @xmath248 , @xmath254 a.s . , for any @xmath143 and any @xmath255 .
4 . @xmath256 if and only if @xmath170 . in the latter case , @xmath257 for any @xmath143 .
it follows from theorem [ thm : main ] that @xmath258 , so theorem [ thm : oseledecfull ] implies that one can not have at the same time @xmath259 .
hence , if @xmath249 , then @xmath260 and it follows from theorem [ thm : main ] that @xmath253 . in this case ,
@xmath168 for every @xmath167 .
+ on the other hand , if @xmath256 , then @xmath261 and it follows from theorem [ thm : main ] that @xmath247 . in this case ,
@xmath262 for every @xmath167 .
+ finally , if @xmath263 , then @xmath264 , and @xmath173 for every @xmath167 .
[ cor : ruelle ] suppose that the measure preserving mapping @xmath42 is invertible .
then , if @xmath265 , then @xmath266 , @xmath267 , @xmath77 a.s .
, for every @xmath81 . also ,
if @xmath249 , then @xmath268 , and @xmath269 , @xmath77 a.s . , for every @xmath167 . finally , if @xmath270 , then @xmath173 for every @xmath167 .
[ ex : periodic2][periodic probabilities ] suppose that the processes @xmath50 are periodic for all @xmath91 , and denote by @xmath220 the least common multiple of their respective periods .
further set @xmath271 .
then @xmath237 is the number of eigenvalues of @xmath272 less than or equal to @xmath40 in absolute value , while @xmath243 is the number of eigenvalues of @xmath272 greater of equal to @xmath40 in absolute value .
game c is an example of a @xmath56 with full rank and periodic probabilities . in this case
@xmath57 , @xmath273 , and @xmath274 . as a result , from corollary [ cor : ruelle ] , @xmath275 , @xmath77 a.s .
, for every @xmath167 .
this explains why game c has a paradoxical behavior .
suppose now that one alternates between games a and b , i.e. , one plays game
c. then @xmath65 , @xmath63 , @xmath276 , and @xmath274 . from corollary [ cor : ruelle ] , @xmath277 , @xmath77 a.s . , for every @xmath167
. hence , in this case , the game does not have a paradoxical behavior .
suppose that @xmath56 has rank @xmath279 .
one can assume , without loss of generality that @xmath280 , where @xmath281 has rank @xmath278 , and @xmath282 , so that @xmath283 . further let @xmath284 be the @xmath285 diagonal matrix formed with the first @xmath278 rows and columns of @xmath286 , and let @xmath287 stands for the diagonal matrix formed with the last @xmath288 rows and columns of @xmath286 .
finally , let @xmath289 be the matrix composed for the first @xmath278 columns of @xmath290 and set @xmath291 for the remaining @xmath288 columns of @xmath290 . for any @xmath94 , set @xmath292 and @xmath293 .
[ hyp : inv ] with probability @xmath40 , @xmath294 and @xmath295 are invertible . under this assumption
, one can define , for any @xmath94 , @xmath296 and @xmath297 , where @xmath298 .
note that both @xmath294 and @xmath295 are stationary ergodic sequences since @xmath299 is not random . as before
, it follows from oseledec s theorem ( theorem [ thm : oseledec ] ) that with probability 1 , the random sets @xmath300 and @xmath301 are subspaces with deterministic dimensions @xmath302 and @xmath303 respectively , provided @xmath304 is integrable .
also , with probability 1 , the random sets @xmath305 and @xmath306 are subspaces with deterministic dimensions @xmath307 and @xmath308 respectively , provided @xmath309 is integrable .
the proof of the following result is given in appendix [ pf : oseledecgen ] .
[ thm : oseledecgen ] suppose that hypothesis [ hyp : inv ] holds and assume that @xmath304 and @xmath310 are integrable . 1 . if @xmath311 , then @xmath247 and @xmath248 , @xmath77 a.s . , for any @xmath143 and any @xmath141 .
2 . @xmath312 if and only if @xmath250 . in the latter case , @xmath251 for any @xmath143 .
3 . if @xmath313 , then @xmath253 , and @xmath248 , @xmath254 a.s . , for any @xmath143 and any @xmath255 .
@xmath314 if and only if @xmath170 . in the latter case , @xmath257 for any @xmath143 .
it follows from theorem [ thm : main ] @xmath258 , so theorem [ thm : oseledecgen ] implies that one can not have at the same time @xmath315 .
hence , if @xmath312 , then @xmath260 and it follows from theorem [ thm : main ] that @xmath253 . in this case , @xmath168 for every @xmath167 . on the other hand ,
if @xmath314 , then @xmath261 and it follows from theorem [ thm : main ] that @xmath247 . in this case , @xmath262 for every @xmath167 . finally , if @xmath316 , then @xmath264 , and @xmath173 for every @xmath167 . using our methodology when @xmath317 , one recovers theorem [ thm : alili ] due to @xcite .
[ cor : alili1 ] suppose @xmath317 and set @xmath318 .
then @xmath101 if and only if @xmath319 ; @xmath98 if and only if @xmath320 ; @xmath103 if and only if @xmath321 . note that is this case , @xmath322 , where @xmath323 . since @xmath324 , @xmath325 , and @xmath326 , where @xmath327 .
next , it is easy to check that @xmath328 and @xmath329 , as @xmath330 .
it then follows from oseledec s theorem ( theorem [ thm : oseledec ] ) that @xmath101 if and only if @xmath319 , @xmath331 , @xmath332 and @xmath333 .
just take @xmath334 to get @xmath335 . here
@xmath336 . also take @xmath337 to obtain @xmath338 .
similarly , @xmath98 if and only if @xmath339 , @xmath340 , @xmath341 and @xmath342 . finally , @xmath103 if and only if @xmath343 , and @xmath344 , since in this case , taking @xmath345 , one gets @xmath346 .
[ cor : ruelle2 ] suppose that the mapping @xmath347 is invertible .
then , if @xmath348 , then @xmath349 , @xmath267 , @xmath77 a.s .
, for every @xmath81 . also ,
if @xmath312 , then @xmath350 , and @xmath269 , @xmath77 a.s . , for every @xmath167 . finally , if @xmath351 , then @xmath173 for every @xmath167 .
[ ex : weird ] suppose @xmath352
. then @xmath353 . in this example , @xmath354 , and + @xmath355 .
it follows that @xmath356 .
hence @xmath294 is invertible if and only if @xmath357 a.s . what happens if hypothesis [ hyp : inv ] is not met ?
in this case , one proceeds almost as before , reducing the dimension of @xmath358 instead .
for example , taking @xmath359 in example [ ex : weird ] , i.e. , games 1 and 2 are the same , @xmath360 , @xmath361 , where @xmath362 , and @xmath363 .
hence , @xmath364 .
thus hypothesis [ hyp : inv ] does not hold .
however , setting @xmath365 , one has @xmath366 and @xmath367 also , @xmath368 .
it then follows from that @xmath369 and @xmath370 .
hence we are back to the first case considered , i.e. , the rank @xmath40 case .
the following statement of the celebrated oseledec s theorem is taken from @xcite .
[ thm : oseledec ] let @xmath371 be a stationary ergodic sequence of @xmath372 matrices such that @xmath373
. then there exists constants @xmath374 with the following properties : 1 . with probability 1 , the random sets @xmath375 are subspaces .
the map @xmath376 is measurable and if @xmath42 is the measure preserving map for which @xmath377 , then @xmath378 .
set @xmath380 and let @xmath381 be the unique indices at which @xmath382 jumps , i.e. , @xmath383 . then for @xmath384 , one has @xmath385 4 .
the sequence of matrices @xmath386 converges almost surely to a limit matrix @xmath387 with eigenvalues @xmath388 .
the orthogonal complement of @xmath389 in @xmath390 is the eigenspace of @xmath387 corresponding to @xmath391 .
if @xmath392 and @xmath393 with probability 1 , then @xmath394 and @xmath395 , the subspace corresponding to @xmath396 is a proper nonempty subspace of @xmath397 .
we give first the formula for the @xmath398-th visiting time of @xmath0 to a state @xmath176 . to this end , let @xmath399 denotes the time of the k - th visit to @xmath130 and define @xmath400 as the number of visits to site @xmath130 .
[ prop : kthvisit ] for @xmath130 given , set @xmath401 while recalling that , for every @xmath178 , @xmath94 , @xmath402 .
then for any @xmath149 , @xmath403 and @xmath404 with @xmath405 .
[ pf : kthvisit ] set @xmath406 . then for any @xmath178 and any @xmath94 , one has @xmath407 the result then follows by induction on @xmath398 . [ lem : main0 ] the following statements hold : * @xmath408 for every @xmath167 if and only if for every @xmath409 , @xmath410 * @xmath411 for every @xmath167 if and only if for every @xmath409 , @xmath412 * @xmath413 for every @xmath167 , if and only if for every @xmath414 , @xmath415 first , @xmath416 for every @xmath167 if and only if for every @xmath409 , @xmath417 . by lemma [ lem :
main ] , this is equivalent to and .
this proves ( i ) .
next , ( ii ) follows from ( i ) applied to @xmath418 . to prove ( iii ) , note that @xmath419 for every @xmath167 if and only if for every @xmath420 , @xmath421 .
the latter implies that @xmath422 whenever @xmath423 , using . using this result with @xmath418 ,
one obtains that @xmath424 for every @xmath167 implies that for every @xmath409 , @xmath425 whenever @xmath138 .
then yields that @xmath426 . to complete the proof , suppose now that holds true .
proposition [ prop : time ] with @xmath427 , combined with proposition [ prop : time3 ] then implies that @xmath428 for any @xmath429 . by using
, it suffices to show that @xmath430 for every @xmath143 .
now , by hypothesis , @xmath431 , defined by @xmath432 , @xmath178 , is a stochastic matrix , so @xmath433 for every @xmath149 , by using . as a result , @xmath434 .
[ pf : mainlem ] set @xmath435 . then , @xmath436 as a result , for @xmath138 , one obtains @xmath437 while if @xmath141 , then @xmath438 first , suppose that @xmath439 for any @xmath167 .
then , according to , @xmath440 for any @xmath143 and any @xmath138 .
next , if @xmath141 , then @xmath439 implies that @xmath441 .
hence holds true .
+ suppose now that holds true . then combining and , one obtains @xmath442 therefore , to complete the proof , it suffices to show that @xmath443 for any @xmath189 . to this end , note that using , one gets @xmath444 \nonumber\\ & = & \sum_{\beta=1}^m \sum_{\gamma=1}^m q_{\alpha \beta } p_{\beta \ell } p^\omega_{\beta,\ell+1}(\tau_\ell < \infty , g_{\tau_\ell}=\gamma ) p^\omega_{\gamma\ell}({\mathcal{a}}_\ell ) \nonumber\\ & & \qquad + \sum_{\beta=1}^m \sum_{\gamma=1}^m q_{\alpha \beta } q_{\beta \ell } p^\omega_{\beta,\ell-1}(\tau_\ell < \infty , g_{\tau_\ell}=\gamma ) p^\omega_{\gamma\ell}({\mathcal{a}}_\ell ) \nonumber\\ & = & \sum_{\gamma=1}^m p^\omega_{\alpha \ell}(\tau_\ell <
\infty , g_{\tau_\ell}=\gamma ) p^\omega_{\gamma\ell}({\mathcal{a}}_\ell),\label{eq : main11}\end{aligned}\ ] ] since @xmath445 by hypothesis , @xmath446 for any @xmath91 , so by it follows that @xmath447 . further set @xmath448 .
then , from , @xmath449 , so @xmath450 . since @xmath451
, it follows that @xmath452 . hence the result .
let @xmath77 be given .
first , is obviously true for @xmath453 .
if is not true for some @xmath454 , then for some @xmath455 , and some @xmath456 , @xmath457 next , for @xmath456 and @xmath133 , set @xmath458 .
it then follows that for any @xmath133 , and @xmath459 , @xmath460 while @xmath461 , and @xmath462 .
it follows that for every @xmath77 , there is a sub - stochastic matrix @xmath463 on @xmath464 , so that @xmath465 , for any @xmath466 .
note that @xmath467 if @xmath459 , while @xmath468 , and @xmath469 .
next , @xmath470 if @xmath471 , and by induction , @xmath472 if @xmath473
. therefore , for any @xmath77 , there exists @xmath474 so that @xmath475 . as a result , using ( * ? ? ?
* theorem 8.4 ) , one obtains that @xmath476 , contradicting the hypothesis that @xmath477 for some @xmath466 . hence the result .
one only proves the proposition for @xmath138 , the proof of the case @xmath141 being similar . if @xmath478 for @xmath479 , then for some @xmath91 , @xmath480 .
hence , for any @xmath481 so that @xmath482 , one has @xmath483 , according to proposition [ prop : time ]
. therefore @xmath484
. next , if @xmath478 for some @xmath485 , then proposition [ prop : time ] implies that @xmath486 .
this proves the first part of the proposition .
suppose now that @xmath428 , @xmath77 a.s .
since @xmath54 is irreducible , for a given @xmath481 , one can find @xmath68 so that @xmath487
. as a result , it follows from proposition [ prop : time ] that @xmath488 a.s .
hence , using stationarity , @xmath489 , @xmath77 a.s . , entailing that @xmath490 , @xmath77 a.s .
one first proves that @xmath258 .
in fact , it follows from proposition [ prop : time4 ] that @xmath491 .
one now shows by contradiction that it is impossible to have @xmath260 and @xmath261 .
so suppose that @xmath492 .
since holds by proposition [ prop : exit ] , it follows that @xmath493 which is impossible since @xmath260 and @xmath261 .
hence , @xmath258 .
it follows from propositions [ prop : time3]-[prop : time4 ] that if @xmath250 , then for any @xmath91 and any @xmath494 , @xmath495 , @xmath77 a.s .
similarly , if @xmath261 , then for any @xmath91 and any @xmath496 , @xmath495 , @xmath77 a.s .
next , if @xmath253 , then @xmath497 a.s . by proposition [ prop : time4 ] .
it then follows from proposition [ prop : time3 ] that @xmath498 , @xmath77 a.s .
, for any @xmath91 and any @xmath138 . similarly , if @xmath247 , then @xmath499 a.s . by proposition
[ prop : time4 ] , so proposition [ prop : time3 ] entails that @xmath440 , @xmath77 a.s .
, for any @xmath91 and any @xmath141 . as a result
, using lemma [ lem : main0 ] , one gets ( 1 ) , ( 2 ) , and ( 3 ) . for any @xmath99 , set @xmath500 . since @xmath501 , one gets @xmath502 , where @xmath503 .
note that @xmath504 and @xmath505 .
by oseledec s theorem ( theorem [ thm : oseledec ] ) , there exists constants @xmath506 with the following properties : 1 . with probability 1 , the random sets @xmath507 are subspaces .
the map @xmath508 is measurable and @xmath509 .
set @xmath511 and let @xmath512 be the unique indices at which @xmath382 jumps , i.e. , @xmath383 .
then for @xmath513 , one has @xmath514 since each @xmath233 is invertible , it then follows from oseledec s theorem that for any @xmath515 , the product @xmath516 can be replaced with the product @xmath517 and one still gets the same constants @xmath518 and the dimensions of the associated subspaces are exactly the same .
let @xmath130 be given .
for @xmath141 , set @xmath519 $ ] , @xmath141 , @xmath211 . using , one can write @xmath520 since @xmath187 for every @xmath521 , it follows that the vectors @xmath522 are linearly independent , and they belong to @xmath523 . as a result , @xmath524 .
also , @xmath525 , so @xmath526\in \bar{v}_{0}$ ] .
. then from almost every environment , there exists a random vector @xmath527 so that @xmath528 .
in particular , @xmath529 a.s . , and @xmath530 a.s .
hence @xmath531 , since by definition , @xmath532 .
it then follows that @xmath533 , and one can deduce from that for any @xmath141 , @xmath534 .
hence , for any @xmath91 and any @xmath141 , one has @xmath535 as a result , @xmath247 .
this completes the proof of ( 1 ) .
suppose first that @xmath536 .
recall that for any @xmath91 , @xmath537 . then @xmath538 . let @xmath539 be a ( possibly random ) basis of @xmath540 .
the proof proceeds as in @xcite .
+ ( a ) the last @xmath541 components of @xmath539 are linearly independent .
if not , there exists a member @xmath542 of @xmath543 so that its last @xmath541 components are zero . for @xmath141 , define @xmath544 .
then @xmath545 $ ] , with @xmath546 .
set @xmath547 , @xmath91 , @xmath548 .
let @xmath549 the markov chain starting at @xmath550 associated with @xmath72 but absorbed on @xmath551 , and set @xmath552 . since @xmath553 , @xmath554 .
then @xmath555 is a bounded martingale with @xmath556 .
because of its random walk nature , either @xmath549 is absorbed or @xmath557 . as a result , by the martingale convergence theorem
, it follows that @xmath558 , so @xmath559 . since the latter is true for any @xmath141 and any @xmath189
, one may conclude that @xmath560 , contradicting the assumption .
thus , one may conclude that the last @xmath541 components of @xmath539 are linearly independent .
+ ( b ) based on ( a ) , there exists @xmath553 such that its last @xmath541 components are @xmath40 .
then , for @xmath141 , define @xmath544 , so @xmath561 $ ] , with @xmath562 .
further set @xmath563 , @xmath91 , @xmath564 .
then , since @xmath553 , one has @xmath554 .
next , define @xmath565 $ ] , @xmath548 .
then , using , one can write @xmath566 , @xmath567 . note also that @xmath568 .
one will prove that @xmath569 for all @xmath548 .
suppose this is not true .
then there exists @xmath570 so that @xmath571 . as before , let @xmath549 the markov chain starting at @xmath550 associated with @xmath72 but absorbed on @xmath572 , and set @xmath573 .
then @xmath555 is a bounded martingale with @xmath574 , and @xmath575 since @xmath576 on @xmath577 . as a result , @xmath578 .
finally , @xmath579 is also a bounded martingale and @xmath580 on @xmath581
. therefore @xmath582 . finally , @xmath583 showing that @xmath584 .
hence @xmath569 for any @xmath548 .
it then follows that @xmath250 , and so @xmath251 , @xmath77 a.s .
, for any @xmath143 . to complete the proof of ( ii ) , suppose now that @xmath260
. then @xmath585 . to see this , note that @xmath586 for every @xmath211 , since @xmath587 and holds .
as a result , @xmath588 . since @xmath589
, it follows that @xmath585 .
the proofs of ( 3 ) and ( 4 ) are similar .
in fact , setting @xmath590 $ ] , @xmath138 , @xmath211 , and using , one can write @xmath591 as a result , @xmath592 . using the same notations as in the proof of theorem [ thm : oseledecfull ] , it follows from ( * ? ? ?
* theorem 3.1 ) that the lyapunov exponents @xmath593 associated with @xmath594 are @xmath595 .
hence , @xmath596 and @xmath597 .
in particular , if @xmath246 , then @xmath598 and if @xmath536 , then @xmath252 . recall that from , @xmath599 with @xmath187 .
also , since @xmath600 , it follows that @xmath601 , @xmath182 , with @xmath602 .
let @xmath603 and @xmath604 represent respectively the first @xmath278 components and the last @xmath288 components of @xmath605
. for simplicity , set @xmath606 .
it then follows from the representation of @xmath56 and that for any @xmath182 and any @xmath607 , @xmath608 as a result , one can write , for any @xmath182 and any @xmath607 , @xmath609.\ ] ] furthermore , since @xmath602 and @xmath610 , it follows from that for any @xmath94 , @xmath611.\ ] ] next , using ,
one can write , for any @xmath612 and any @xmath607 , @xmath613 while for any @xmath182 , @xmath614 also , for any @xmath607 , @xmath615 now , set @xmath616 $ ] , where @xmath617 is the vector composed of the first @xmath278 components of @xmath618 .
next , taking linear combinations of the functions @xmath619 , @xmath607 , one can define , for any @xmath182 and any @xmath620 , @xmath621 where @xmath622 is the first @xmath278 components of @xmath623 $ ] .
also @xmath624 , and @xmath625 , for any @xmath94 . under hypothesis
[ hyp : inv ] , one can then apply oseledec s theorem .
define @xmath626 as the dimension of the set @xmath627 of vectors @xmath628 so that @xmath629 . next , let @xmath130 be given . for @xmath141 , set @xmath630 $ ] , @xmath141 , @xmath631
, one can write @xmath632 since the vectors @xmath633 are linearly independent for every @xmath634 , it follows that the vectors @xmath635 are linearly independent as well , and they belong to @xmath636 . as a result , @xmath637 .
also , @xmath638 , so @xmath526\in \check{v}_{0}$ ] .
then one proceeds as in the proof of theorem [ thm : oseledecfull ] , with a few minor changes .
set @xmath639 , @xmath640 and @xmath641 . using @xcite notations , for a given @xmath642 and a given stochastic matrix @xmath643 , for @xmath644 , define @xmath645 , where @xmath646 .
it is not easy to compute @xmath647 for a small @xmath648 , but if the probabilities are periodic with period @xmath19 for example , then @xmath649 , @xmath650 , @xmath651 , @xmath652 , @xmath653 and @xmath654 . in (
* theorem 1 ) , the authors claim that the limit @xmath655 exists and is independent of @xmath643 .
one crucial hypothesis for the proof is the existence of only one communication class ( a.e . ) .
as noted in remark [ rem : bolthausen ] , this is usually not the case here , especially for games c and c. we show next that ( * ? ? ?
* theorem 1 ) does not hold for game c , because either the limit does not exist , or it depends on the initial value @xmath643 .
in fact , starting from @xmath656 , the limit does not exist .
now , @xmath657 and @xmath658 .
next , if one starts from @xmath659 , then @xmath660 .
durrett , r. , kesten , h. , and lawler , g. ( 1991 ) . making money from fair games . in _
random walks , brownian motion , and interacting particle systems _ , volume 28 of _ progr .
_ , pages 255267 .
birkhuser boston , boston , ma .
pyke , r. ( 2003 ) . on random walks and diffusions related to parrondo s games . in _ mathematical statistics and applications : festschrift for constance van eeden _ , volume 42 of _ ims lecture notes monogr .
_ , pages 185216 .
inst . math .
, beachwood , oh . | combining losing games into a winning game is a century old dream .
an illustration of what can be done is the so - called parrondo s paradox in the physics literature .
this
paradox " is extended to regime switching random walks in random environments .
the paradoxical behavior of the resulting random walk is explained by the effect of the random environment
. full characterization of the asymptotic behavior is achieved in terms of the dimensions of some random subspaces occurring in the famous oseledec s theorem .
basically , these models have the same asymptotic behavior as simple random walks , in terms of transience and recurrence . |
in the currently favored cosmological scenario , the matter content of the universe is dominated by a cold and dark component of unknown particle species whose only significant interaction with the smaller baryonic matter fraction occurs through the gravitational force .
this `` cold dark matter '' ( cdm ) picture is most strongly required by observations on the largest scales @xcite .
the cdm scenario holds that galaxies form within the potential wells of extended dark - matter halos which began their collapse while baryonic matter was still ionized @xcite .
this view is supported by direct evidence for dark matter in disk galaxies from the observation of rotational velocities that remain approximately constant out to radii at which the stellar galactic component makes a diminishing contribution ( e.g. * ? ? ?
* ; * ? ? ?
* ) . unlike disk galaxies , early - type galaxies ( e and s0 ) are generally pressure supported and lack bright kinematic tracers at large radii such as h@xmath1 .
as such their kinematics are more difficult to measure and interpret in terms of mass density profiles .
the density structure of early - type galaxies is nevertheless of great interest for numerous reasons .
first , their structure is the `` fossil record '' of their formation and evolution processes ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
hierarchical cdm galaxy - formation theories hold that early - type galaxies are built through the merging of late - types @xcite , which should have predictable consequences for the structure of the merger products
. the most stringent test of these theories will require precise observational measurements of early - type mass profiles .
for example , detailed measurement of the structure of high - surface - brightness early - type galaxies will enable quantitative tests of the cdm theory on scales where baryonic and radiative processes have significant effects upon the structure of the host dark - matter halo ( e.g. through adiabatic contraction
, e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , altering it significantly relative to the form expected to result from pure collisionless dark - matter collapse ( e.g. * ? ? ?
* ; * ? ? ?
second , early - type galaxies exhibit great regularity in their photometric , spectroscopic , and kinematic properties , as described , e.g. , by the well - known fundamental plane ( fp ) relation between velocity dispersion , effective radius , and surface brightness @xcite . the `` tilt '' of the fp relative to the simple expectation of the virial theorem
can be understood in terms of a dependence of galaxy structure and mass - to - light ratio on total galaxy mass @xcite . however ,
further constraints on the mass structure of early - type galaxies are needed in order to distinguish between the various effects of differing stellar populations ( the explanation given by , e.g. , @xcite ) , differing density profiles ( `` structural nonhomology '' or `` weak homology '' ) , and differing dark - matter fractions ( the explanation of , e.g. , @xcite ) in giving rise to the fp . despite great progress
, observational results remain uncertain because of the small number of individual galaxies suitable for study .
stellar - dynamical measurements of local early - type galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the statistics of early - type gravitational - lens galaxies ( e.g. * ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and combined lensing and dynamical measurements of the few systems amenable to such study ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* hereafter kt ) generally argue for the presence of a significant amount of dark matter even on the scale of the half - light radius , leading to an approximately linear increase of enclosed mass with radius and thus to a flat equivalent rotation curve as in disk galaxies .
however , in some cases the presence of dark matter is not required , and the observed kinematics can be reproduced with a constant mass - to - light ratio ( e.g. * ? ? ?
* ; * ? ? ?
* ) , corresponding to a declining rotation curve .
furthermore , @xcite has pointed out an apparent conflict between isothermal mass profiles ( i.e. flat rotation curves ) and some gravitational - lens time delays under the assumption of @xmath2 70 km s@xmath3 mpc@xmath3 . due to this persistent uncertainty about the structure and diversity of early - type galaxies
( see also * ? ? ?
* ) , it is important to collect precision data for a larger number of objects , possibly spanning a large range of redshifts , environments , and mass .
strong gravitational lensing provides the most direct probe of mass in early - type galaxies : a measurement of the mass enclosed within the einstein radius .
when combined with a spatially resolved measurement of the line - of - sight velocity - dispersion profile and the surface brightness of the lens galaxy , this mass measurement can be used to constrain the luminous and dark - matter mass profiles simultaneously through the jeans equation ( e.g. kt ) .
unfortunately , only a handful of known strong lenses are amenable to this type of analysis , often because the back - ground source is a bright qso whose images outshine the lens galaxy .
we have therefore initiated the sloan lens acs ( slacs ) survey ( see * ? ? ?
* ) to discover a much larger sample of new early - type strong gravitational lenses suitable for detailed photometric and dynamical study and thereby realize the full potential of gravitational lensing as a probe of early - type galaxy structure .
slacs uses the method described by ( * ? ? ?
* hereafter b04 ) to select candidate galaxy - scale gravitational lens systems from within the sloan digital sky survey ( sdss ) spectroscopic database on the basis of multiple higher - redshift emission lines in the spectrum of lower - redshift target galaxies .
for discussion of other spectroscopic lens surveys and lens discoveries based on the identification of anomalous emission lines , see @xcite .
the sdss data also provide stellar velocity dispersion measurements for all slacs lens candidates . the most promising candidates ( see [ sec : selection ] )
are observed through f435w and f814w using the advanced camera for surveys ( acs ) aboard the _ hubble space telescope _
( _ hst _ ) .
acs images enable us to confirm bona fide lenses , measure detailed photometric and morphological parameters of the lensing galaxies , and obtain accurate astrometry and surface - brightness measurements of strongly lensed features with which to constrain gravitational - lens models .
finally , confirmed lenses from the program are being targeted for deep ground - based spectroscopy in order to obtain spatially resolved velocity - dispersion profiles of the lensing galaxies , which will provide more precise dynamical constraints on their mass structure .
this paper ( paper i ) is the first in a series presenting the results of the current cycle13 _ hst _ imaging component of the slacs survey .
this paper presents the catalog of new lenses confirmed by the slacs _
snapshot survey and their properties as measured by sdss , describes our methods of image processing and lens - galaxy image subtraction , and provides an analysis of the possible selection biases that bear upon whether or not our lenses are a representative sample of similar sdss early - type galaxies .
we also present ground - based integral - field spectroscopy of several slacs systems which supports a strong - lensing interpretation of the observed features .
paper ii presents photometric and morphological measurements of the slacs lens sample from ( _ hst _ ) imaging and places the sample within the context of the fp , paper iii presents gravitational - lens and dynamical modeling results , and paper iv focuses on the properties of the lensed source galaxies .
the outline of this paper is as follows : in [ sec : survey ] , the survey selection procedure and the available data of each system is described . in
[ lenscat ] , nineteen newly discovered lens systems and nine unconfirmed systems or non - detections are presented . in [ sec : stats ] , the statistics and possible selection effects of the sample are discussed .
[ sec : future ] discusses some of the implications of the slacs survey for future surveys and in [ sec : conclusions ] conclusions are drawn and future work is discussed . throughout this paper , we assume a cosmological model with @xmath4 , @xmath5 , and @xmath6kms@xmath3mpc@xmath3 ( with @xmath7 ) .
lccccccccc sdss [email protected] & 0655 - 52162 - 392 & 18.00,16.81,16.39 & @xmath9 & 0.1954 & 0.6322 & @xmath10 & lrg & yes + sdss [email protected] & 0668 - 52162 - 428 & 19.07,17.46,16.90 & @xmath11 & 0.3317 & 0.5235 & @xmath12 & lrg & yes + sdss [email protected] & 0541 - 51959 - 145 & 19.38,17.84,17.15 & @xmath14 & 0.3223 & 0.5812 & @xmath15 & lrg & yes + sdss [email protected] & 0441 - 51868 - 108 & 18.63,17.51,17.07 & @xmath16 & 0.1943 & 0.4462 & @xmath17 & main & ?
+ sdss [email protected] & 0472 - 51955 - 429 & 17.31,16.22,15.78 & @xmath18 & 0.1642 & 0.3239 & @xmath19 & lrg & yes + sdss [email protected] & 0902 - 52409 - 577 & 18.57,17.61,17.22 & @xmath20 & 0.1310 & 0.4670 & @xmath21 & main & ?
+ sdss [email protected] & 0902 - 52409 - 068 & 18.41,17.17,16.62 & @xmath22 & 0.2405 & 0.4700 & @xmath23 & lrg & yes + sdss [email protected] & 0572 - 52289 - 495 & 18.52,17.48,17.02 & @xmath24 & 0.1260 & 0.5350 & @xmath25 & main & yes + sdss [email protected] & 0272 - 51941 - 151 & 17.07,16.03,15.57 & @xmath26 & 0.1589 & 0.2764 & @xmath27 & lrg & ?
+ sdss [email protected] & 0835 - 52326 - 571 & 18.72,17.56,17.12 & @xmath28 & 0.2285 & 0.8232 & @xmath29 & main & ?
+ sdss [email protected] & 0969 - 52442 - 134 & 18.40,17.22,16.65 & @xmath30 & 0.2150 & 0.4808 & @xmath31 & main & yes + sdss [email protected] & 0847 - 52426 - 549 & 18.40,17.26,16.77 & @xmath32 & 0.2318 & 0.7946 & @xmath33 & main & yes + sdss [email protected] & 0337 - 51997 - 480 & 18.58,17.59,17.24 & @xmath34 & 0.2243 & 0.7843 & @xmath35 & main & yes + sdss [email protected] & 0783 - 52325 - 279 & 18.80,17.46,17.01 & @xmath36 & 0.2333 & 0.4488 & @xmath37 & lrg & ?
+ sdss [email protected] & 0910 - 52377 - 503 & 18.34,17.45,17.05 & @xmath38 & 0.0808 & 0.7115 & @xmath39 & main & yes + sdss [email protected] & 0605 - 52353 - 503 & 18.26,16.98,16.49 & @xmath40 & 0.2046 & 0.4814 & @xmath41 & lrg & yes + sdss [email protected] & 0788 - 52338 - 605 & 16.39,15.56,15.17 & @xmath42 & 0.0629 & 0.5350 & @xmath43 & main & yes + sdss [email protected] & 0617 - 52072 - 561 & 17.94,16.84,16.39 & @xmath44 & 0.1883 & 0.3955 & @xmath45 & lrg & ?
+ sdss [email protected] & 0815 - 52374 - 337 & 18.78,17.60,17.09 & @xmath46 & 0.1989 & 0.6656 & @xmath47 & main & yes + sdss [email protected] & 0364 - 52000 - 084 & 18.54,17.29,16.89 & @xmath48 & 0.2076 & 0.5241 & @xmath49 & lrg & yes + sdss [email protected] & 0626 - 52057 - 518 & 18.84,17.41,16.92 & @xmath50 & 0.2479 & 0.7933 & @xmath51 & lrg & yes + sdss [email protected] & 0627 - 52144 - 464 & 18.92,17.68,17.18 & @xmath52 & 0.2282 & 0.6745 & @xmath53 & main & ?
+ sdss [email protected] & 0973 - 52426 - 464 & 18.01,16.91,16.40 & @xmath54 & 0.1784 & 0.4357 & @xmath55 & lrg & ?
+ sdss [email protected] & 0352 - 51789 - 563 & 16.86,15.97,15.54 & @xmath56 & 0.0899 & 0.7367 & @xmath57 & main & yes + sdss [email protected] & 0677 - 52606 - 520 & 18.97,17.63,17.14 & @xmath58 & 0.2285 & 0.4635 & @xmath59 & lrg & yes + sdss [email protected] & 0743 - 52262 - 304 & 17.58,16.39,15.96 & @xmath60 & 0.1553 & 0.5170 & @xmath61 & lrg & yes + sdss [email protected] & 0645 - 52203 - 517 & 16.07,15.21,14.82 & @xmath62 & 0.0819 & 0.5324 & @xmath63 & main & yes + sdss [email protected] & 0684 - 52523 - 311 & 19.81,18.51,17.95 & @xmath64 & 0.4168 & 0.7145 & @xmath65 & lrg & ? [ galtab ] the method by which we select our lens candidates is essentially that described by b04 .
briefly , we subtract best - fit pca templates ( a byproduct of the redshift pipeline ) from the observed sdss target - galaxy spectra .
we require the sdss continuum to be well - fit by the template , which effectively yields a parent sample of galaxies with well - behaved absorption - dominated spectra and very secure redshifts , which we denote @xmath66 ( for `` foreground '' ) .
the residual spectra are then scanned for nebular line emission at redshifts greater than @xmath66 .
spectra in which such emission is significantly detected for at least three separate common atomic transitions at a single background redshift ( @xmath67 ) are taken as lens candidates for having emission at two different redshifts along the same line of sight as sampled by the 3-arcsec diameter sdss spectroscopic fiber .
taking @xmath66 , @xmath67 , and the measured stellar velocity dispersion @xmath68 from sdss allows us to determine to first order the strong - lensing cross section for each system in the source plane using a singular isothermal sphere ( sis ) model as @xmath69 , with the einstein radius given by @xmath70 . in the image plane
the region of multiple imaging is four times larger than the source - plane strong - lensing cross section , in the sense that all images within a radius @xmath71 of the potential center will show strongly - lensed counterimages . to maximize the number of strong lenses in our survey , we formed our cycle13 _ hst _ target list from the candidates with the highest predicted lensing cross section . thus our _ hst _ target sample is velocity - dispersion selected to leading order , with an additional selection bias in favor of systems with larger angular - diameter distance ratios @xmath72 ( a function of the redshifts @xmath67 and @xmath66 ) . we take 20 of our targets from the list of lens candidates published in b04 , which were selected from within the luminous red galaxy spectroscopic sample of the sdss ( lrg , * ? ? ? * ) .
the lrg sample is defined by photometric selection cuts that very efficiently select massive early - type galaxies in the redshift range @xmath73 , as confirmed by sdss spectroscopy .
these galaxies are very homogeneous in their spectral , photometric , and morphological properties , and we place no further requirements on our lrg - sample lens candidates beyond their photometric selection as described in @xcite and subsequent spectroscopic confirmation as galaxies with redshifts @xmath74 .
the remaining 29 targets on our snapshot list are selected with the same spectroscopic algorithm from within the main galaxy sample of the sdss @xcite .
the main sample is much more heterogeneous , and we impose a quiescent , absorption - dominated spectral criterion by requiring our lens candidates to have rest - frame equivalent widths in h@xmath75 of @xmath761.5 ( with a few exceptions made in the interest of maximizing the total number discovered lenses in the program : see sdssj1251 - 0208 below . ) .
although most of the systems we target do indeed exhibit early - type morphology , the selection of slacs lens candidates from within the lrg and main galaxy samples of the sdss is purely spectroscopic .
our acs snapshot observing strategy is discussed in @xcite , and consists of one 420s wide - field channel ( wfc ) exposure through each of the two filters f435w and f814w .
the slacs lens candidates are selected to yield bright lenses with faint background sources .
this facilitates the photometric , morphological , and kinematic study of the lens galaxy , but can also make the relatively faint lensed features difficult to detect and even more difficult to use in lens modeling .
the key to success is effective subtraction of the image of the lens galaxy by fitting a smooth model to the image data of the candidate lens galaxy ( e.g. * ? ? ?
* ; * ? ? ?
given the extremely regular isophotal structure of most early - type galaxies , this is a reasonable proposition .
the most common parameterized model for early - type brightness distributions is the generalized de vaucouleurs or srsic law @xcite .
we first attempted to use elliptical srsic model - fitting and subtraction in our analysis , but generally found that the global systematic residuals of the fit were large compared to the surface - brightness levels of the lensed features that we hope to use to constrain gravitational - lens models .
significant galaxy - core residuals are especially pronounced in the srsic - subtracted f814w data .
the shortcomings of the srsic model led us to implement a more generalized galaxy - model fitting procedure involving a b - spline fit to the radial profile with a low - order multipole dependence to fit the angular structure , which we describe in appendix [ bsplineapp ] .
here we present the catalog of 28 candidate lens systems observed by the slacs survey through 2005 march 22 .
photometric and spectroscopic parameters for the sample as measured by sdss are given in table [ galtab ] .
figure [ specfig ] presents the sdss discovery spectra of all targeted systems , focused on the background - redshift line emission . for definiteness , we enumerate the several possible explanations for any one of our spectroscopically selected lens candidates as follows : ( 1 ) a multiply imaged background galaxy ( `` strong lens '' ) , ( 2 ) a singly imaged though possibly magnified background galaxy ( a `` non - lens '' in our classification ) , ( 3 ) a multiple - image system corresponding to a projection of multiple singly imaged sources with similar colors , and ( 4 ) spurious noise features in the sdss spectrum .
given the adopted significance threshold for selection and the careful treatment of the noise properties of sdss spectra ( b04 ) , we expect ( 4 ) to be an unlikely explanation , though it merits consideration in the case of any lower signal - to - noise candidates that do not produce detections in follow - up observations .
systems falling into category ( 3 ) could in principle be mistaken for bona fide strong lenses , but can be expected to have ill - fitting lens models , since real lenses occupy only a small subspace of all conceivable image configurations . for simplicity , this paper only categorizes acs - observed systems into `` strong lenses '' on one hand and `` everything else '' on the other , with `` everything else '' including systems for which a definitive explanation can not be made based on the data in hand ( i.e. acs non - detections ) . to determine the incidence of strong lensing , we examine the f814w and f435w residual images formed by subtracting smooth models of the foreground galaxy constructed as described in appendix [ bsplineapp ] .
if the residual image of a target shows multiple images with similar colors that can be reproduced by a simple lens model with the potential center fixed at the optical center of the foreground galaxy we classify the system as a strong gravitational lens .
the lens models for these systems are presented in paper iii , but we summarize the lens - modeling process here .
the procedure is based on the method of @xcite , with an implementation as described in @xcite and @xcite .
we parametrize the lens galaxy as a singular isothermal ellipsoid ( sie , * ? ? ?
* ) , with three free parameters corresponding to lens strength ( i.e. einstein radius ) , ellipticity , and major - axis position angle .
the source - plane brightness distribution of the candidate lensed galaxy is represented on a pixel grid .
we determine the best - fit lens parameters and source - plane pixel brightnesses by minimizing @xmath77 with respect to the galaxy - subtracted @xmath78-band image - plane data , while also imposing a brightness - dependent regularization on the source - plane distribution to suppress curvature .
unsuccessful strong - lens models are characterized by an inability to map putative multiply - imaged features to a single region of the source plane for any set of sie model parameters .
sdssj1025 is an example of a system with an apparent arc and counterimage which nonetheless can not be reproduced by an sie lens model , and is thus not classified as a strong lens .
acs data for target galaxies classified as lenses are shown in figure [ acsfig ] .
we note again that _ all _ target systems have line emission in their sdss spectra at a redshift higher than that of the foreground galaxy , and thus the lensing interpretation does not rest on imaging data alone .
figure [ maybefig ] presents all systems other than those that we classify solidly as lenses .
this figure encompasses systems that may indeed be strong lenses as well as systems that appear to be definite _ non_lenses .
we note that a significant number of systems that we consider possible but not definite lenses show a faint candidate counterimage near the center of the foreground galaxy opposite a more prominent image at larger radius . if these features were all due to simple residual error in the foreground galaxy subtraction , this configuration would not be expected .
the features could conceivably be explained by the prominent image `` pulling '' the foreground - galaxy model to the side , leading to an under - subtraction of the foreground - galaxy flux near the center on the opposite side .
we believe this explanation is unlikely since we have taken care to mask neighboring images when fitting the foreground - galaxy model ( appendix [ bsplineapp ] ) , and we do not see any corresponding over - subtraction on the near side of the core .
the most asymmetric double lenses will in general have faint counterimages at small angular offset from the lens center and will be the most difficult lenses to confirm , although their status as lenses or non - lenses can have a significant impact on statistical inferences based on the lens sample .
integral - field spectroscopy provides the best chance to associate background - redshift line emission with these faint putative counterimages near the lens center and thereby confirm or reject a strong - lensing hypothesis .
many of the lens galaxies shown in figure [ acsfig ] show such striking features in their residual images that their interpretation as strong lenses is effectively certain when considered in combination with their sdss spectroscopic detection and gravitational - lens models ( paper iii ) . for putative lenses with less dramatic morphology , further evidence as to their status
may be obtained with integral - field spectroscopy , which can confirm or deny the spatial coincidence between the high - redshift line emission and the candidate lensed features seen with _
hst_. for several of our slacs target galaxies , we have obtained spatially resolved spectroscopy with the integral - field units ( ifus ) of the inamori magellan areal camera and spectrograph ( imacs , * ? ? ?
* ; * ? ? ?
* ) on the 6.5 m walter baade
( magellan i ) telescope at las campanas observatory and the 8 m gemini - north multi - object spectrograph ( gmos - n , * ? ? ?
* ; * ? ? ?
* ) at mauna kea ; we present these data in figure [ slacs_ifu ] . in all cases ,
the ifu emission - line images are coincident with the acs residual - image morphology , supporting the strong - lensing interpretation .
we note here that the background sources in our lens systems are generally faint galaxies with irregular morphology .
as evidenced by the emission - line flux by which they are originally detected , they should also be regarded as starforming .
the absence of significant broadening of the balmer emission lines at the @xmath79kms@xmath3 resolution of the sdss further suggests that they do not host quasars .
for a recent observational study of the likely source population of emission - line galaxies , see @xcite .
the statistics of strong gravitational lensing are the subject of a great body of literature ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , most of it dealing with the lensing of quasars by galaxies .
the rigorous statistics of the slacs survey , though involving similar considerations , would be sufficiently distinct to require their own detailed treatment : our original sample is formed from potential lensing objects rather than from potentially lensed sources , and we must consider the effects of an extended source image and finite fiber sampling .
we defer a full lens - statistical analysis of the slacs survey to a future paper . here
we address the selection effects of the slacs survey in terms of the robust statistics of the distribution of lens - galaxy observables .
as discussed above , the slacs target selection involves an explicit velocity - dispersion and redshift selection .
this should not in itself present any complication in the comparison of our lenses to other galaxies with similar velocity dispersions and redshifts . after controlling for this selection
, we may ask whether our lenses are a representative sample of the populations from which they have been drawn : luminous red galaxies and quiescent main - sample galaxies from the sdss .
if so , their relation to early - type galaxies selected by other means is simply determined by the sdss lrg and main sample target selection function @xcite .
the question can be framed between two alternatives : do we select simply for spectral superposition caused by a chance foreground - background coincidence , or is there some significant bias by which galaxies with certain properties are more likely to be selected as lens candidates based on their sdss spectra ? to test this possibility
we exploit the parent sample of @xmath80 sdss galaxies ( early data release through data release 3 ) . for each lens
we construct a control sample of sdss galaxies with the same redshift and velocity dispersion ( within the uncertainties)quantities for which we explicitly select and test whether the lens galaxy is typical of galaxies in the control sample in terms of its magnitude , color , effective radius , and isophotal ellipticity
quantities for which we do not explicitly select .
although this is not a definitive test of the representative nature of our lenses ( it will not uncover any `` hidden variable '' bias ) , it is straightforward , robust , and informative . by forming control samples at the redshift of the lenses ,
we avoid the necessity of applying evolution- and @xmath81-corrections to the broadband magnitudes .
we may perform our comparison sensibly for galaxies with velocity dispersions well - measured by sdss .
we exclude the lens sdssj1251@xmath80208 from the analysis , both because its spectral signal - to - noise ratio is too low for a confident velocity - dispersion measurement and because its h@xmath75 equivalent width exceeds our quiescent threshold .
( in fact , it is a bulge - dominated spiral - galaxy lens ) .
the double - lens galaxies sdssj1618@xmath134353 and sdssj1718@xmath136424 are also excluded , as is the lens sdssj1205@xmath134910 , which has significant flux from a neighboring galaxy within the sdss spectroscopic fiber .
this leaves a well - defined sample of 15 single early - type lenses .
the comparison sample for each lens consists of unique galaxies from the sdss database with redshifts within @xmath82 of the lens redshift , velocity dispersions within @xmath83kms@xmath3 ( approximately the median velocity - dispersion error of the sample ) of the lens galaxy , and median signal - to - noise per spectral pixel greater than 8 . for lrg - sample targets
, we also require the comparison sample to pass the photometric lrg cuts , and for the main - sample lenses , we require a rest - frame @xmath84 .
we make a luminosity - distance correction to the broadband magnitudes and an angular - diameter - distance correction to the effective radii within the redshift slice to place all galaxies closer to the exact redshift of the lens .
the resulting individual comparison samples have as few as 17 and as many as 1793 galaxies , with a mean of 451 and a median of 337 control galaxies per lens galaxy .
we would like to answer the question of whether our lenses have observables ( luminosities , colors , effective radii , and ellipticities ) consistent with having been drawn from the distribution of those observables seen in the control samples . since we only have one lens at each redshift , a straightforward kolmogorov - smirnov test of the lens sample against the control samples is not possible : even under the null hypothesis , each individual lens has a different parent distribution .
we may however put all these parent distributions on common footing by using the fact that the test is invariant under a monotonic rescaling of the variable under consideration . specifically , for each individual lens s control sample , there exists a transformation of the observable of interest that converts the control distribution into a uniform distribution between the minimum and maximum values . the correspondingly transformed lens observable is simply equal to the normalized rank ( between 0 and 1 ) of the lens quantity within the cumulative distribution of the control sample in that quantity .
thus we may perform a one - dimensional test of the distribution of lens - observable ranks against a uniform distribution over the interval 0 to 1 .
this test is in some sense like a rank - correlation test within the formalism : does ranking as a lens correlate with ranking in luminosity , color , size , or ellipticity ?
figure [ ksplot ] shows the cumulative distributions of lens rankings in @xmath85 ( magnitude ) , @xmath86 ( color ) , @xmath87 ( effective radius ) , @xmath88 ( isophotal axis ratio ) , and @xmath89 ( @xmath85-band effective surface brightness : @xmath90 ) within their control samples . from these distributions
we compute the statistic @xmath91 , equal to the maximum difference between the cumulative rank distribution and the null - hypothesis distribution .
since the parent distribution under the null hypothesis is known exactly by construction ( uniform probability for any rank between 0 and 1 ) , and since the total number of control galaxies is much larger than the number of lenses , we compute the significance of @xmath91 for a distribution of 15 values against a known parent distribution ( e.g. * ? ? ?
( the statistical significance of an outlying lens will be limited by a smaller control sample in that its rank is at least @xmath92 and at most @xmath93 , where @xmath94 is the number of galaxies in its control sample . )
the resulting probabilities of the lens sample having been drawn at random from the control - sample distributions in @xmath85 , @xmath86 , @xmath87 , @xmath88 , and @xmath89 are 0.097 , 0.655 , 0.264 , 0.550 , and 0.085 respectively . in terms of color and axis
ratio , the slacs lenses seem to be a representative sample of the parent distribution of sdss galaxies . however ,
the distribution of slacs galaxies in brightness and effective radius ( and hence in surface brightness ) is somewhat significantly skewed towards brighter and more centrally concentrated systems .
this trend is in the same sense as that of the fp : at fixed velocity dispersion , brighter galaxies have smaller effective radii .
paper ii shows that the slacs lenses fall on the local fp corrected for stellar evolution , and further examines the degree of lens bias towards higher effective surface brightness in the context of the fp .
although the statistical significance is not great , the trend is at first glance suggestive of the effect that more centrally condensed objects are more efficient gravitational lenses ( e.g. * ? ? ?
* ) , to the extent that stellar mass constitutes the dominant lensing component .
a larger sample of similarly selected lenses will allow us to test this selection effect with greater statistical significance .
possible interpretations of this finding as discussed in detail in papers ii and iii include observational selection effects due to the finite size of the sdss fiber which may bias our survey towards the highest surface brightness lens galaxies .
in paper iii we will see that , within the context of powerlaw mass models , the slacs lens sample shows great homogeneity in the slope of the total mass density profile ( @xmath95 ) and no significant correlation between @xmath95 and @xmath96 .
this uniform degree of mass concetration within the lens sample itself argues somewhat against a lens - selection bias due to mass concentration , but this argument is limited by the lack of an equally direct probe of the mass distribution in the non - lens control sample .
the large number of new strong lenses confirmed by the slacs survey demonstrates the technical feasibility of carrying out spectroscopic surveys for strong gravitational lenses , and suggests that similar gravitational lens surveys should be considered as explicit goals of future large - scale spectroscopic surveys .
the slacs results are particularly notable when one considers that this type of lens survey was never one of the science or design goals of the sdss , and the survey is thus a testament to the performance of the sdss telescope and spectrograph , and to the optimal quality of the sdss spectroscopic data pipeline . with the clarity of hindsight , we may identify those particular features of the sdss that were beneficial to the spectroscopic lens survey project and those that could conceivably be improved in any future survey to increase the yield of spectroscopically selected strong lenses suitable for specified science goals .
the large parent sample of galaxies observed is a clear advantage afforded by the sdss for this project or any other rare object survey .
empirically , we have found that spectroscopic lenses and lens candidates appear with a frequency on the order of one in a thousand in the sdss : starting with @xmath80 early - type galaxies is therefore one key to success . at the same time , the large number of highly magnified galaxies that we have found close to the spectroscopic detection limit ( note the generally low emission - line signal - to - noise ratio in the discovery spectra of figure [ specfig ] ) suggests that we may only be seeing the tip of the iceberg .
we conclude that ( 1 ) a spectroscopic lens survey will always be at an advantage if it can `` piggy back '' on a more broadly conceived spectroscopic survey that does not have lenses as its sole scientific goal , and ( 2 ) future spectroscopic surveys that go significantly deeper than the sdss should , all other factors being equal , discover a significantly larger fraction of gravitational lenses
. one might suspect that a significant factor for slacs survey is the large ( 3arcsec diameter ) aperture of the sdss spectroscopic fibers .
certainly a larger fiber will always collect more photons , but for lenses with image splittings of @xmath97 ( i.e. @xmath98 einstein radii ) such as we have presented here , a 3arcsec fiber could afford an advantage over smaller fibers disproportionate to its larger aperture due to the concentration of lensed flux away from the center of the fiber .
we can test the importance of this aperture - size effect by simulating observations with varying fiber apertures .
we assume an einstein radius of @xmath99 ( an approximate median value from paper iii ) and an exponential - disk background galaxy with a disk scale length of @xmath100 .
we compute lensed images for a range of source - galaxy impact parameters using a singular isothermal sphere lens model .
we then smear the images with @xmath101 seeing ( the median sdss spectroscopic seeing for our targets ) and integrate over the 3arcsec diameter sdss fiber . if we reduce the fiber aperture to 2arcsec in diameter , we find that the flux collected from the lensed galaxy is reduced to a fraction of @xmath102@xmath103 , depending on impact parameter .
this is essentially the factor of @xmath104 corresponding to the diminished aperture size .
thus for lenses of the angular size typical of slacs lenses , the advantage of a large fiber is just the simple advantage of a large aperture due to the smearing effects of seeing ; a smaller fiber would not have lost a disproportionate amount of line flux due to lensing geometry . however , for lenses with larger image splittings ( which would occur for higher redshift sources behind the same foreground galaxies ) , a small fiber would likely be a qualitative disadvantage to a spectroscopic lens survey . because a fiber - diameter much greater than 3 arcsec would probably increase spurious non - lensed interlopers , we conclude that the sdss fiber - size represents a very good compromise , when searching for massive early - type lens galaxies at these redshifts .
for two principal reasons , our spectroscopic survey owes much success to having targeted lens candidates with background redshifts confirmed by multiple emission lines .
first , the incidence of false - positive spectroscopic detections and emission - line mis - identifications is negligible .
second , to fixed line - flux limits , intermediate - redshift emission from 3727 , 5007 , and the hydrogen balmer series is more common than lyman-@xmath75 emission at high redshift @xcite , and thus we see an abundance of oxygen / balmer lenses .
however , for optical surveys , this survey strategy limits the background redshifts to @xmath105 , beyond which h@xmath106 and 5007 move out of the observable band .
there is a definite incentive to discover significant numbers of lenses with higher lens and source redshifts , both to probe evolution in the lens population and to observe lensed images at larger physical radii within the lens galaxy in order to obtain greater leverage on dark - matter halos .
therefore in designing future surveys to discover gravitationally lensed emission - line galaxies at higher redshift , two obvious considerations are increased survey depth to detect a fainter population , and increased spectroscopic resolution beyond the @xmath107 resolution of the sdss in order to split the 3727 doublet and resolve the characteristic asymmetric profile of lyman-@xmath75 emission ( thus permitting more secure single - line redshifts )
. one of the greatest uncertainties at the outset of the slacs survey was simply the fraction of strong lenses within the candidate list .
one can attempt to calculate the lensing probabilities of the candidates given @xmath66 , @xmath67 , @xmath68 , assuming a model for the sdss spectroscopic observations , and assuming a luminosity function ( lf ) and size / shape for the background emission - line galaxies , as was done in @xcite , but the ingredients are all uncertain . to put it simply
, we did not know beforehand whether or not the majority of our targets would owe their high - redshift line emission to the psf - smeared wings of bright unlensed galaxies at large impact parameter rather than to einstein - ring images of faint galaxies closely aligned with the sdss target galaxy .
the high ratios of lenses to non - lenses and of quads / rings to doubles that we see are suggestive of a large magnification bias in our selection , with highly magnified faint lensed galaxies detected with greater frequency than less magnified lensed galaxies or unlensed projections .
future work will quantify the extent of any such magnification bias , and will derive statistical constraints on the lens and source populations based on the incidence of lensing within the survey .
the _ hst _ cycle13 slacs survey has produced a catalog of 19 previously unknown early - type strong gravitational lens galaxies .
the @xmath108% fraction of genuine lenses within the the first 28 candidates observed shows that the survey strategy
spectroscopic candidate selection from within the sdss , followed by _
hst_-acs snapshot observations of the systems with the largest estimated lensing cross sections is an efficient means for the discovery of new gravitational lenses ; similar lens surveys should be considered as an explicit goal of future spectroscopic surveys . by targeting galaxies with the
highest estimated lensing cross sections , the cycle13 slacs survey has effectively selected high - mass lensing galaxies ( strong lensing cross section is simply proportional to lens mass within the einstein radius ) .
the natural selection of lenses in other surveys is also weighted by lensing cross section , and thus strong lenses are typically massive galaxies . during _ hst _
cycle14 , we will extend our survey to galaxies of lower mass in order to use strong lensing and stellar dynamics to measure the mass dependence of early - type galaxy structure and mass - to - light ratio .
these galaxies will have a lower lensing cross section and hence should have a lower lensing rate , although magnification bias may skew the distribution of candidates in favor of lenses over non - lenses .
the final combined lens sample will be a unique resource for the detailed measurement of the mass profile of early type galaxies within the effective radius .
the slacs sample represents the largest single catalog of uniformly selected early - type gravitational lens galaxies assembled to date .
the lensed images provide an aperture mass constraint within a typical scale of @xmath109 .
the slacs lens galaxies are all brighter than their lensed background galaxies by many magnitudes , and are thus ideally suited to detailed photometric measurement as with the current acs data .
they are also excellent candidates for spatially resolved spectroscopy to constrain dynamical galaxy models in combination with the mass constraints from strong lensing .
in fact , nearly all of the slacs galaxies already have well - measured luminosity - weighted velocity dispersions inside the 3arcsec sdss fiber ( see table [ galtab ] ) , and thus we have already nearly tripled the number of known gravitational lenses with known stellar velocity dispersions .
paper iii will present the constraints on the logarithmic slope of the total radial density profile ( luminous plus dark matter ) that can be obtained by combining lens models with the aperture - integrated sdss velocity dispersions .
we may derive tighter constraints and a significant decomposition into luminous and dark components by obtaining spatially resolved line - of - sight velocity dispersion measurements within the lens galaxies ( see kt ) .
an analysis of the distribution of slacs lens - galaxy photometric and structural parameters within control samples from the sdss database shows that the slacs lenses are typical of their sdss parent sample with regard to color and ellipticity .
the location of slacs lens galaxies within the fp is the subject of paper ii .
however , the slacs lens galaxies exhibit a somewhat significant bias towards brighter and higher surface - brightness galaxies at fixed redshift and velocity dispersion ( [ sec : stats ] ; see also paper ii ) .
the significance of this bias can be confirmed with a larger sample of similarly selected lenses , which we plan to obtain with the continuation of our acs survey during _ hst _ cycle14
. results in this paper are based in part on data from the sloan digital sky survey ( sdss ) archive .
funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , princeton university , the united states naval observatory , and the university of washington .
these results are also based in part on observations obtained with the 6.5-m walter baade telescope of the magellan consortium at las campanas observatory .
these results are also based in part on observations obtained under program gn-2004a - q-5 at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the particle physics and astronomy research council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , cnpq ( brazil ) and conicet ( argentina ) .
these results are also based in part on observations made with the nasa / esa hubble space telescope , obtained at the space telescope science institute ( stsci ) , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these observations are associated with program # 10174 .
support for program # 10174 was provided by nasa through a grant from stsci .
the work of lam was carried out at jet propulsion laboratory , california institute of technology , under a contract with nasa .
the authors thank the anonymous referee for a careful reading of the manuscript and for a constructive report .
the b - spline technique is a well - known method for fitting a piecewise - defined polynomial of arbitrary order to the dependence of a series of data values upon an independent variable ( e.g. * ? ? ?
the coefficients of the polynomial change at breakpoints in the independent - variable domain , whose spacing may be chosen to allow more or less freedom depending upon the level of detail to be fit .
a b - spline of order @xmath110 ( where by convention @xmath111 is piecewise constant , @xmath112 is linear , @xmath113 is quadratic , @xmath114 is cubic , and so on ) has continuous derivatives to order @xmath115 across the breakpoints . for
our galaxy modeling we use b - splines of order @xmath114 , which have continuous 0th , 1st , and 2nd derivatives .
the coefficients of the b - spline are determined by fitting to the data in a least - squares sense , given the breakpoint spacing and derivative - continuity conditions .
a b - spline model may be reinterpreted in terms of a number of localized basis functions within the domain whose shapes are set by the order of the b - spline and by the breakpoint spacing and whose amplitudes are determined by the fit to the data .
the fit itself is entirely linear , and the localized influence of each basis function within the domain implies that only a banded - diagonal matrix need be inverted in the solution for the coefficients .
the radial b - spline technique for galaxy images permits smooth fitting of arbitrary radial brightness profiles .
since we do not have multiple dithered exposures to combine , we perform our b - spline galaxy - model fitting in the native acs pixel coordinates of the images .
we account for the distortion in the acs by using the solution provided in the image headers to compute relative tangent - plane ra and dec values for all pixels in the image , which are taken as the independent variables for the fit .
we fit for surface brightness as a function of position using cosmic - ray - masked flat - fielded images ( i.e. calibrated by the calacs software pipeline to measure surface brightness rather than counts - per - pixel ) , so our fits are not biased by slight variations in pixel area across the images .
the one non - linear step in our b - spline galaxy - model fitting is the determination of the center of the lensing galaxy ; we describe our centering method further below .
the model - fitting procedure is carried out using the adopted center for the lens galaxy within a suitable subsection of the acs field ( typically @xmath116 ) , along with an error image , a mask specifying cosmic - ray and other zero - weight pixels , and a mask corresponding to stars , neighboring galaxies , and any apparent background - galaxy features .
the cosmic - ray masks are initially generated by the lacosmic software @xcite and adjusted manually over the image subsection surrounding the lens ; the neighboring - object masks are created manually . for each pixel in the sub - image ,
the radial offset @xmath117 from the galaxy center ( in arcseconds ) and the azimuthal angle @xmath118 relative to a fixed position angle are computed .
a set of breakpoints in @xmath117 is chosen for the fit , typically every @xmath119@xmath120 in the central @xmath121 to @xmath122 , with increased spacing further out .
( we note again that the acs pixel scale is 005 per pixel . ) for the case of complete einstein rings , breakpoints at the ring radius can be removed to allow smoother interpolation of the lens - galaxy model over the ring region .
a multipole angular dependence is incorporated into the fit as follows .
if the purely one - dimensional b - spline fit is represented as @xmath123 where the @xmath124 are the localized basis functions that are only non - zero over a small range of breakpoints and the @xmath125 are their amplitudes , then the two - dimensional fit is represented by @xmath126 f_k ( r ) ~~.\ ] ] the number of multipole orders to be fit is chosen individually for each galaxy , with @xmath127 ( monopole ) and @xmath128 ( quadrupole ) always present and higher - order terms added if necessitated by systematic angular structure in the residual images . with the occasional exception of an @xmath129 ( dipole ) term to capture slightly disturbed morphologies or mild mis - centering in the central regions ,
only even multipole orders are used .
this form of angular dependence is ideal for fitting early - type galaxies : the fit remains linear , the global symmetries seen in early - type galaxies are naturally reflected in the low - order terms , and effects such as isophotal twists , varying ellipticity with radius , and diskiness / boxyness can be captured with minimal effort ( with only the last of these effects requiring multipole orders beyond the quadrupole ) .
we note also that the monopole term can automatically include a fit to the sky background . following the completion of the initial b - spline model - fitting step ( which for our @xmath130 , 240@xmath131240-pixel images generally takes 23 seconds on a 2.53ghz pentium 4 linux pc ) ,
we examine the residual ( data@xmath8model ) image for faint features not associated with the smooth galaxy model and not initially masked , and perform a second fit with an updated mask , also adding multipole orders as needed .
since at this stage we are principally concerned with generating high - quality residual images and not with the measurement of lens - galaxy parameters , we fit directly to the images without convolving the b - spline model with the _ hst _ point - spread function . as an example
, we describe the fit to the sky - subtracted image of the e3 lens galaxy sdssj0912@xmath130029 to give the f814 residual image shown in figure [ acsfig ] , which uses breakpoints spaced every @xmath120 in @xmath117 and multipole terms of order 0 , 2 , and 4 ( monopole , quadrupole , and octopole ) . the ratio of the @xmath128 amplitudes ( defined as @xmath132 ) to the @xmath127 amplitudes rises from @xmath133 in the center to @xmath134 at the effective radius ( about @xmath135 ) , and the ratio of the @xmath136 to the @xmath127 amplitudes rises from @xmath137 to @xmath138 over the same range .
some of this increase in angular structure with radius is due to our fitting directly to the psf - blurred data .
the centering necessary for the b - spline galaxy fit is accomplished by first fitting with respect to a best - guess center with monopole , dipole , and quadrupole terms .
this fit is then evaluated in a ring and used to determine a flux centroid , which is adopted as a new best - guess center for another iteration of the same procedure . within a few iterations this process which amounts to minimizing the dipole term
converges to a position that we adopt as the constrained center for the final model fit .
the angular resolution of _ hst _ reveals this system to be a pair of foreground galaxies .
the pair lenses a compact background source into a 3@xmath131 quad configuration .
the extra compact image to the west of the counterimage is more blue in color than the lensed images and thus does not pertain to the lensed configuration .
the imaging depth is not great enough to unambiguously show whether the putative lensed features appear in both bands , but the general geometry is plausibly similar to j0912 + 0029 , which has an extremely solid gravitational lens model ( paper iii ) .
the curvature of the arcs is suggestive of ring geometry , and the knot - counterknot features along the north - south axis may be images of each other .
deeper observations would reveal whether or not this system is a real einstein ring , but it is not certain from the current data . ,
b. c. , & dressler , a. m. 2003 , in instrument design and performance for optical / infrared ground - based telescopes . edited by iye , masanori ; moorwood , alan f. m. proceedings of the spie , volume 4841 , pp .
1727 - 1738 ( 2003 ) .
, 17271738 , i. , et al . 2003 , in instrument design and performance for optical / infrared ground - based telescopes . edited by iye , masanori ; moorwood , alan f. m. proceedings of the spie , volume 4841 , pp .
1645 - 1656 ( 2003 ) . ,
16451656 , g. j. , et al .
2003 , in instrument design and performance for optical / infrared ground - based telescopes . edited by iye , masanori ;
moorwood , alan f. m. proceedings of the spie , volume 4841 , pp .
1750 - 1759 ( 2003 ) . , 17501759 , j. , dodsworth , g. n. , content , r. , & allington - smith , j. r. 2004 , in uv and gamma - ray space telescope systems .
edited by hasinger , g " unther ; turner , martin j. l. proceedings of the spie , volume 5492 , pp .
624 - 633 ( 2004 ) . , | the sloan lens acs ( slacs ) survey is an efficient _ hubble space telescope _ ( _ hst _ ) snapshot imaging survey for new galaxy - scale strong gravitational lenses .
the targeted lens candidates are selected spectroscopically from within the sloan digital sky survey ( sdss ) database of galaxy spectra for having multiple nebular emission lines at a redshift significantly higher than that of the sdss target galaxy .
the slacs survey is optimized to detect bright early - type lens galaxies with faint lensed sources , in order to increase the sample of known gravitational lenses suitable for detailed lensing , photometric , and dynamical modeling . in this paper ,
the first in a series on the current results of our _ hst _ cycle13 imaging survey , we present a catalog of 19 newly discovered gravitational lenses , along with 9 other observed candidate systems that are either possible lenses , non - lenses , or non - detections .
the survey efficiency is thus @xmath0% .
we also present gemini 8 m and magellan 6.5 m integral - field spectroscopic data for 9 of the slacs targets , which further support the lensing interpretation .
a new method for the effective subtraction of foreground galaxy images to reveal faint background features is presented .
we show that the slacs lens galaxies have colors and ellipticities typical of the spectroscopic parent sample from which they are drawn ( sdss luminous red galaxies and quiescent main - sample galaxies ) , but are somewhat brighter and more centrally concentrated .
several explanations for the latter bias are suggested .
the slacs survey provides the first statistically significant and homogeneously selected sample of bright early - type lens galaxies , furnishing a powerful probe of the structure of early - type galaxies within the half - light radius .
the high confirmation rate of lenses in the slacs survey suggests consideration of spectroscopic lens discovery as an explicit science goal of future spectroscopic galaxy surveys . |
quoting from the final p5 report@xcite , of the particle physics project prioritization panel ( p5 ) , `` many extensions of the standard model , including supersymmetry , have additional sources of cp non - conservation . among the most powerful probes of new physics that does not conserve cp are the electric dipole moments ( edm s ) of the neutron , electron and proton . '' various schemes for using frozen spin storage rings to measure electric dipole moments ( edm ) of various baryons have been suggested .
an early design first proposed by farley and others@xcite has , by now , evolved into a design by the international storage ring edm collaboration@xcite ; important steps in this evolution are described in various reports and papers .
they include : overall strategy , semertzidis@xcite ; polarimetry , stephenson@xcite ; squid magnetometry , kawall@xcite ; magnetic shielding , morse@xcite ; error compensation , orlov@xcite ; lattice design , j. and r. talman@xcite@xcite ; spin coherence time ( sct ) , haciomeroglu@xcite and talmans@xcite .
the main recent experimental advances toward the goal of storage ring edm measurement have occurred at the cosy storage ring in juelich , germany , by lehrach@xcite , rathmann , stephenson and others , and described , for example , by stroeher@xcite@xcite .
the storage ring edm measurement proposed here has continued this evolution with major modifications .
the most important modification is the introduction of resonant polarimetry , which permits the edm measurement to be `` moved into the frequency domain '' . this , and other significant differences ,
are spelled out in detail in what follows .
much of the early discussion assumes electrons , but most results are common to both electrons and protons ; even deuterons in some cases .
the current upper limit@xcite@xcite for the electron edm is about 10 in our nominal @xmath1e - cm edm units .
this measurement , using thorium oxide molecules , exploits a 9 order of magnification factor of internal to applied electric field for this molecule , helped also by a conveniently small molecular mdm . by constrast
our storage ring trap measurement will measure the free electron edm , using a much larger external electric field , but without benefit of the molecular polarizability magnification factor .
the expected instrumental precision ( i.e. not including systematic errors ) corresponds to an edm value of @xmath0e - cm .
this is 100 times smaller than the current electron edm upper limit , but with systematic errors not yet included .
( for the proton the expected edm precision is about the same , and is almost 6 orders of magnitude smaller than the current proton upper limit . )
it would not be realistic to claim , at this stage , that all systematic errors at this level of precision have been identified , much less eliminated .
however , even if the @xmath0 precision is overly optimistic for ultimate accuracy , _ high precision is important , since it establishes the precision with which systematic errors can be investigated and ameliorated .
_ the most promising possibility , as regards physics reach , continues to be measuring the edm of the _ proton _ using an all - electric ring .
it is important to reverse the beam circulation direction frequently , with no ring modification ( other than injection direction ) .
this is required for a significant reduction in systematic error . in previous proposals , such as reference@xcite
, the two beams are required to counter - circulate simultaneously , as in a colliding beam ( with `` collisions '' calculated to have negligible effect on the measurement ) . in the present proposal
just one beam circulates at a time , but with frequent circulation reversals .
the justification for dispensing with simultaneous beams is explained in connection with discussion of magnetic shielding , the purpose for which is to suppress the spurious edm signal caused by poorly - known radial magnetic fields .
it has been realized only quite recently that charged hadron particle edm s can be directly measured with accuracy comparable to or , actually , far better than , neutron edm s .
in particular there have been a continuing series of neutron measurements over time , but no direct proton measurements .
the history of neutron edm measurement , up to 1982 , is described ( and very clearly explained ) by ramsey@xcite .
modern neutron edm measurements use the so - called `` ramsey method of separated oscillatory fields '' .
this method is discussed briefly in a later section on magnetic field shielding , needed to reduce a source of systematic error common to neutron and charged particle edm measurement .
the analogy is much closer than magnetic shielding , however , since , like the proposed rolling spin method , the ramsey neutron method can be regarded as shifting the edm sensitivity `` into the frequency domain '' . at their `` magic '' kinetic energies in an all - electric storage ring , 14.5mev kinetic energy for electrons , 235mev for protons
, the beam polarization precesses at exactly the same rate as the beam momentum .
when viewed at any fixed point in the ring , the polarization appears `` frozen '' , for example always parallel to the beam orbit . by intentionally superimposing a (
precisely - controlled ) torque around the local radial @xmath2-axis , the beam polarization will `` roll '' around that axis at frequency @xmath3 .
the first of the two purposes for the roll is to make resonant polarimetry possible .
a resonant polarimeter responds directly to the longitudinal magnetization of a polarized beam . with beam polarization
frozen forward , the magnetization signal consists of harmonics of the revolution frequency . as such , the magetization signal would compete unfavorably with direct coulomb excitation of the resonator by the beam charge , no matter how carefully the resonator is designed to be insensitive to that excitation .
the rolling polarization alteration ( first introduced by koop , for different reasons , as a `` spin wheel '' ) overcomes this by shifting the magnetization signal frequency far enough away from revolution harmonics for the resonator quality factor @xmath4 to be high enough to suppress the direct resonator response to the passing beam charge . with ideal stabilization and the magic condition
exactly satisfied , the beam polarization stays always in the local @xmath5 plane where @xmath6 is `` up '' and @xmath7 is longitudinal . in this condition ,
any torque due to the electric bending field acting on the electron edm also lies in the @xmath5 plane and causes a systematic shift in the roll frequency .
_ it is this frequency shift that is to be measured and ascribed to the electron edm .
shifting the edm sensing from polarimeter amplitude to polarimeter phase greatly improves the ultimately - achievable edm precision .
_ both the longitudinal component of beam polarization vector and the roll frequency will be measured using a longitudinally - aligned , resonant polarimeter , sensitive to the longitudinal beam polarization . as discussed in reference@xcite , and in greater detail in a paper under preparation
, this polarimeter responds linearly to the slowly oscillating @xmath7-component of beam magnetization , and is insensitive to both radial @xmath2 and vertical @xmath6 polarization components , as well as being insensitive to beam charge .
( of course the magnetization signal _ is _ proportional to the beam charge .
it is only because the polarized beam charge can be large in a storage ring ( @xmath8 particles , more or less aligned , for example ) that makes resonant polarimetry practical . ) to keep the spin wheel properly aligned will require phase - locked loop stabilization of the other two beam polarization degrees of freedom .
transverse components of beam polarization will be fed back to provide the required stabilizing torques . for protons
this polarimetry can be performed by measuring both the left - right and up - down scattering asymmetries of extracted particles .
alternatively , resonant polarimeters sensitive to transverse components of beam polarization could be used .
such phase - locking has not yet been achieved in practice .
but the jedi collaboration@xcite at the cosy ring in juelich germany , appears to be on the verge of performing this feat .
the proposed storage ring layout is shown in figure [ fig : singlering ] .
spin wheel , polarimeters , and stabilizers are shown .
feedback to a wien filter holds the roll plane perpendicular to the local radial axis by providing left - right `` steering stabilization '' ( though it is the polarization rather than the orbit that is being steered ) .
this amounts to holding the average beam energy exactly on the magic value .
the design lattice has no intentional solenoids .
however a trim solenoid will be required to cancel possible solenoidal fringe field components .
this solenoid also provides the `` wheel - upright '' stabilization torque .
the stabilizing torques are shown in figure [ fig : by - out - steer ] .
even if the proton edm measurement is ultimately more promising as `` physics '' , the electron edm measurement is also important . and it is sufficiently similar for electron and proton measurements to be described together in this paper
. also , because the electron magic momentum is so much smaller , the electron ring will be much cheaper .
a sensible first step might therefore be to first build , as a prototype , an all - electric , frozen - spin storage ring for electrons of kinetic energy @xmath9mev ( which is the `` magic energy '' at which the electron spin is `` frozen '' because the `` spin tune '' @xmath10 vanishes ) .
as well as serving as a prototype for an eventual proton ring , such an electric ring could also be used to measure the electron edm with unprecedented precision , and to develop techniques applicable to the proton measurement .
the spins in a beam of deuterons can not be frozen in an all - electric ring ; superimposed electric and magnetic fields are required to cancel the spin tune .
the presence of magnetic bending excludes the possibility of simultaneously counter - circulating beams .
since counter - cirulating beams are not required to be simultaneous for the edm measurement method proposed in the present paper , the same method will be applicable for deuterons , with the only added complication being that reversing the beam circulation direction will require reversing the magnetic fields . nevertheless , for brevity
, the deuteron measurement is not discussed in this paper , except for pursuing a suggestion by senichev@xcite , concerning the possibility of a deuteron ring with disjoint electric and magnetic bend sectors .
this is pursued later in the section titled `` geometric phase errors '' .
i. a. koop@xcite has proposed a `` spin wheel '' in which , by applying a radial magnetic field , the beam polarization rolls in a vertical plane with a frequency in the range from 0.1 to 1 hz .
the rolling polarization i propose is similar , except my proposed roll frequency is greater by a factor as great as one thousand .
as shown in the upper left corner of figure [ fig : singlering ] , the beam polarization `` rolls '' at uniform rate in the plane defined by the vertical @xmath6 and ( local ) longitudinal @xmath7-axes .
this rolling action causes the longitudinal polarization , which is what the polarimeter senses , to vary sinusoidally with frequency @xmath11 .
@xmath3 will be adjustable , to a value such as 100hz .
the ring revolution frequency is @xmath12hz for electrons .
the polarimeter natural frequency will be tuned to one or the other sidebands of harmonic number @xmath13 times the revolution frequency . @xmath14 where the roll can be either `` forward '' or `` backward '' .
the polarimeters have high @xmath15-values , such as @xmath16 and fractional frequency selectivity of about @xmath17 . in theory
the polarimeter is sensitive primarily to beam magnetization and not to beam charge .
however it is anticipated that , if tuned exactly to any harmonic of @xmath18 , excitation due to direct beam charge or beam current would be likely to dominate the polarimeter response .
the rolling of the polarization shifts the polarization response frequency by an amount large compared to the resonator selectivity .
it is also essential for the rolling polarization to not suppress the edm signal .
this is only possible if the bunch polarization stays in a plane normal to the electric field , which is the source of the edm torque .
since the electric field is radial ( @xmath2 ) , the roll plane has to be vertical / longitudinal ( @xmath19 ) .
the effect of the edm torque is then to alter the rate of roll .
the instrumentation has to distill this `` foreground '' effect from the `` background '' of intentional ( and unintentional ) sources of roll .
just mentioned at this point , and discussed later in connection with suppressing systematic errors , it can be mentioned that the rolling polarization will also have the beneficial effect of helping torques due to field errors to average to zero over times much less than one second .
the ring is racetrack - shaped with most of the instrumentation in the long straights .
`` wien filters '' are crossed electric and magnetic fields which cause no beam deflection . with no deflection in the laboratory
there is no electric field in the electron rest frame . as a result
these elements apply zero torque to the electron s edm .
the horizontal and vertical wien filter strengths are @xmath20 and @xmath21 . in a wien filter ( stripline terminated by appropriate resistor ) such as @xmath20 , both the electric and magnetic fields are produced by the same current , @xmath22 .
@xmath23 the electric field is produced by the voltage in the precision terminating resistor carrying the current @xmath24 .
the remaining stabilization field is @xmath25 which is the field of a solenoid aligned longitudinally . in all cases , though labelled as if magnetic fields , it is actually the currents through these elements that control the applied torques . as with the wien filters ,
the solenoid strength is established by a single current .
the beam magnetization observed at a fixed point in the ring consists of a `` comb '' of periodic time domain impulses at the beam revolution frequency @xmath18 , but with pulse amplitude varying sinusoidally at the roll frequency .
the frequency domain representation of the magnetization then consists of upper and lower sidebands of the frequency domain comb of all harmonics of the revolution frequency . a resonant polarimeter sensitive to longitudinal polarization and with axis aligned along @xmath7 and tuned to frequency @xmath26 , where @xmath13 is an integer harmonic number , will respond to the upper or lower sideband excitation signal .
even when designed to be insensitive to direct excitation due to beam charge or beam current , a polarimeter tuned to any harmonic of the revolution frequency would likely be overwhelmed by direct coulomb excitation .
the roll frequency therefore has to be high enough , and the quality factor @xmath4 of the resonator large enough , to reject background response due directly to the beam charge or the beam current .
this is what makes the rolling beam polarization essential for the edm measurement .
only upper limits are presently known for the electric dipole moments of fundamental particles .
as regards physics , this is unsatisfactory . as regards describing the apparatus to be used in measuring edm s
it is a nuisance . to simplify discussion
it has become conventional to define a nominal edm value and define @xmath27 to be the edm expressed in units of the nominal edm .
magnetic parameters are given for electrons and protons in table [ tbl : magneticparams ] .
for an ideal , non - relativistic dirac particle in a uniform magnetic field the momentum precession rate and the spin precession rate are equal all energies are `` magic '' . for a non - ideal particle with anomalous magnetic moment @xmath28 and magneton value @xmath29 , the difference between these precession rates is the anomalous precession rate @xmath30 . by a numerical coincidence though the electron s magnetic moment is three orders of magnitude greater than the proton s , its anomalous magnetic moment is three orders of magnitude less the anomalous precession rates for electron and proton having the same relativistic @xmath31 factor are roughly the same .
the same rough equality holds also with electric bending , with @xmath32 replaced by @xmath33 .
this provides a handy mnemonic when electron and proton edm rings are being contemplated at the same time .
any precession in a storage ring that is due to a particle s edm competes with the anomalous precession due to its mdm , which is roughly the same for electrons and protons . for assessing their relative importance , one can compare the absolute precession due to its edm with the anomalous precession due to its mdm . .[tbl
: magneticparams]e and p magnetic parameters .
the units are si , but with energies expressed in ev .
anomalous precession rates are in angular units of radians / second . in some cases
the numbers are given to many - many places to indicate the accuracy to which the values are known , _ not _ to indicate the values should be trusted as authoratively correct . [ cols="^,^,^,^,^",options="header " , ] the resonant polarimeter theoretical analysis seems solid .
nevertheless , the method has not yet proven out experimentally .
obviously the lack of experimental experience with resonant polarimetry introduces major uncertainty into the design of an experiment for which it is the main component .
one thing is certain though ; the signal level , relying as it does on stern - gerlach force , will be small . both derbenev s and my paper calculate signal to noise ratios , assuming the noise to be dominated by thermal noise at the resonator ambient temperature .
both conclude that high - q , cryogenic , superconducting resonators are needed to bring the signal convincingly out of the noise .
certainly stored polarized beams with as many as @xmath34 particles should give clean resonator output signals . for the electron edm experiment
the predicted signal level is very safely above the thermal noise level .
this is also true for the proton edm experiment , though less so because of the smaller proton mdm .
this section analyses the influence of the noise on the polarimetry and produces an estimate of the resulting precision limit .
consider a particle circulating in a storage ring magnetic field with rotation frequency @xmath18 . applying the @xmath35 ratio calculated previously ,
the edm - induced precession frequency is @xmath36 . with revolution frequency of @xmath3710mhz ,
this is @xmath38r / s .
0ne has to plan on measuring a `` nominal '' edm - induced precession of order @xmath39r / s , or about 3mr / day .
as a vector angle to be measured this seems pretty small but , as a clock phase angle in a modern , high precision clock , it is not so small .
a virtue of frequency is that its measurement proceeds by counting cycles , or even fractions of cycles ; for example @xmath40 , if one tenth percent of a cycle can be distinguished . here
a cycle is being referred to as a `` fringe '' as that term is applied in optical length measurements .
it will be convenient for data analysis to allow the number of fractional fringes @xmath41 to be non - integer and to interpret @xmath42 as a statistical error parameter ( such as r.m.s .
value ) made in determining @xmath41 .
one can then refer to @xmath41 as the digitised polarimeter phase advance , expressed in units of the statistical error parameter .
modern technology has enabled extremely accurate frequency comparison , for example as described for a commercial device by kramer and klische@xcite .
they use the `` allan variance '' @xmath43 as a quantitative measure of the fractional frequency variance after measuring for time @xmath44 . for the multi - mhz range we are interested in , their apparatus achieves @xmath45 for runs of the duration @xmath44=1000s we are assuming .
they characterize this performance as `` better than 0.2 degrees related to 10 mhz '' , though allan variance @xmath46 is not simply converible into our @xmath42 parameter , this corresponds roughly to @xmath47 .
this suggests our choice of @xmath42=0.001 , for a numerical example , is conservative .
this conservativism reflects our current lack of understanding of the phase noise that limits the achievable precision .
kramer and klische@xcite suggest using coherent detection for further noise reduction . this technique would be made available to us using the coherent responses of multiple polarimeters .
the resonant polarimeter frequency is necessarily quite close to a harmonic of the revolution frequency @xmath18 . in other words
@xmath48 is approximately equal to an integer , such as 10 .
the rate at which fractional edm fringes accrue in the polarimeter is then @xmath49 in a pair of runs , each of duration @xmath44 , the total number of fractional fringe shifts , after time @xmath50 , is @xmath51 where @xmath52 is the electric dipole moment in units of @xmath53e - cm , and @xmath54 . in this numerical example , with @xmath55 , after 2000s , @xmath52 will have been measured with one unit of statistical precision . by averaging over some number of runs , @xmath56 , or by performing longer runs one will have measured @xmath52 with higher precision .
exactly how the precision scales with @xmath56 depends on how we interpret the @xmath42 parameter .
the most optimistic assumption possible is that there is no random run - to - run error whatsoever , in which case the precision improves proportional to @xmath56 .
realistically , one expects some random run - to - run error . we can choose to interpret @xmath42 as defining the r.m.s .
run - to - run uncertainty , in which case the precision improves only proportional to @xmath57 .
this is a pessimistic result . even in this case
it will remain necessary to determine @xmath42 phenomenologically .
as has been mentioned already , the dominant contributor to @xmath42 may be irreducible phase noise in the resonant polarimeter . to properly reduce the ambiguity between optimistic and pessimistic approaches
it would be necessary to replace @xmath42 by two independent statistical parameters . but this will be quite difficult , for example because the split will depend on the run length and other uncertainties . to be conservative
we continue by simply following the pessimistic route . continuing our earlier numerical example , and expecting to make @xmath58 pairs of runs ( over about one year ) one will achieve a one sigma edm precision of @xmath0e - cm .
the dominant fields in an electron storage ring are radial lab frame electric @xmath59 and/or vertical lab magnetic field @xmath60 .
they give@xcite transverse electron rest frame field vectors @xmath61 and @xmath62 , and longitudinal electric and magnetic components @xmath63 and @xmath64 , all related by @xmath65 a particle in its rest system , with angular momentum @xmath66 and magnetic dipole moment @xmath67 , in magnetic field @xmath62 , is subject to torque @xmath68 . here
@xmath29 is the `` magneton '' value for a particle of that particular mass and charge . by newton s angular equation , substituting from eq .
( [ eq : lorentz.3b ] ) to express the fields in laboratory coordinates , @xmath69 customarily the rest frame angular momentum is represented by @xmath70 rather than by @xmath66 ( because @xmath70 is a true 3-vector only in that frame ) and the laboratory time interval @xmath71 is used instead of rest frame time interval @xmath72 . furthermore , the magnitude @xmath73 is known to be constant . with the laboratory field being purely electric and purely radial , the vertical component of @xmath70 is conserved . the normalized horizontal component
@xmath74 satisfies @xmath75 as jackson explains@xcite , relativistic effects cause the electron axis to precess in the laboratory , irrespective of any static moments the electron may have .
this thomas precession causes the polarization vector to precess even if there is no torque acting on the magnetic or electric moments .
this precession has to be allowed for when ascribing precession to mdm s or edm s . in our case , the beam direction advances uniformly , by @xmath76 during one revolution period @xmath77 and the thomas precession term is @xmath78 adding this term to the electric part on the rhs of eq .
( [ eq : newton.1 ] ) , @xmath79 this agrees with jackson s eq . ( 11.170 ) .
we try a solution of the form @xmath80 for circular motion in an electric field , @xmath81 , which leads to @xmath82 accelerator spin physicists define the `` spin tune '' @xmath83 in an electric field by @xmath84 where @xmath85 is the angle between spin vector and particle velocity .
the two definitions of `` spin tune '' have therefore been inconsistent . unlike the beam direction , which advances by @xmath76 each turn , and
could therefore be described as having a `` tune '' value of 1 , the spin tune is reckoned relative to the particle velocity rather than relative to a frame fixed in the laboratory .
this accounts for the `` + 1 '' on the rhs of eq .
( [ eq : newton.10 ] ) . )
the content of this section has therefore amounted to being a derivation of eq .
( [ eq : spinprecess.4 m ] ) , for the spin tune in an electric ring .
a particle at rest , with angular momentum @xmath70 and electric dipole moment @xmath86 , in electric field @xmath61 , is subject to torque @xmath87 . using eq .
( [ eq : lorentz.3a ] ) , if the laboratory field is purely electric , the induced precession satisfies @xmath88 this precession is small enough to be treated as a perturbative addition to otherwise - inexorable polarization evolution . to calculate edm - induced precession
we can use evolution formulas to describe dependence on @xmath89 , of the beam angle , which advances by @xmath76 every turn . with @xmath90 being cartesian coordinates fixed in the lab , setting @xmath91 , as appropriate for frozen spin motion , the frenet and spin vectors advance as @xmath92 here @xmath93 is the polar angle of the polarization relative to the vertical axis . substituting these on the right hand side of eq .
( [ eq : edmprecssion.1 ] ) , the evolution of @xmath94 is the driven response given by @xmath95
the achilles heel of storage ring frozen spin edm measurement has , until now , been spurious precession due to any non - zero radial magnetic field average @xmath97 .
acting on each frozen particle s mdm , this magnetic field applies torque that perfectly mimics the edm effect . without rolling polarization ,
the whole determination of systematic error boils down to finding the maximum spurious edm precession that can come from time - dependent variation of @xmath96 . starting with a discussion of frozen spin operation , much of this section is devoted to showing that rolling polarization largely eliminates this source of systematic edm error . in the bnl proposed@xcite
proton edm experiment this precession error is limited by precision bpm - measured cancellation of the vertical displacement of simultaneously counter - circulating , exactly superimposed beams . unlike spin precession due to @xmath97 , which is independent of beam direction
, there is a differential displacement of the counter - circulating beams , one up , one down .
cancelling this differential displacement cancels both the spurious mdm - induced precession and the average magnetic fields produced by the two beams , not just on the beam centerline , but also nearby .
using ultra precise squid beam position monitors , located as near as possible to the beam orbits , this method is explained by kawall@xcite .
the produced signals are proportional to the vertical beam separation .
the effectiveness of this approach depends on the coherent summing of field amplitudes in each one of many bpm s , along with the `` null '' beam centering this enables .
according to that proposal , this procedure limits the systematic edm accuracy from this source to about @xmath98e - cm after running for one year .
a goal for the present proposal has been to reach at least the same level of accuracy , but without depending on simultaneously circulating beams .
the rolling polarization method renders this cw / ccw beam reversal less essential .
the new method continues to use counter - circulating beams , but they circulate consecutively , not concurrently .
this produces many of the same cancelations , but lack of simultaneous beams is less effective for two reasons .
one is that the vertical beam displacement is being measured as the small difference of two positions , each of which is being measured using lower precision bpm s ( because of their larger required dynamic range ) .
the other is that , to the extent the magnetic field errors depend on time , their cancelation is impaired by measuring them at separate times . to be able to claim comparable insensitivity to @xmath96 errors , without depending on simultaneously circulating beams
, we have to overcome this loss of edm selectivity .
a `` selectivity factor , s.f . '' will be used to reduce tedious circumlocution in the following discussion .
small s.f . is good , s.f.=0 is perfect . for ultimate precision one will , in any case , try to situate the ring away from unpredictable magnetic sources in a city . with active and passive shielding ,
fields of several ft/@xmath99 ( femto - tesla per root hertz ) have been obtained in shielded rooms in city environments . averaged over a 1000s run ,
this gives about @xmath100 t .
it is essentially impossible for this field to be always radial in the storage ring .
averaging over the full ring is likely to give a reduction factor of perhaps 10 .
the magnetic field error is in competition with the `` magnetic equivalent '' of an electric field of roughly @xmath101v / m .
dividing by @xmath102 , this is equivalent to @xmath103 t .
the ratio @xmath104 has to be compared to the relative effectiveness factor @xmath105 . by this estimate
the spurious mdm signal is 10 times greater than the nominal edm signal .
we have been copying , so far , from reference@xcite , assuming the beam polarization to be truly frozen .
this is no longer the case in the present proposal .
now the beam polarization rolls with a frequency of , say , @xmath106hz . even if this frequency were imposed by a control voltage oscillating at frequency @xmath3 , it would produce substantial reduction of the spurious precession caused by @xmath96 .
but any correlation between roll - torque and roll - phase would limit the edm selectivity improvement .
in fact , the polarization roll - frequency is `` autonomously '' self - generated from an applied dc control current @xmath24 .
there is no source for the rolling polarization to be coherent with . as a result the spurious torque due to @xmath96
will truly average to zero over times of order @xmath107=10milliseconds .
furthermore , even if , by chance , there were accidental synchronism at one roll frequency , it would not be present at another .
this rolling polarization improvement has not come without cost . in our proposed method ,
the edm measurement comes from differencing forward and backward roll rates . canceling the effect of external magnetic fields has come at the cost of introducing new possible sources of error via this subtraction .
this will be addressed shortly .
the discussion so far is largely applicable also to time - dependent @xmath108 fields generated inside a shielded room or tunnel .
current leakage or sparking would be candidate sources .
permanent magnetization would not be serious , but hysteretic or temperature dependent magnetization could be .
repeated precision measurements with the same circulation direction for both beams will give reliable information concerning the time - dependence of these magnetic field errors .
this information will indicate the extent to which the edm error can be reduced by averaging over multiple runs . at a minimum
this sort of control experimentation will permit an objective determination of systematic error .
another concern is that a sufficiently large magnetic disturbance could destroy the phase lock .
pulses of this magnitude seem not to have been observed in recent polarization measurements with deuterons in the cosy ring at juelich@xcite@xcite .
if such a pulse is large enough for one or more of the phase - locked loops to lose lock the run would obviously have to be discontinued and discarded .
_ as well as spotting two typos in this section ( which have been fixed ) jrg pretz has questioned the treatment of thomas precession and pointed out that the averaging in eq .
( [ eq : ntlo.9p ] ) ( the main result in this section ) is incorrect .
more careful treatment has introduced a factor @xmath109 into eq .
( [ eq : ntlo.9p ] ) .
this factor is unimportant for proton edm measurement , but important for the electron .
the averaging error pointed out by pretz has not been corrected however .
it is left as an easy exercise to figure out why the averaging is incorrect ( for the radial magnetic field component at issue)and a hard exercise what to do about it .
the eventual success of the edm measurement will depend on arranging conditions to cancel this source of spurious precession .
_ _ in compensation for the reduction in precision acknowledged in the previous paragraph , i am pleased to report a significant improvement in suppression of radial magnetic field systmeatic error@xcite . by using the self - magnetometry feature of an octupole focusing electric storage ring `` bottle ''
, the average radial magnetic field error can be cancelled with accuracy @xmath110 t . _ for comparison purposes , we continue to develop formulas applicable to both frozen spin and rolling spin methods . for an apparatus intended to measure edm s
, one has to be prepared to suppress any mdm - induced precession that mimics edm - induced precession .
the leading source of systematic error is ( unintended ) rest frame radial magnetic field @xmath111 which , acting on the mdm , mimics the effect of radial electric field acting on the edm .
it is only the rest frame magnetic field @xmath62 that causes mdm - induced precession , but both laboratory frame components , @xmath112 and @xmath113 , contibute to @xmath62 .
these are also the lab frame field components capable of causing the closed orbit to deviate from the ideal , horizontal , design plane .
because the guide field is radial electric , but with possible error from not quite vertical electrodes , the dominant field error can be expected to be a vertical laboratory electric component @xmath113 . transformed to the electron rest frame , this contributes a radial rest frame magnetic component @xmath114 . to the extent the beam _ does _
move out of the horizontal plane , the cross product in the transformation from lab to rest frame can also produce a radial magnetic field .
this i neglect , relying on precise beam steering , and particle trap operation , and hoping that the tendency to average to zero will not be defeated by conspiring correlations . substituting into eqs .
( [ eq : lorentz.3b ] ) , the fields being retained are , for electrons @xmath115 where the error fields depend on @xmath89 which is the angular position around the ring . here
the upper of the @xmath116 and @xmath117 signs refer to a clockwise ( cw ) and the lower to a counter - clockwise ( ccw ) beam ( _ not _ roll ) direction .
in the rest frame , since the velocity vanishes , only eq . ( [ eq : ntlo.1 ] ) influences the vertical motion of the particle by the lorentz force , @xmath118 where the sign is appropriate for electrons , and @xmath113 and @xmath112 are unknown lab frame field errors .
also we are temporarily ignoring the fact that the @xmath112 term will have caused cw and ccw closed orbits to differ . knowing that the beam stays more or less centered vertically over long times , averaging this equation over @xmath89 yields the result @xmath119 in particular , if @xmath120 then @xmath121 .
( this analysis has neglected gravitational forces .
there is certainly a vertical laboratory gravitational force @xmath122 acting on each electron corresponding to its total energy @xmath123 .
neglecting the issue of difference between cw and ccw beams , one can define an `` equivalent '' laboratory magnetic field @xmath124 such that @xmath125 since this is far smaller than the smallest conceivable uncertainty in the true magnetic field error @xmath112 , i will continue to neglect gravity gravitational acceleration accounted for here , to be of concern such an effect must imply a general relativistic effect capable of mimicking an observable cp - violating effect .
it seems likely to me that any such mechanism could plausibly be responsible for the observed matter / anti - matter imbalance that has motivated the search for non - vanishing edm s in the first place . by this reasoning
, general relativistic considerations can not really alter the motivation for attempting to measure electric dipole moments .
in other words , _ if it looks like cp violation , it is cp - violation . _ ] . )
we will concentrate , for example , on a beam polarized in the @xmath5 plane , perpendicular to the local radial direction .
the rest frame polarization vector @xmath70 ( conventionally written without a prime in spite of relating to the rest frame ) is given by @xmath126 the error torque acting on the mdm is proportional to @xmath127 ; @xmath128 and the design torque acting on the edm is proportional to @xmath129 ; @xmath130 consider first the case of a truly frozen spin beam , meaning @xmath93 is constant .
the error torque and the design torque are always parallel or antiparallel .
the magnetic error causes a vertical kink in the orbit .
had this kink been caused by electric field it would have caused no spurious spin precession because the particle energy is magic .
for magnetic bending , the energy is _ not _ magic .
the resulting precession error is proportional to the magnetic field error @xmath112 , but the effect is more serious for the not - very - relativistic magic protons than for the fully - relativistic electrons . comparing eqs .
( [ eq : ntlo.6 ] ) and ( [ eq : ntlo.8 ] ) , the discrimination against spurious precession can be expressed by an averaged `` edm selectivity factor '' , @xmath131 the @xmath109 factor ( evaluated at the `` magic '' beam energy ) allows for the fact that it is the difference between magnetic and electric precession that needs to be accounted for .
it can be shown , at the magic momentum , that the difference between magnetic and electric spin tune precession rates ( per radian angular momentum deflection ) is given by @xmath132 .
though this factor is unimportant for the proton edm measurement , it is important for the electron measurement , for which @xmath133 .
because @xmath134 can have either sign , there will be significant reduction by averaging of the numerator factor . nevertheless , because the mdm is so huge compared to the edm , in this truly frozen case , this s.f .
will be large enough to require huge reduction of precession caused by @xmath134 field errors due to external sources by magnetic shielding .
consider next the case of rolling polarization beam , meaning @xmath135 varies sinusoidally .
in this case , replacing @xmath136 , the selectivity factor becomes @xmath137 because the roll frequency is uncorrelated with the rotation frequency , the average will be essentially zero over times of order the roll period , say 10ms , or longer .
what has previously considered to be the most serious source of systematic error in edm determination has been completely eliminated by the rolling polarization .
the same cancellation will occur for all field errors .
this `` miracle '' is only brought about by what is the true miracle , namely the phase - locked , rolling - polarization , trapped beam . _ as noted at the start of this section , this result is incorrect , but has not yet been repaired .
_ though the spurious spin precession due to @xmath96 has been eliminated , this magnetic field error causes clockwise and counterclockwise orbits to differ .
this magnetic field violates the time reversal symmetry of the ring and causes the cw and ccw beams to separate vertically .
this separation will be limited however by the opposite electric field components @xmath138 that the beams encounter because of their different orbits . transformed to the electron rest frame , these yield @xmath139 here @xmath113 and @xmath112 are the same lab frame field errors as before and @xmath140 are laboratory frame electric fields that develop ( differently cw and ccw ) to limit the orbit deviations caused by @xmath112 . averaged over longitudinal coordinate @xmath141 the fields @xmath140 will be opposite for cw and ccw beams but the equality is not guaranteed locally at every longitudinal position .
( this is because the field errors do not respect the lattice mirror symmetries .
as it happens , since the edm lattice will have only mild beta function dependence on @xmath141 , the symmetry will be only mildly broken , causing the fields @xmath140 to be more or less equal and opposite locally . )
the rest frame vertical electric fields are @xmath142 upon averaging these yield @xmath143 the rest frame magnetic fields are @xmath144 upon averaging and substituting from eqs .
( [ eq : cw - ccw.5 ] ) and ( [ eq : cw - ccw.6 ] ) these yield @xmath145 the fact that the signs are the same shows that the spin precession caused by @xmath146 field error is the same for cw and ccw beams .
the reason for alternating between clockwise and counterclockwise circulating beams is _ not _ , therefore , to cancel spurious @xmath146-induced precession .
it is to cancel other sources of systematic error .
if the radial magnetic field causes a net - upward force on an electron beam then its effect on a counter - circulating electron beam will be downward .
one can therefore measure @xmath146 by measuring the vertical separation between counter - circulating beams .
the vertical deflection at bpm @xmath147 caused by @xmath148 vertical deflections @xmath149 at lattice positions @xmath150 is given by@xcite @xmath151 where @xmath152 is the lattice tune , @xmath153 is the vertical betatron phase advance from @xmath150 to @xmath147 , and @xmath154 and @xmath155 are the corresponding vertical twiss function values . for the edm ring the @xmath156-functions will not depend strongly on position .
this , simplifies eq .
( [ eq : bpm.1 ] ) to @xmath157 for _ true frozen spin operation _ , to increase rejection of radial magnetic field error @xmath112 , the vertical tune @xmath158 would be adjusted intentionally so that @xmath159 .
making the vertical focusing weak amplifies the vertical separation caused by a given radial magnetic field@xcite .
this improves the sensitivity with which the field error can be compensated away by eliminating the beam separation .
the rolling polarization modification has made this unnecessary . by allowing vertical tune @xmath158 to be comparable to horizontal tune
@xmath160 greatly simplifies the storage ring design .
this will inevitably lead to improved dynamic aperture , and better suppression of emittance dilution due to intrabeam scattering@xcite .
but the rolling polarization brings with it inevitable deflection errors caused by wien filter imperfection .
this makes it important to know the orbit shift caused by local @xmath112 magnetic field errors .
one of the important strategies for reducing the edm systematic error associated with roll - reversal , is to have the roll caused by the reversal of a _ single isolated source_. nevertheless , for generality , we continue to treat multiple @xmath112 error sources , since precision orbit centering will also be important .
the ultimate edm accuracy limit will probably concern the accuracy with which the polarization can be reversed , which will depend on closed orbit precision .
we have been pretending the ring has no field errors , which will obviously not be the case .
so precise orbit smoothing will be important .
the more accurately the beam can be held on the design orbit , the more accurate the wien reversal will be ( because deviant closed orbit reflects deviant wien deflection ) . assuming @xmath148 deflection errors are distributed more or less uniformly around the ring , an average value for the cosine factor is @xmath161 .
the vertical deflection error @xmath149 caused by magnetic field error @xmath162 in an element of length @xmath163 is @xmath164 and the summation gives approximately @xmath165 also @xmath166 . with these approximations ,
( [ eq : bpm.2 ] ) becomes @xmath167 the momentum factor can be expressed in terms of @xmath168 as @xmath169 , yielding @xmath170 the main purpose for the @xmath171 vertical bpm s is to accurately measure the average deflection @xmath172 ( or rather , the difference in vertical deflections of counter - circulating beams ) in order to infer @xmath146 . defining the precision of each such measurement by @xmath46 , the r.m.s .
field error is given by @xmath173 substituting from eq .
( [ eq : bpm.5 ] ) into eq .
( [ eq : ntlo.9 ] ) and using parameter values to be spelled out shortly , @xmath174 for truly frozen spin operation this is the factor to be contrasted with the disadvantage factor @xmath175 the mdm has over an edm value of @xmath53e - cm .
a priori one does not know @xmath46 , but @xmath176 nm bpm resolution has been achieved , for example in connection with the interanational linear collider@xcite@xcite@xcite@xcite@xcite .
this measure suggests the smallest upper bound that can be set on the electron edm will with truly frozen spin operation will be about @xmath177e - cm .
( this is why simultaneous countercirculating beams were considered to be necessary in the reference@xcite . )
the same formulas can be used for estimating the precision with which closed orbit deviations from the ideal orbit can be minimized for our actual rolling polarization case .
it has been emphasized that the combination of rolling polarization and phase - locked trap have greatly reduced the sensitivity of the edm measurement to electric and magnetic field uncertainties .
these claims are , of course , only valid if successful phase - locked trap operation is actually achieved .
magnetic field errors can certainly prevent phase - locking of the spin wheel and an occasional magnetic field transient can break the lock , forcing the run in progress to be aborted .
this would be true even assuming perfect polarimetry .
electromagnetic field transients can also cause runs to be aborted due to beam loss or emittance dilution , but much ordinary storage ring experience exists , making further discussion of particle loss unnecessary . with electric fields of several million volts per meter
just about everywhere along the beam line , the most likely electric transient is a spark .
since this could be catastrophic for sensitive electronics nearby , this has to be made essentially impossible by appropriate engineering design .
so nothing more will be said about this possibility either .
steady , ultralow frequency field errors do not constitute a problem for phase - locked operation , since the compensation elements can comfortably cancel their effects .
this would include the electric and magnetic fields due to miniscule dc leakage currents .
it would also include tiny patch effect magnetic fields associated with microscopic mosaic spreads in conductor surface structure .
the only serious problem to be faced is transient external magnetic fields , which are notoriously difficult to shield against .
the polarization roll frequency is autonomously ( i.e. self - generated ) , not externally , imposed . with no external source setting the roll frequency , no error can result from the correlation of an external field with the polarization vector
. there could be accidental synchonism .
for example the roll frequency could be 60hz .
but this possibility can be recognized and avoided .
the dominant source of error for all previous edm experiments has come from unknown spurious spin precession caused by wandering dc ( sub - mhz ) and ( less important ) low frequency ( sub - khz ) magnetic fields remaining in spite of all attempts to shield them .
the differences among competing experimental methods largely come down to differences in approach to minimization of the errors caused by such unknown magnetic field variation .
other edm error sources can only become significant once this source has been reduced by several orders of magnitude .
one significant source of man - made noise in this frequency range comes from 50 or 60hz hum and its harmonics .
this noise is sufficiently narrow - banded to be avoidable , but only if frequency selectivity is available , by avoiding these frequencies .
more significant are transients due to starting and stopping of electrical machinery , crane and elevator movements , and vehicles passing nearby .
these noise sources are discussed , for example , in reference @xcite , and in references given in that paper .
numbers to be used below concerning residual magnetic fields are obtained from this reference .
the dc earth magnetic field of about @xmath178mt is not as important as its r.m.s .
variation over typical edm run durations .
after serious four layer passive magnetic shielding , the rms error will be about @xmath179nt .
this five orders of magnitude reduction is a somewhat ambiguous combination of ac / dc fraction in the range from 0.01 to 0.1 , and a passive permalloy shielding reduction ratio in the range from 1000 to 10,000 .
active shielding reduces the noise level to about @xmath180nt .
this further reduction of 10 due to active field compensation is accurately measureable in any fixed geometry , but depends strongly on the particular geometry of coils and experimental apparatus . the effectiveness of active shielding decreases rapidly with increasing frequency . with resonant polarimetry
we can concentrate on higher , but still very low frequency , non - man - made , terrestrial , magnetic noise . according to bianchi and meloni@xcite , natural terrestrial noise in the frequency range from 1 to 100
hz is dominated by radiation produced in lightning strikes , several million per day world - wide and trapped between earth and ionosphere , especially at schumann resonant frequencies , 7.8hz and its lowest few harmonics .
see figure [ fig : terrestrialb - noise ] which shows noise levels of about @xmath181 .
assuming permalloy shielding ( which continues to work well in this frequency range ) an estimated residual magnetic field noise in this frequency range , is @xmath182 \approx \pm1\,{\rm ft/\sqrt{hz}}. \label{eq : magnoise.2}\ ] ] active magnetic shielding would be of little value over large volumes in this frequency range . natural terrestrial noise in the frequency range important for rolling polarization edm measurement .
vastly greater noise at ultralow ( mhz ) frequencies is not shown . ]
the experience with magnetic transients most easily applicable to our storage ring measurement of charged particle edm s comes from the numerous experiments that have been performed to measure the neutron edm . a thorough review of all such experiments to find the most sophisicated , lowest possible residual field procedures , is unnecessary .
it is more important , at this early stage of development of a novel method , to acquire prejudice as to the seriousness of the issue , and to estimate the performance to be achieved at reasonable cost . in this spirit
i review a single recent neutron edm experimental publication , chosen , more or less , at random@xcite , which describes experience at the paul scherrer institute ( psi ) .
their shielded volume is roughly a 2.5 m cube .
the shielded volume required for our electron edm experiment would be a square room , perhaps 7 m on a side , 3 m or 4 m high .
a proton edm experiment would perhaps require a shielded 2.5 m diameter toroidal volume of circumference 300 m or greater .
these dimensions seem close enough for neutron edm experience to be applied to the edm experiments under discussion . to obtain neutron edm measurements to present day accuracy of roughly @xmath183e - cm accuracy ,
requires not just the passive magnetic shielding described previously , but also the further factor of 10 by active compensation , using compensation coils located outside the magnetic shielding , controlled by precision magnetometer measurements inside the shield .
the psi investigation found , with this _ active _ compensation , that `` disturbances of the magnetic field are attenuated by factors of 5 to 50 at a bandwidth from 1mhz up to 500mhz , which corresponds to integration times longer than several hundreds of seconds and represent the important timescale for neutron edm measurement . ''
proton edm run durations will be about the same , or somewhat longer . with passive magnetic shielding at psi ( an accelerator environment ) over times of a few hours ,
they find nighttime dc magnetic @xmath32 field levels changing over a @xmath184nt range and increasing to @xmath185nt in the daytime . with active shielding this is reduced by more than a factor of 10 . for round numbers
let us say the residual dc noise , after maximal shielding is at the @xmath178nt level ; @xmath186 \approx
\pm 1\,{\rm nt}. \label{eq : magnoise.1}\ ] ] for the electron and proton edm measurements proposed here i assume similarly high quality passive permalloy magnetic shielding , but not necessarily active shielding . for the electron edm measurement this would probably be accomplished by putting the whole experiment in a magnetically shielded room .
this would be impractical for the proton edm measurement .
rather the ring lattice would be enclosed in a shielded toroidal tube . as regards active magnetic compensation , based on psi experience ,
i assume it would be ineffective and will not be used . to simplify the continuing discussion
i will abbreviate the `` ramsey neutron edm '' method as `` nedm '' , the previously proposed storage ring proton edm experiments as `` pedm - frozen '' , and my proposed new proton method as `` pedm - resonant '' . as mentioned earlier , the nedm , ramsey separated rf oscillator method , can be categorized as a frequency domain measurement . in a constant magnetic field
b , the neutron polarization precesses with larmor frequency proportional to b. for runs of several hundred second duration one measures an edm - induced phase shift between rf induced polarization induced at the run start , from that observed at the end .
( the famous ramsey before and after rf pulses improve the precision by converting the center of the fringe pattern into a zero crossing .
) a spectral pattern is obtained from a large number , such as 100 , of sequential runs with rf frequency progressively advanced by a tiny amount each run .
( in this paper i am referring to patterns like this as fringe patterns . )
repeated with electric field off , any non zero edm would cause a shifted fringe pattern or , in more literal terms , a frequency shift . in the pedm - resonant method
being proposed , the resonant polarimeter monitors the precession phase continuously .
( this reduces the number of runs from 100 to 1 , but this is not the point here . )
the accumulating phase is the sum of the edm phase shift and a ( vastly greater ) artificially - imposed roll phase .
the edm contribution can only be extracted by subtracting the result of a subsequent run taken with exactly equal , but opposite - sign , artificial roll . in the nedm method ,
as already described , the edm measurement comes from subtracting data taken in sequential runs having different edm precession conditions ( such as reversed electric field ) .
otherwise , the nedm method and the pedm - resonant method are quite similar . but both are quite different from the previously proposed storage ring pedm - frozen methods . as estimated above in eq .
( [ eq : magnoise.2 ] ) , the r.m.s . magnetic field uncertainty for the nedm method , with both active and passive magnetic shielding ,
is @xmath180nt .
for the pedm - resonant method the r.m.s .
magnetic field uncertainty for the pedm - resonant method depends on the resonator bandwidth , which we take , conservatively , to be 100hz . using eq .
( [ eq : magnoise.1 ] ) which assumes passive , but not active shielding , we get the the r.m.s .
magnetic field uncertainty for the pedm - resonant method to be @xmath187 ft . the nedm method and the pedm - resonant methods are sufficiently similar , as regards sensitivey to unknown magnetic field for the previous two estimates to be directly comparable . by this comparison
the pedm - resonant can be expected to be five orders of magnitude more accurate than the nedm experiment . since the quoted nedm accuracy is @xmath188e - cm , the pedm - resonant accuracy ( coming from this source of error )
is estimated to be @xmath189e - cm .
the fact that this error is so small almost guarantees that the external magnetic field error source will not , in fact , be the dominant error source in the pedm - resonant method .
( it can be noted , in passing , that this procedure for estimating the accuracy , when applied to the pedm - frozen method , seems superficially to imply an accuracy limit in the range from @xmath190 to @xmath188e - cm , depending on assumption about the effectiveness of active magnetic shielding .
this is because , the random walk of the frozen spins protons is just like the random walk of polarized free neutrons ( except for their minor mdm difference ) and is dominated by the wandering dc magnetic field error .
what makes this inference incorrect is that a different method of suppressing magnetic field error is used in reference@xcite .
( as already explained ) measuring and then eliminating to exquisite accuracy , the relative vertical displacement of simultaneously counter - circulating beams , has the effect of suppressing the wandering dc error in the pedm - frozen method .
this is what permitted the pedm - frozen , @xmath1e - cm edm upper limit error estimate . )
as expected , the rolling polarization with resonant polarimetry almost completely suppresses all spurious polarization precession caused by magnetic noise .
magnetic fields at frequency lower than @xmath3 average to zero after multiple rolls .
the effects of magnetic fields at frequency greater than @xmath3 average to essentially zero even during a single polarization roll
. only magnetic fields with frequencies close to @xmath3 can produce noticeable spurious precession that would emulate an edm effect .
there is the possibility of an occasional loss of phase lock due to a short magnetic pulse .
but the precession impulse angle needed to lose phase lock is huge compared to the value that would give a seriously wrong edm value . during a run a magnetic impulse at this level
would necessarily force the run to be aborted immediately .
because the anomalous precession rates and magic velocities of proton and electron are different , their wien filter designs are not the same .
but , since the essential issues are much the same , only the electron case needs to be discussed here .
the critical wien filter , labeled @xmath191 in figure [ fig : singlering ] , causes the electron beam polarization to roll with frequency @xmath3 , a frequency tentatively in the range from 10 to 100hz .
a schematic wien filter design is shown in figure [ fig : striplinewienfilter ] . powered from upstream , the poynting vector @xmath192 points in the beam direction , and the electric and magnetic forces tend to cancel
. with @xmath193 horizontal and @xmath194 vertical , using relations ( [ eq : lorentz.4 ] ) , the electron rest frame fields , expressed in terms of laboratory frame quantities , are @xmath195 with wien filter tuned for zero transverse acceleration in the laboratory there is also zero transverse force in the rest frame , so @xmath196 , from which @xmath197 . the laboratory frame time duration within a wien filter of length @xmath198
is @xmath199 and the rest frame duration is @xmath200 . while within the wien filter in the rest frame , using eq .
( [ eq : newton.1 ] ) , the magnetic moment precession angle laboratory advance @xmath93 , which is the same as in the rest frame , is given by @xmath201 one of the denominator @xmath31 factors has come from the second of eqs .
( [ eq : wien.1 ] ) , and the other from the time dilation factor . to produce the desired roll frequency @xmath3 in a ring with circulation frequency @xmath18
, @xmath93 needs to be given by @xmath202 in the wien filter , assuming electrode width @xmath203 is much greater than gap @xmath204 , the electric field @xmath168 , the current @xmath205 and the magnetic field @xmath206 are given , in terms of the voltage @xmath207 , and wien filter parameters by @xmath208 then the net transverse force on an electron is @xmath209 for exact cancellation @xmath210 ( as a check , setting @xmath211 , one obtains @xmath212 which is the well known characteristic impedance of the stripline as a transmission line . even under dc conditions
one can describe the setup as a wave of infinite wavelength propagating along the line at the speed of light and being perfectly absorbed in the terminating resistor .
this is especially apt for the electron case , since the electron magic velocity is close to @xmath102 . )
combining formulas , the required length times magnetic - field product and current times length product , for electrons , are given by @xmath213 for @xmath214mhz , @xmath106hz , @xmath215 m , and @xmath216 m , the termination resistor is @xmath217 , and the current times length product is @xmath218a - m .
five amperes is a convenient current for precise control and reversal .
a 5a reference current source from the lhc accerator control system is dc amplified to produce a @xmath219 ring magnet supply current@xcite .
the current calibrator is described by fernquist , halvarsson , and pett@xcite .
this is , in turn , calibrated from a precise cern 10ma standard current , whose specification is year - to - year stability of 1 part in @xmath220 and short term reproducibility of 2 parts in @xmath221@xcite .
typical variation over 1000 seconds is about 5 parts in @xmath222 ; its temperature coefficient is @xmath223k .
meeting these specifications was made more difficult by the requirement for the unit to be mobile . in a single shielded location
higher precision will be practical .
furthermore , it will be possible to record every run current to higher precision by comparison with a standard current , for example using a current bridge circuit like that copied from reference@xcite and shown in figure [ fig : currentbridge ] .
run - by - run current reversal accuracy to several parts in @xmath34 should be achievable in this way .
of course the rest of the ring has to be shielded from the magnetic field caused by these 5a currents .
note though , that being a dc magnetic field , it causes no polarimeter response frequency shift .
stripline wien filter dimensions . with electromagnetic power and beam
traveling in the same direction , the electric and magnetic forces tend to cancel .
termination resistance @xmath224 is adjusted for exact cancelation . ]
current bridge used for high precision current monitoring . copied from cern pbc reference@xcite .
one current can be the active control current , the other a highly stable reference current .
even hand - held , 1 part in @xmath222 precision is obtained . ] establishing phase - locked , rolling - polarization electron trap operation will be a tour de force of accelerator technology .
but the task of setting @xmath191 to produce the desired roll frequency @xmath3 to adequately high accuracy is not very challenging , for example because the electron mdm is known to such high accuracy .
the significant challenge in measuring the electron edm to high accuracy will be in guaranteeing that forward and backward roll frequencies ( not counting the small edm effect ) are identical to exquisitely high accuracy .
the basic edm measurement comes from measuring , say , the forward roll frequency ( as a sideband deviation from the revolution frequency ) .
if the mdm - induced roll frequency were perfectly calculable ( which is far from the case ) it could be subtracted off . for @xmath225hz , and nominal edm value of @xmath1e - cm ,
the ratio of imposed roll to edm roll is about 10 orders of magnitude , making an absolute subtraction unthinkable .
on the other hand , both the electron @xmath226-factor and the electron mdm @xmath227 are known to parts in @xmath228 .
the sideband deviation of roll frequency minus revolution frequency can also be measured to high accuracy .
so , from eq .
( [ eq : wien.6 ] ) , with all factors accurately known , the @xmath229 product is well known , at least to the extent that the electron edm contribution can be neglected or , more usefully , when the electron edm effect has been arranged to cancel out , irrespective of its magnitude .
the fundamental measurement comes , therefore , from the difference between forward and backward roll frequencies .
this transfers the accuracy requirement to a reversal symmetry requirement .
the reversal symmetry has to be better than 1 part in @xmath8 .
( optimistically , for the proton edm measurement , this may be reduceable to 1 part in @xmath222 . )
the leading task is therefore to minimize the reversal inaccuracy when the @xmath191 wien filter is reversed . by temperature control
the wien dimensions can be controlled to good accuracy . by interferometry
the wien gap width @xmath204 can also be kept quite stable .
the magnetic field nearly on axis can also be measured and held constant to high accuracy . to be discussed shortly ,
the best measure of the accuracy with which reversal is being done is to readjust @xmath191 and @xmath230 to make the polarization roll in the horizontal , rather than vertical , plane .
this completely suppresses the edm effect . for best accuracy over shortest time interval
, this suggest that sets of 4 runs , forward and backward roll , edm effect present and suppressed , should form the basic data collection block .
interleaving these four configurations repeatedly and in varying order for multiple cycles may produce smallest systematic error .
cw / ccw reversal of beam direction would then be relatively infrequent , perhaps once per day .
other reversals are possible too complicated for thorough analysis here .
beam rotation can be cw or ccw . since this requires reversing powered end and termination end , it is not as clean as one would wish .
the connections can be switched or , with the apparatus mounted on a turntable , as in neutron edm practice , the wien filter can be switched end - for - end .
rolling the wien filters around their axes represents another way to cancel reversal bias . in combination
these reversals can reduce the reversal error . reversing the electric field ( as is done in neutron edm measurement )
is clearly impossible .
but , in the electron case , there is the possibility of measuring the positron edm .
the electron - positron edm difference is probably the most accurate lepton measurement possible
. it will be valuable to be able to measure the accuracy with which the polarization roll reversal can be accomplished , irrespective of how good that accuracy may be .
one can not test the reversal accuracy using an unpolarized beam to suppress the edm contribution there would be zero resonator response .
one can , however , arrange for the polarization to roll in a horizontal plane .
this would force the edm precession to average to zero , while leaving the resonator unchanged .
any residual asymmetry could only be ascribed to being instrumental . to achieve stable horizontal polarization roll
will require biasing the @xmath230 wien filter while leaving @xmath191 set for truly frozen spin .
( this will also entail replacing @xmath230 by @xmath191 for wheel stabilization . ) in this case the reversal symmetry can be checked with full digital precision .
this may turn out to be the best way to quantify the reversal symmetry .
much has been made of the important of the isolation of the @xmath191 wien filter .
but , for the reason just given , it will be just as important for @xmath230 to be cleanly isolated from @xmath191 and from the rest of the ring . irrespective of the absolute accuracy of each single roll reversal
, the overall edm accuracy can be improved by averaging over large numbers of runs with conditions reversed in ways that should cause asymmetries to vanish .
in the proposed ring there will be no intentional longitudinal field components @xmath232 .
but there will be bend field errors @xmath231
. there will therefore have to be at least one longitudinal trim field @xmath233 . the rest frame longitudinal fields to be discussed are then @xmath234 where @xmath231 represents an _ unknown _ longitudinal field component , and @xmath233 is a _ known _ longitudinal trim field .
it is unecessary to introduce a time - dependent magnetic rf term , because the rf electric field does not interact ( on - axis ) with the mdm .
( off - axis there is a time - dependent magnetic field that has be discussed separately , along with other issues such as rf misalignment . )
various non - zero contributions to @xmath235 have also been suppressed , since they give only edm - induced precession small compared to the dominant edm effect .
consider a more or less vertically polarized beam , with rest frame polarization vector @xmath70 given by @xmath236 mdm and edm torques are proportional , respectively , to the cross products @xmath237 the only important torque here is magnetic since the electric rf torque is neglible compared to the bend field torque given in eq .
( [ eq : ntlo.8 ] ) .
the magnetic torque can , potentially , be large but the leading vertically - directed precession @xmath238 will be cancelled by the trim solenoid .
@xmath239 since both @xmath231 and @xmath112 contribute to unbalancing the equilibrium , though at right angles , they both have to be adjusted empirically .
if the @xmath232 trimming altered @xmath129 it would give an unknown edm effect .
but there is no such coupling .
spurious precession can result from the failure of commutation of successive rotations , proportional to the product of the two precession angles .
the mdm precession angles are themselves so small their squares are negligible .
but the artificially imposed rotation angles are vastly greater .
should a nonzero edm value actually be detected , it might need to be corrected for failure of commutation errors .
but this source of error could not , by itself , produce a statistically significant non - zero edm value of either sign .
any spurious systematic phase shift from this source would include the sign of the artificial roll and would cancel in the subtraction .
geometric phase errors are a significant issue for measuring the deuteron edm .
an ideal frozen spin deuteron ring would have electric and magnetic fields homogeneously superimposed . but segregating the electric and magnet fields into disjoint sectors , as suggested by senichev@xcite , would make the ring construction vastly easier to construct and commission .
it has usually been assumed that segregation into electric and magnetic sectors in this way would be made unacceptable for edm measurement by geometric phase shifts caused by the commutation failure of the quite large mdm - induced precessions in successive sectors .
cancelation on the average of the spurious precessions may allow the method proposed here to be immune from this systematic error .
this section assumes the polarization phase locking described previously has been successful . with the beam polarization locked one can take advantage of the precisely - known electron magnetic moment @xmath227 and anomalous moment @xmath240 as given in table [ tbl : magneticparams ] . the spin tune
@xmath10 relates to precession around the vertical axis . in an all - electric ring @xmath241
is given in terms of relativistic factor @xmath31 by @xmath242 at the `` magic '' value , @xmath243 , @xmath244 , where @xmath245 solving eq .
( [ eq : emdm.2 ] ) for @xmath31 , ( and requiring @xmath246 ) , @xmath247 we then obtain @xmath248 though this formula is exact , and quite simple , it can usefully be expanded ( to unnecessarily high precision ) to @xmath249 here @xmath10 has been re - expressed in terms of the frequency deviation from magic , @xmath250 , of the polarization around a vertical axis .
this formula is intended for use only near @xmath251 , with the ratio @xmath252 being a tiny number , less than @xmath253 for example .
further terms are easily obtained , but for values of @xmath254 large enough to require them one can simply use eq .
( [ eq : emdm.6 ] ) . with the rf frequency
known to arbitrarily high accuracy and the beam polarization locked to the revolution frequency , measuring the polarimeter response frequency establishes @xmath31 to correspondingly high accuracy .
one is accustomed to expecting a spread of spin tunes due to finite betatron and synchrotron amplitudes and energy deviation of each individual particle from the central beam @xmath31 value . in a certain sense
the beam conditions are better defined than this .
as long as decoherence can be neglected ( i.e. for times short compared to sct ) the net angular precession of each spin vector can not exceed @xmath255 . during this time @xmath244 for every particle on the average .
one sees that phase locking of the polarization establishes the mean beam parameters to exquisitely high accuracy .
of course this remarkable behavior begins to break down for times approaching sct . in effect ,
loss of phase lock and decoherence of the beam polarization are two manifestations of the same spin evolution .
achieving large sct and maintaining the possibility of phase locking the polarization are equivalent tasks .
this analysis has assumed a purely electric lattice . in practice
there will be some average vertical magnetic field component @xmath256 that will cause the @xmath31 value determined by eq .
( [ eq : emdm.7 ] ) to be not quite correct .
but it has also been anticipated that keeping the polarimeter loop locked will require a wien filter which compensates for the ( unknown ) total angular deflection @xmath257 by magnetic fields .
the spin tune ascribable to this lattice modification is @xmath258 which compensates for fraction @xmath259 of each turn being caused by magnetic rather than electric field .
the ( well , but not perfectly , known ) strength @xmath24 of the wien filter excitation can be interpreted as a measurement of the angular bend caused by unknown magnetic field component @xmath256 averaged over the ring .
the ultimate claimed @xmath31 precision depends on this correction . in the actual experiment ,
since the roll is around the radial axis , the spin tune will necessarily be exactly zero , meaning the phase locking requires the wien filter to exactly cancel @xmath256 .
one problematical phase - locked loop detail concerns the box labelled `` polarity reversal '' in figure [ fig : singlering ]
. it would be splendid if the roll reversal could be completed without losing phase lock .
but reversing the roll introduces a discontinuous phase shift in the polarization signal inputs to the resonant polarimeters which will force the circuits to ring down and then ring up again , unlocking the closed loop synchronization .
perhaps the roll drive phase could be judiciously reversed at just the right time to avoid this effect and preserve all phase locking ?
if this could be done then the roll orientation could be reversed at a quite high rate within each fill cycle , greatly reducing systematic frequency shift errors .
if not , then roll reversal may have to be limited to once per fill .
however there is an excellent reason for _ not _ attempting to synchronize any control function with the roll frequency it would contradict the claim made earlier in the paper that no control sources are synchronous with the roll frequency .
violating this seems sure to cause trouble .
this seems to guarantee loss of phase - lock during roll reversal .
the purpose of the paper has been to describe a method for measuring edm s of electrons and protons with far higher accuracy than they are presently known .
the essential change from earlier proposals comes about from the introduction of resonant polarimetry , which permits phase - locked loop , rolling - polarization trap operation , and shifts the edm signal into the frequency domain .
important contentions , conclusions and conjectures are included in the following long list : * successful application of the method depends on two not yet established experimental methods : resonant polarimetry ( promising theoretically ) and `` rolling polarization trap '' operation meaning stable , phase - locked , rolling polarization operation(promising experimentally ) .
* it is thermal noise in the resonant polarimeter that limits edm measurement _ precision_. * a successful single beam fill will include at least one forward / backward reversal of the roll ( not beam ) direction , with roll frequency precisely measured both before and after .
this will produce one _ precise _ edm measurement .
* it is frequency domain , digital , frequency scaling , that enables at least one reversal per fill , giving one precise edm measurement for every roll reversal .
* a successful single beam run will consist of four data sets , edm effect on and off , roll forward and backward .
* cw / ccw beam direction reversal will reduce systematic error . *
expressed as edm upper limit , measurement _ precision _ of @xmath0e - cm after year - long running , for either electrons and protons , can be expected . *
_ accuracy _ at the same level as the _ precision _ will require average sign reversal accuracy of wien filter length / strength product at the level of one part in @xmath260 , also averaged over one year .
lower reversal nulling accuracy will give proportionally lower edm measurement accuracy .
* apparatus constituting a single wien filter will all be contained in a single , temperature regulated , limited vibration , magnetically shielded , highly isolated , etc . box . *
`` rolling polarization trap '' operation greatly improves edm selectivity ( essentially eliminating spurious external field induced precession ) and permits the rest of the ring to have relaxed requirements . *
identical beam distribution requirement for oppositely - directed beams are similarly relaxed by the requirement of precise edm measurement for every fill .
* the greatly reduced sensitivity to unknown radial magnetic field eliminates the need for counter - circulating beams and ultra - low vertical tune @xmath158 .
this , in turn , allows relatively strong alternating gradient ring design , providing larger aperture and higher stored charge .
* emittance growth due to intrabeam scattering ( ibs ) has been seen as a serious impediment to edm measurement . with ultra - low @xmath158
not required , running `` below transition '' , which tends to suppress beam growth due to ibs , will be possible . *
electron cooling has proved ( at juelich ) to be highly effective in increasing sct .
though long sct would improve edm accuracy , electron cooling has previously been rejected out of hand , because of its uncontrolled electric and magnetic field errors it would introduce .
rolling polarization may remove this impediment to the use of electron cooling .
* spin coherence time sct is greatly increased by mbius ring operation .
recapitulating the recapitulation , all of these wonderful edm measurement improvements have been brought about by the phase - locked loop , storage ring polarized - beam trap , made possible by resonant polarimeter .
these advances have not yet been achieved in practice .
the most immediate task is to successfully build and test resonant polarimetry .
i have profited especially from conversations with bill morse , yuri orlov , frank rathmann , yannis semertzidis , and john talman , as well as with jim alexander , mei bai , ivan bazarov , mike blaskowitz , peter cameron , bruce dunham , ralf gebel , tom kinoshita , andreas lehrach , alfredo luccio , nikolay malitsky , bob meller , maxim perelstein , michael peskin , thomas roser , dave rubin , anders ryd , yuri senichev , valeri shemelin , eric smith , ed stephenson , and hans stroeher . and others .
s. afach , _ dynamic stabilization of the magnetic field surrounding the neutron electric dipole moment spectrometer at the paul scherrer institute , _ arxiv.org>physics>arxiv:1408.6752[physics-ins-det ] , 2014 | precise measurement of the electric dipole moments ( edm ) of fundamental charged particles would provide a significant probe of physics beyond the standard model .
any measurably large edm would imply violation of both time reversal and parity conservation , with implications for the matter / anti - matter imbalance of the universe , not currently understood within the standard model .
a frequency domain ( i.e. difference of frequencies ) method is proposed for measuring the edm of electrons or protons or , with modifications , deuterons . anticipated precision ( i.e. reproducibility )
is @xmath0e - cm for the proton edm , with comparable accuracy ( i.e. including systematic error ) .
this would be almost six orders of magnitude smaller than the present upper limit , and will provide a stringent test of the standard model .
resonant polarimetry , made practical by the large polarized beam charge , is the key ( most novel , least proven ) element of the method .
along with the phase - locked , rolling polarization `` koop spin wheel '' , resonant polarimetry measures beam polarization as amplitude rather than as intensity .
this permits all significant observables to be directly measureable as coherent frequencies .
the same apparatus can be employed to measure magnetic dipole moment ( mdm ) values with high accuracy . but this capability is more usefully exploited to determine ring parameters with otherwise unachievable accuracy , using mdm values that are already known to high precision . also novel , though less essential , is the mbius storage ring lattice modification , which greatly increases the spin coherence time ( sct ) , with correspondingly improved accuracy .
important sources of edm error , statistical or systematic , are considered , along with measures to be taken for improved accuracy .
their effects can be expressed as the edm upper limits they imply .
the polarization roll , at 100hz for example , and adjustable by a constant control current , causes spurious torques due to field errors to average to zero to high accuracy .
since these torques have been considered to be the dominant source of systematic error in truly frozen spin operation , this is a major improvement resulting from the rolling polarization .
important sources of systematic errors remain , the main one being due to wien filter reversal uncertainty .
both electron and proton spins can be `` frozen '' in all - electric storage rings , and their edm precisions should be comparable .
freezing the deuteron spin requires a superimposed electric and magnetic guide field ; otherwise the rolling spin method and precision should be similar .
but the deuteron option is not discussed in this paper . |
clusters are important materials both from the point of view of basic research and from an application point of view .
their , partly controllable , unique physical and chemical properties can be related to the large surface - to - volume ratio or to finite - size or quantum - confinement effects .
these properties make them interesting for use in , e.g. , nanoelectronics and in catalysis .
the central issue is to understand and predict how the properties of interest depend on the size of the cluster .
however , addressing this question is complicated by a number of serious problems .
first , clusters contain typically from some 10s to several 100s of atoms , so that they are neither small , well - characterized molecules , nor macroscopic , approximately infinite crystals .
moreover , in experimental studies , one often considers an ensemble of more or less mono - disperse clusters whose size distribution is largely unknown .
second , experimental and theoretical studies focus often on different systems : in experiment the clusters may be dispersed in some solvent and may possess surfactants and only sometimes are produced in the gas phase , whereas isolated clusters in this gas phase almost exclusively are the subject of theoretical studies .
third , in theoretical works the fairly large size of the individual clusters makes it necessary to invoke one or more approximations : either highly accurate methods on selected clusters with pre - chosen structures are applied , or less accurate , parameterized methods are used on more different clusters , but also then the structures are often chosen ` reasonably ' . in this work we shall present results of our theoretical studies on ni@xmath0 clusters with @xmath1 up to 150 .
our approach is based on the approximate embedded - atom method ( eam ) in calculating the total energy of a given structure and we shall show that this method provides a good compromise between accuracy and computational speed .
moreover , we have used our own _
aufbau / abbau _ method in determining the structures of the lowest total energies .
ni clusters provide one of the most studies sets of clusters and , therefore , is an excellent system for exploring new theoretical or experimental approaches .
this is our main reason for focusing on that system .
moreover , through our unbiased structure - optimization approach we are able to identify the structures of more low - energy structures for each size , and , simultaneously , obtain structure and total energy for all values of @xmath1 up to @xmath6 , whereby size - specific properties can be extracted .
once the structure of a given cluster has been chosen , the variational principle allows for a systematic improvement of the calculated electronic properties . a similar approach does not exist for the structural properties where one ,
instead , has to use more or less biased methods for determining the structure of the lowest total energy . here
, three issues make such theoretical studies very demanding : first , the computational efforts for a single structure scales with @xmath1 to some power from 2 and upwards ; second , the number of structural parameters that have to be determined scale linearly with @xmath1 ; and , third , the number of meta - stable structures scales faster than any power of @xmath1.@xcite therefore , any theoretical study employs one or more approximation that may or may not be crucial . accordingly , parameter - free calculations on ni@xmath0 clusters@xcite have been applied mainly for smaller clusters as well as on single , high - symmetric , larger ones .
alternatively , unbiased structure optimizations for larger clusters are all based on more approximate total - energy methods like the eam or related methods,@xcite whereby only for @xmath7 the structure has been optimized completely unbiased , or the very simple @xmath8-body gupta,@xcite sutton - chen,@xcite finnis - sinclair,@xcite murrell - mottram,@xcite or morse potential@xcite has been used , whereby unbiased structure optimizations up to @xmath9 have been carried through.@xcite only for very simple and material - unspecific potentials , like the lennard - jones potential , largely unbiased structure optimizations up to @xmath10 have been carried through.@xcite finally , ni@xmath0 clusters have also been studied with a tight - binding method,@xcite whereby the structure only for clusters with @xmath11 was optimized unbiased . of the experimental studies on ni@xmath0 clusters,@xcite the chemical - probe experiments@xcite have given very valuable information on the structure of the smaller clusters .
on the other hand , results for medium and large ni clusters,@xcite obtained by performing near - threshold photoionization and time - of - flight mass spectroscopy can be used in identifying particularly stable clusters and , subsequently , in providing information on growth modes .
> from the earlier studies on nickel clusters it has been found that highly stable clusters occur for @xmath12 and @xmath13 for which the structures are multilayer icosahedra , i.e. , the first and second mackay icosahedra . moreover , in several theoretical studies@xcite a _ fcc _ truncated octahedron was found for ni@xmath14 which is in accord with experimental results .
thus , it can be assumed that for these three cluster sizes a ` structural consensus ' has been achieved , which in turn provides a useful check for any further study . on the other hand , for @xmath15
some theoretical studies@xcite have predicted marks decahedron to be the most stable structure , in disagreement with experiment .
moreover , the cluster growth pattern for intermediate sizes , i.e. , for @xmath16 and for @xmath17 , remains unclear and the obtained results seem to depend strongly on the used approach . in the present study the structure and the energetics of the four most stable isomers for small and medium - sized ni@xmath0 clusters with @xmath1 from 2 to 150
have been determined for each cluster size using a combination of the eam ( for calculation of the total energy for a given structure ) , the _ variable metric / quasi - newton _ method ( for the determination of local total - energy minima ) , and our own _
aufbau / abbau _ method ( for the determination of the global total - energy minimum ) .
except for two cases ( @xmath15 and @xmath18 ; here we included the decahedron as a starting configuration ) our structure - determination is completely unbiased .
in particular , as we shall see , it is important to go beyond @xmath9 ( the upper limit in the earlier unbiased structure - optimizations ) in order to identify growth patterns .
in addition , we shall present results for clusters that are obtained as relaxed spherical parts of the _ fcc _ crystal with @xmath1 up to some 1000s .
the paper is organized as follows . in sec .
ii we briefly outline the embedded - atom method , and in sec .
iii we present our structural - determination methods .
the main results are given in sec .
iv , and a brief summary is offered in sec .
v. finally , for the sake of completeness we mention that some preliminary results on clusters with up to @xmath19 and results for the energetically lowest isomer were presented earlier@xcite and in the discussion of the results we shall use ev and as units of energy and length , respectively .
we use the eam of daw , baskes , and foiles @xcite for the calculation of the total energy of a given cluster with a given structure .
the main idea of daw and baskes@xcite is to split the total energy of a ( metallic ) system of interest into two components .
the embedding energy is obtained by considering each atom as an impurity embedded into a host provided by the rest of the atoms .
the remaining part is written as a sum of pair potentials .
accordingly , @xmath20 where @xmath21 is the local electron density at site @xmath22 , @xmath23 is the embedding energy , i.e. , the energy required to embed an atom into this density , and @xmath24 is a short - range potential between atoms @xmath22 and @xmath25 separated by the distance @xmath26 , @xmath27 here the effective charges @xmath28 and @xmath29 depend on @xmath26 . the local electron density at site @xmath22 is assumed being a superposition of atomic electron densities @xmath30 where @xmath31 is the spherically averaged atomic electron density provided by atom @xmath25 at the distance @xmath26 . in the present approach , the atomic densities have been taken from hartree - fock calculations of refs . [ ] .
moreover , adjustable parameters that define @xmath23 and @xmath24 have been obtained by fitting to known bulk properties such as sublimation energy , lattice constant , the heat of solution of binary alloys and , additionally , to the universal equation of rose,@xcite which describes the sublimation energy of the most metals as a function of lattice constant .
the values for @xmath32 , @xmath23 and @xmath33 that we have used are available in numerical form for ni , cu , ag , au , pd and pt.@xcite our main reason for choosing the eam is that the eam provides a computationally efficient , parameterized many - particle method that allows for many calculations of also larger systems .
thus , with this method we can perform unbiased structure optimizations for clusters with well above 100 atoms , which is not possible using more accurate , parameter - free methods .
moreover , by comparing ( see sec .
iv a ) with results of more accurate studies for the smallest clusters we can estimate the accuracy of the method . since the method has been developed first of all for macroscopic metallic systems , in particular the smallest clusters may be those for which the largest inaccuracies show up .
using expression ( [ 1 ] ) we can calculate the total energy of any cluster with any structure as a function of atomic coordinates @xmath34 , @xmath35 . in order to obtain the closest local total - energy minimum we use the _ variable metric / quasi - newton _
method.@xcite we found that this was significantly more efficient than the _ conjugate gradient _ method . for searching the global minima
we have developed our own _
aufbau / abbau _ method .
it consists of the following steps : \1 ) we consider two cluster sizes with @xmath1 and @xmath36 atoms with @xmath37 . for each of those we study a set of randomly generated structures , @xmath38 . using the _ quasi - newton _ method
the @xmath39 relaxed structures are identified and the structures of the lowest total energy selected .
each of the @xmath39 starting structures for a cluster with @xmath40 atoms was generated using a random - number generator for positions within a sphere or a cube of volume @xmath41 , where @xmath42 is the nearest - neighbor distance of bulk ni and @xmath43 , i.e. we considered slightly compressed , normal , and slightly expanded structures .
we included the constraints that the smallest allowed inter - atomic distance was 0.5@xmath44 and each atom has to interact with at least two others .
\2 ) one by one , each of the @xmath40 atoms is displaced randomly , and the closest local minima is determined .
if the new structure has a lower total energy than the original one , this new one is kept , and the old one discarded .
this is repeated approximately @xmath45 times ( depending on cluster size ) .
\3 ) this leaves us with two ` source ' clusters , ni@xmath0 and ni@xmath46 with their lowest total energies .
one by one an atom is added at a random position to the structure with @xmath1 atoms ( many hundred times for each size ) , and the structures are relaxed . in parallel ,
one by one an atom is removed from the structure with @xmath36 atoms for each intermediate cluster with @xmath47 atoms we consider _ all _ @xmath48 possible configurations , that one can obtain by removing one atom from the ni@xmath49 cluster .
> from the two series of structures for @xmath50 those structures of the lowest energies are chosen and these are used as seeds for a new set of calculations .
first , when no lower total energies are found , it is assumed that the structures of the global - total - energy minima have been identified , and we proceed to larger clusters . moreover , by keeping information on not only the single energetically lowest isomer , but more low - lying ones , we have been able to identify the energetically four lowest - lying isomers .
our method combines randomness with regularity and is significantly more efficient compared to the case when only random starting geometries are used .
in addition , it is completely unbiased . however , as for any structure - determination method , there is no guarantee that we have identified the true global total - energy minima , although we believe that we in most cases have very good candidates .
but , unfortunately , in two cases , @xmath15 and @xmath18 ( both cases correspond to decahedral structures , which often have been difficult to find in unbiased structure optimizations ) , our approach failed and we had to include that in our study explicitly , and in another , @xmath51 , we only got the previously obtained structure after considerably more attempts than otherwise used . we add that similar problems have been observed and analysed in studies on lennard - jones clusters.@xcite
the smallest clusters , i.e. , ni@xmath0 for @xmath52 , provide an excellent test for our approach of two reasons : the clusters are so small that the number of ( meta-)stable structures is small , which in turn means that the global total - minimum within a given approach can be identified . and since the eam has its foundations in considerations for infinite , extended metallic systems with largely delocalized electrons , the small clusters should be those that with most difficulties can be treated by our approach .
therefore , we have in table [ tab00 ] tabulated our obtained structural parameters for @xmath53 and @xmath12 in comparison with results of other theoretical or experimental studies .
the table shows that the structure , here given through the average bond length @xmath54 , is most often given within a couple of % of the results of experiment or of parameter - free electronic - structure calculations .
moreover , none of the other parameterized methods , i.e. , the tight - binding approach or model potentials of the sutton - chen , gupta , or effective - medium type , gives results that are systematically more accurate than our approach does .
this is important since for the larger clusters that we shall discuss here it is not possible to apply accurate parameter - free methods when simultaneously attempting to optimize the structure .
the table also shows that the relative ordering of the different isomers may differ for different methods .
hence , for @xmath55 most parameterized methods predict the d@xmath56 isomer to have a lower total energy than the c@xmath57 structure , whereas experiment and parameter - free theoretical studies predict a reversed ordering .
therefore , it can be important to calculate more different , energetically low - lying isomers when studying theoretically the properties of not very small clusters identifying a different isomer as the one of the lowest total energy may suggest a significant disagreement between different studies . in order to emphasize this point
further we show in fig .
[ fig01 ] the structures of the two energetically lowest isomers for @xmath58 .
it is very clearly seen that in some cases the two isomers are indeed very different and identifying the ` wrong ' one as the most stable one may give unexpected results .
it may also surprise that for the two isomers for @xmath55 that look very different we find a total - energy difference of only 0.03 ev / atom . to our knowledge there exist two dft calculations of ni@xmath59 cluster.@xcite both studies predict for the two energetically lowest isomers a capped octahedron and a pentagonal bipyramid , respectively . the corresponding average binding energy per atom
are found to differ by only 0.050.07 ev / atom , i.e. , very similar to our findings .
table [ tab00 ] and fig . [ fig01 ] give one further result for @xmath55 : ni@xmath59 is the smallest cluster that incorporates the fivefold symmetry axis , an important element of icosahedral and decahedral symmetries .
for @xmath60 our two energetically lowest isomers differ in energy by 0.016 ev / atom , in good agreement with the value of 0.02 ev / atom found by desmarais _
_ in a dft study.@xcite there are some fundamental differences between different _ ab initio _ studies on ni@xmath61 cluster.@xcite according to one study @xcite a cube is the most stable structure for ni@xmath61 , whereas in other studies@xcite a cube was only the 5_th _ isomer .
further , in another work@xcite the second isomer was found to be a bicapped trigonal prism ( c@xmath62 ) and a capped pentagonal bipyramid was only isomer number four .
et al._@xcite has found for the eight - atom nickel cluster only one isomer with the tetrahedral symmetry t@xmath63 .
this gga result is inconsistent with all other studies .
it seems that the lsda calculations of desmarais _
et al._@xcite have the best agreement with both experimental and semiempirical studies . to our best knowledge
there are only dft studies@xcite on ni@xmath0 , @xmath64 , for @xmath12 . at @xmath12 the first mackay icosahedron ( i@xmath65 )
is constructed .
the second isomer is 1.16 ev less stable .
this is a rather large value and the icosahedron as the most stable structure for ni@xmath66 was found almost in all theoretical studies ( the only exception being an extended - hckel study@xcite ) .
once again we obtain good structural agreement with _
ab initio _
studies.@xcite the obtained bond lengths of 2.36 ( center to vertex ) and 2.48 ( vertex to vertex ) agree well with those derived by calleja _
et al._,@xcite i.e. , 2.41@xmath67 and 2.53@xmath67 , and with those of reuse and khanna,@xcite ,
i.e. , 2.23 and 2.34 , respectively
. furthermore , when comparing the differences in the binding energies per atom between the ni@xmath66 and ni@xmath68 clusters , our value of 1.525 ev / atom agrees very well with the corresponding value of 1.558 ev / atom obtained using the self - consistent dft program siesta.@xcite closing this subsection we emphasize that for our purpose , the calculation of energetic and structural properties of ni@xmath0 clusters , the present eam approach appears to be sufficiently accurate . in fig .
[ fig02 ] we show the binding energy per atom for the optimized clusters for the four energetically lowest isomers together with the binding energy per atom for clusters that were constructed as relaxed , spherical parts of the _ fcc _ crystal structure ( with an atom at the center of the sphere ) for @xmath1 up to almost 2600 . with few expectations
the four total - energy curves are so close that their difference can not be resolved in the figure . on the other hand , except for @xmath4 ,
fcc_-derived clusters lie all below the other structures , indicating that for @xmath1 up to around 150 the _ fcc _ structure is not playing a significant role as a structural motif ( an exception occurs for @xmath51 for a fragment of the _ fcc _ crystal with the center not at an atom this case is not shown in the figure ) .
[ fig02](b ) shows that the apparent saturation is not real : first when considering considerably larger clusters , the convergence towards an average binding energy around 4.45 ev / atom can be recognized . in order to identify the particularly stable clusters we shall consider two criteria .
first , a cluster may be considered very stable if its binding energy per atom is much larger than that of the two neighboring clusters .
this can be quantified through the stability function , @xmath69 , where @xmath70 is the total energy of the energetically @xmath71-lowest isomer of the ni@xmath0 cluster .
this function , that has maxima for particularly stable clusters , is shown in fig .
[ fig03 ] . here
we can identify a large number of particularly stable clusters , i.e. , so - called magic clusters .
these are found for @xmath1 = 4 , 6 , 10 , 13 , 15 , 19 , 23 , 26 , 28 , 32 , 36 , 39 , 43 , 46 , 49 , 55 , 58 , 61 , 64 , 67 , 71 , 75 , 77 , 79 , 83 , 86 , 89 , 92 , 95 , 99 , 101 , 104 , 108 , 110 , 114 , 116 , 119 , 122 , 125 , 127 , 129 , 131 , 137 , and 147 .
the most pronounced peaks occur at @xmath1 = 13 , 19 , 23 , 46 , 49 , 55 , 71 , 75 , 116 , 131 , 137 , and 147 and most of them contain an odd number of atoms and , as seen in table [ i ] , these clusters possess high symmetry , including the highest icosahedral symmetry for ni@xmath66 , ni@xmath72 , and ni@xmath3 . according to our other criterion for a particularly stable cluster ,
such a cluster occurs if the energy difference between the two energetically lowest isomers @xmath73 is large .
this energy difference together with the energy differences to the energetically higher ones are shown in fig .
[ fig04 ] .
it is striking that the energetically higher - lying isomers become energetically less and less separated and , as seen when comparing with fig .
[ fig03 ] , many of the clusters that are particularly stable according to the first criterion is it also according to the second one . a central issue of our work is to identify general properties of the ni@xmath0 clusters from the large amount of information that we have obtained from the calculations . instead of discussing the individual clusters , in particular their structure ( a brief summary of the structures
is given in the appendix ) , we shall introduce different quantities that are devised to reduce the available information to some few key numbers .
first we shall consider the overall shape of the clusters . as we discussed in our earlier report,@xcite
it is useful to examine the @xmath74 matrix containing the elements @xmath75 with @xmath76 = 1 being our length unit , and @xmath77 and @xmath78 being @xmath79 , @xmath80 , and @xmath81 , and with @xmath82 being the center of the cluster .
the three eigenvalues of this matrix , @xmath83 , can be used in separating the clusters into being overall spherical ( all eigenvalues are identical ) , more cigar - like shaped ( one eigenvalue is large , the other two are small ) , or more lens - shaped ( two large and one small eigenvalue ) . moreover ,
the average of the three eigenvalues , @xmath84 , is a measure of the overall extension of the cluster . finally , for a homogeneous sphere with @xmath1 atoms , the eigenvalues scale like @xmath85 .
therefore , we show in fig
. [ fig05 ] quantities related to these eigenvalues but scaled by @xmath86 . in the figure we see that only some few clusters have an overall spherical shape ( these are found for the energetically lowest isomer for @xmath1 = 4 , 6 , 13 , 26 , 28 , 38 , 55 , and 147 and for the next one for @xmath1 = 54 , 79 , and 92 ) , which all correspond to high - symmetry isomers ( cf . table [ i ] ) and , for the lowest - energy isomer , also to the class of magic clusters .
moreover , it is interesting to notice that the average value follows more or less the same curve for all four isomers , i.e. , the more compact ones ( with relative low average value ) are found for @xmath1 just above 50 and above 140 ) .
also the largest differences follow more or less the same curve , except for some few cases mainly for @xmath1 below 40 . therefore , except when the eigenvalues all are very similar ( i.e. , the largest difference is very small , which occurs for @xmath1 around 50 , 100 , and 140 ) , the overall shape ( i.e. , lens- or cigar - like ) is the same for all four isomers . in our earlier work@xcite we showed that the structural development of the isomer of the lowest total energy with advantage could be monitored through so - called similarity functions . thus ,
for a given cluster ni@xmath0 we define a radial distance for each of the atoms @xmath87 and sort these in increasing order .
simultaneously we consider a large fragment of a _ fcc _ crystal or , alternatively , a large cluster of icosahedral symmetry .
also for these we define a radial distance for each atom , @xmath88 , that also are sorted . in order to compare the given cluster with any of those systems we calculate subsequently @xmath89^{1/2},\ ] ] giving the similarity function @xmath90 ( @xmath76 = 1 ) , which approaches 1 if the ni@xmath0 cluster is very similar to the reference system . in fig .
[ fig11 ] we show the resulting functions in four cases , i.e. , when comparing with the relaxed ni@xmath91 cluster of icosahedral symmetry , and when comparing with three fragments of the _ fcc _ crystal differing in the position of the center ( the position of an atom , the middle of a nearest - neighbor bond , and the center of the cube , respectively ) . we see that both clusters that clearly resemble _ fcc _ fragments and such that resemble icosahedral clusters can be identified . of obvious reasons ,
the latter is the case for @xmath1 around 13 , 55 , and 147 , but it is interesting to notice that these clusters are not singular : clusters with a large similarity to icosahedral clusters are found over a larger range of clusters ( @xmath9210 ) around the truly icosahedral cluster .
similarly , clusters that are close to being fragments of the _ fcc _ crystal can be identified around @xmath1 = 13 , 28 , 38 , and 79 .
moreover , it is surprising to observe that all four isomers for a given size show very similar similarity functions .
also the radial distances themselves can give some interesting information .
these are shown in fig .
[ fig10 ] for the four different isomers as functions of cluster size .
up to @xmath1 around 40 , only few trends can be identified ( with the exception of @xmath12 ) , but for @xmath1 just above 50 a clear tendency towards shell - construction can be seen . whereas this transition towards shell - construction is fairly smooth as a function of cluster size , independent of whether we start from the larger or the smaller clusters , in the range @xmath93 abrupt changes occur also towards some shell - construction . that this size range is special shall be discussed in more detail below .
also for @xmath1 close to 115 this tendency is seen , but most dramatic are the results for clusters around 140 and above where , for all four isomers , a very clear shell - construction is identified . comparing with fig .
[ fig11 ] we see that this is correlated to the construction of icosahedral - like structures .
thus , since the infinite crystal possesses the _ fcc _ structure , the clear shell - construction in this size range will have to disappear again for larger clusters . in total
, it can not surprise that when analysing the structures of the individual clusters in detail we can identify many structural motifs that are characteristic for clusters of particularly high stability also for clusters with somewhat different sizes .
moreover , such an analysis@xcite shows also that many of our structures for the smaller @xmath1 ( these are the only ones that have been the subject of earlier unbiased structure - optimizations ) are close to those that have been found in earlier studies . here
, however , we shall largely omit the discussion of the individual structures but instead explore the more general trends . in fig .
[ fig06 ] we show some few quantities related to the structure of the clusters , i.e. , coordination numbers and average bond lengths . here , we define two atoms as being bonded if their interatomic distance is less than 3.00 , which is the average of the nearest - neighbor distance ( 2.49 ) and the next - nearest - neighbor distance ( 3.52 ) in bulk ni . moreover , we distinguish between inner atoms with a coordination number of 12 or larger and surface atoms with a coordination number less than 12 .
[ fig06](a ) shows the average coordination number as function of @xmath1 , where an obvious increase as function of size is observed , but with additional oscillations in particular for the smallest clusters .
this function shows clearly a saturation towards the bulk limit of 12 , although one has to remember that even for the largest cluster of our study 94 out of 150 atoms are characterized as surface atoms .
[ fig06](b ) shows the minimum atomic coordination for each cluster size from the range @xmath94 , which is useful information for chemical - probe experiments due to the enhanced chemical activity of clusters containing low - coordination atoms .
in addition , the existence of low - coordination atoms ( i.e. , coordinations of 3 or 4 ) could point to the occurrence of a cluster growth , where extra atoms are added to the surface of the cluster , whereas the higher minimum - atomic - coordination numbers ( _ e.g. _ 5 or 6 ) could indicate a growth where atoms are inserted inside the cluster , or , alternatively , upon a strong rearrangement of the surface atoms .
these considerations follow the so - called `` maximizing the minimum coordination''.@xcite fig .
[ fig06](c ) shows the average bond length as function of cluster size .
the dashed line corresponds to the bulk value of 2.49 . the average bond length for all the structures with the exception of ni@xmath95 , ni@xmath96 , ni@xmath97 , and ni@xmath98 is smaller than the bulk value .
such a reduction is in accordance with the results of most other theoretical studies , but it is interesting to observe that this property approaches the bulk value faster than the other quantities of the figure .
important experimental work in the determination of the structure of individual ni@xmath0 clusters has been presented by riley @xmath99 @xmath100@xcite using chemical - probe experiments .
thereby , the number of n@xmath101 molecules that is bonded to a given cluster is determined , and in order to relate this number to the structure of the cluster , the following empirical rules have been formulated:@xcite @xmath102 nickel atoms with a coordination of four and less will bind two nitrogen molecules at saturation , @xmath103 nickel atoms with coordination of five to eight will bind only one nitrogen molecule at saturation , @xmath104 nickel atoms with a coordination of nine will bind one nitrogen molecule weakly or not at all ( this kind of binding has not been observed@xcite for @xmath105 ) , and @xmath106 nickel atoms with a coordination more than nine do not bind any nitrogen molecules .
these rules allow us in turn to estimate the number of n@xmath101 molecules that will bond to a given cluster .
we stress that the chemical - probe experiments provide information on the complete system , i.e. , cluster plus adsorbants , whereas we consider the naked clusters .
although this difference may result in some deviations between theory and experiment , a comparison should give useful information on the clusters .
this is supported by the studies of parks _
et al._@xcite that show that in many cases the structures of naked and n@xmath101-covered clusters are very similar .
doing so , we obtain the results of fig .
[ fig07 ] , where we also compare with the results of other theoretical works .
the experimental work covers the range @xmath107 , whereas the former theoretical works cover @xmath108 ( ref .
[ ] ) , @xmath109 ( ref . [ ] ) , and @xmath110 ( ref .
[ ] ) , obtained using the sutton - chen potential,@xcite the finnis - sinclair potential,@xcite and the effective - medium theory.@xcite fig .
[ fig07](a ) , where we show the results for the whole range of our study , i.e. , @xmath111 , shows an interesting aspect that obviously is beyond most of the previous experimental and theoretical studies , i.e. , that up to around @xmath112 the number of binding sites increases more rapidly with @xmath1 than for @xmath1 above @xmath9255 , i.e. , the curve is roughly two piecewise linear curves .
the experimental studies [ cf . figs .
[ fig07](d ) and ( e ) ] do in fact show a small tendency in this direction , which , on the other hand , is not found in the earlier theoretical work [ cf .
[ fig07](c ) ] . figs .
[ fig07](b)-(e ) show a comparison of the experimental results with those of the four different theoretical studies to this issue .
it is clear that our approach is at least as accurate as any of the other theoretical ones .
moreover , the sutton - chen potential seems to predict a larger scatter of the number of binding sites as a function of size compared to any other approach .
whereas icosahedral motifs are found to play an important role for very many values of @xmath1 , we found that for @xmath1 around 75 also other structural motifs are important .
the chemical - probe results for n@xmath101 uptake , both the experimental ones ( notice , however , that the experimental studies are limited to @xmath113 ) and the simulated , theoretical ones , suggest also that in this range of @xmath1 values more structural motifs compete : for these values of @xmath1 the results of fig .
[ fig07 ] are particularly scattered .
moreover , as we shall see in the next subsection , also growth processes suggest that these @xmath1 values are special .
therefore , we shall study the structural properties of those clusters in some more detail . for cluster sizes ranging between 70 and 80 atoms two more closed - shell structures other than icosahedral ones can be identified , i.e. , the ( 2,2,2 ) marks decahedron ( giving a 75-atom cluster ) and a 79-atom _ fcc _ cluster , which is the truncated centered octahedron . due to their quasi - spherical shape
they have a particularly low surface energy . moreover , the cluster sizes between 70 and 80 are very different from the values of closed - shell icosahedral clusters that have 55 or 147 atoms .
accordingly , one may expect a strong competition between these three main morphologies in this size range .
our calculations support this suggestion .
[ fig08 ] shows the lowest total energies for the decahedral and _ fcc _ clusters relative to the icosahedral ones for @xmath114 .
the marks decahedron and its derivatives dominate over the icosahedral structures for @xmath115 . at @xmath4 both decahedral and _ fcc
_ motifs are more stable than the icosahedral one .
the decahedron at @xmath4 is only 0.04 ev more stable than the truncated octahedron .
finally , three of four lowest - energy isomers at @xmath15 are decahedra and for @xmath116 there are only decahedral clusters among the four first isomers .
this result suggests that a fundamental structural change takes place for @xmath1 around 75 .
this , however , is only partly the case : for larger @xmath1 values the icosahedral packing is again stabilized ( an exception occurs for @xmath1 around 104 , where decahedral structures are found ) . in order to verify this we computed the total energies for several optimized closed _
fcc _ and decahedral structures , i.e. , truncated noncentered octahedra for @xmath117 and 140 , a noncentered octahedron and a ( 3,2,2 ) marks decahedron for @xmath118 , and a cuboctahedron and an ino decahedron for @xmath119 . in all cases the icosahedral structures
were found to have lower total energies .
the _ fcc _ clusters for @xmath117 , 140 and 146 and _ m_-dh@xmath120 were 0.85 , 2.01 , 5.64 and 2.29 ev , respectively , less stable than the icosahedral - like structures .
the three - shell mackay icosahedron was 5.69 and 4.71 ev more stable than the cuboctahedron and the ino decahedron , accordingly .
thus , our calculations point strongly to icosahedral growth up to cluster sizes of at least 150 .
this statement is supported by mass - spectroscopy studies of nickel clusters@xcite which suggest icosahedral growth over a wide cluster - size region from 50 to 800 .
one of the central issues is to identify how clusters grow , i.e. , whether some fundamental principles determine how ( and whether ) the cluster with @xmath1 atoms is derived from the one with @xmath121 atoms .
we shall here study this issue from a completely static point of view , i.e. , dynamic and kinetic effects , that most likely are important in experiment , will be neglected here .
basic information about this issue is given by the structures of the magic clusters .
the majority of the structures of these are icosahedral .
thus our study supports predominantly icosahedral growth of nickel clusters .
there are two different ways of icosahedral growth .
thus , the first mackay icosahedron , ih@xmath66 , contains 20 triangular faces joined by 30 edges and 12 vertices .
in the first case , mic / mackay or _ fcc _
covering,@xcite additional surface atoms are placed on top of the edges and the vertices .
this kind of growth results ultimately in the second mackay icosahedron , ih@xmath72 .
another possibility , the tic / polyicosahedral@xcite or _ hcp _ or face - capping growth ( see ref . [ ] and references therein ) , corresponds to adding atoms on top of the atoms at the center of each face ( @xmath122 sites ) which leads , after complete covering , to the 45-atom rhombic tricontahedron .
the tic structures are the structures with the shorter average bond length , but at the same time they possess the larger strain energy .
they are favored at the beginning of a covering until some critical size ( depending on details of the atom - atom interaction@xcite ) after which the mic structures become lower in energy . to explain the icosahedral shell filling in mg and ca clusters martin _
et al._@xcite proposed the so - called _ umbrella _ model . according to this model enhanced stability
is achieved each time a vertex and its five surrounding faces are covered ( capping umbrellas ) .
the capping umbrellas for mic and tic structures lead to different magic numbers . by covering of ih@xmath66
the mic umbrellas are completed at @xmath123 , 24 , 28 , 32 , 36 , 39 , 43 , 46 , 49 and 55 , whereas the tic umbrellas give magic numbers for 19 , 23 , 26 , 29 , 32 and 34 . comparing with our set of magic numbers ( see above ) , we find that growth from ni@xmath66 to ni@xmath124 is icosahedral and it can be described according to the umbrella model . for @xmath125
our magic numbers coincide with the mic magic figures .
this confirms seemingly the suggestion@xcite that at @xmath126 there is a transition from tic to mic growth .
however , a detailed analysis of our structures shows that our structures for ni@xmath127 and ni@xmath128 are not icosahedral .
instead , the most stable icosahedral structure appears first at @xmath129 whereas the structures according to the mic growth or to _ fcc _ packing is found up to @xmath130 with a single exception for @xmath51 .
thus the transition from anti - mackay to mackay covering occurs in the region of cluster sizes from 28 to 34 .
cluster growth in this range is complicated and there is a competition between icosahedral , polyicosahedral and tetrahedral structure .
probably this non - trivial growth was responsible for the difficulties in interpreting results of nitrogen - uptake experiments@xcite for @xmath131 .
at @xmath13 we have the next complete icosahedron .
according to the tic growth model we should then have magic number at @xmath13258 , 61 , 64 , 67 , and 71,@xcite which is in excellent agreement with our results . at @xmath133 the tic umbrella growth is completed .
the mic umbrella growth is completed at @xmath133 , 83 , 92 , 101 , 110 , 116 , 125 , 131 and 137 , again in agreement with our calculated magic numbers .
our study allows a more detailed analysis of possible growth processes that ultimately will confirm the complexity of the growth .
thus , in order to quantify whether the cluster of @xmath1 atoms is related to that of @xmath121 atoms we consider first the structure of the lowest total energy for the @xmath5-cluster . for this
we calculate and sort all interatomic distances , @xmath134 , @xmath135 .
subsequently we consider the @xmath1 fragments of the @xmath1-cluster that can be obtained by removing one of the atoms and keeping the rest at their positions .
for each of those we also calculate and sort all interatomic distances @xmath136 , and subsequently calculate @xmath137^{1/2}.\ ] ] of the @xmath1 different values of @xmath138 we choose the smallest one , @xmath139 and calculate the similarity function @xmath140 ( @xmath76 = 1 ) which approaches 1 if the ni@xmath0 cluster is very similar to the ni@xmath141 cluster plus an extra atom . this function
is shown in fig .
[ fig09](a ) .
we see indeed that for @xmath1 up to around 40 , @xmath142 is significantly different from 1 , giving support for the consensus that in this range the growth is complicated . for larger
@xmath1 we see again that around @xmath143 some larger structural changes take place .
[ fig03 ] shows , however , that the energetically lowest isomers differ only little in energy .
therefore , one may ask whether isomers with a slightly higher energy play a role in the growth process .
accordingly , we examined whether the structure of the energetically lowest isomer of ni@xmath0 would resemble @xmath144 of the four energetically lowest isomers for ni@xmath141 in the same sense as above , i.e. , we use the quantity @xmath138 in quantifying the structural difference . out of the four different values for @xmath138 ( for the four different isomers ) we choose the smallest one and construct the similarity function from this . the resulting function is shown in fig .
[ fig09](b ) , and compared with fig . [ fig09](a ) it is clearly seen that in particularly in the region @xmath145 the similarity function now approaches 1 , except for the clusters around @xmath146 , i.e. , the region where we already in analysing other properties have found atypical behaviors . in fig .
[ fig09](c ) we show which of the four energetically lowest isomers of ni@xmath141 leads to largest similarity with the ni@xmath0 cluster , and from the fact that this only for the smaller @xmath1 most often is the first isomer , we learn that the growth process is a complicated process where more different isomers are important , i.e. , not only the energetically most stable one , although these may resemble each other ( cf .
[ fig05 ] ) .
this means that it is difficult to imagine that the growth occurs as a one - by - one addition of atoms to the energetically most stable isomers .
in summary , we have determined the four energetically lowest isomers of nickel clusters ni@xmath0 in the range @xmath147 , using a combination of the embedded - atom method in the version of daw , baskes and foiles ( for the calculation of the total energy for a given structure ) , the _ variable metric / quasi - newton _ method ( for the determination of the closest total - energy minimum ) , and our own _
aufbau / abbau _ method ( for the determination of the global total - energy minimum ) .
although the calculations provide structural information for each individual cluster , separately,@xcite that may be useful information for experimental studies , we have refrained from discussing these separately ( some information is , however , given in the appendix ) , but instead focused on identifying general trends for ni@xmath0 clusters , and reducing the information on the individual clusters to some few key quantities like total energy ( per atom ) , overall symmetry and shape , average bond length and coordination number , and similarity with other structures ( _ fcc _ fragments , icosahedral clusters , or smaller clusters ) .
since our total - energy approach ( the embedded - atom method ) has its roots in arguments for extended solids with delocalized atoms , we first verified that it could be applied also for the fairly open structures of our interest .
we found that for the absolutely most ` critical ' cases , i.e. , the smallest clusters , our results are in excellent agreement with available theoretical and experimental information making us believe that our approach is accurate .
furthermore , we have also tried to use the parameterization of voter and chen@xcite of the eam , which has been produced in order to give accurate descriptions of also finite and more open systems , for some of the smallest clusters . to our satisfaction ,
no structural changes were found and , therefore , we did not pursue this approach . our study predicts a number of particularly stable clusters , i.e. , magic numbers , which in many cases are in agreement with the prediction of other studies and in some cases with ` common expectations ' although we stress that in our study they were found in a completely unbiased approach .
these magic numbers were clearly visible both in the ` stability function ' and in the total - energy difference between the energetically lowest and higher - lying isomers , whereas for the energetically higher - lying isomers , the differences in the total energy become small .
we found also that even for large values of @xmath1 ( above 2500 ) the total energy per atom has still not converged to the bulk limit .
similarly , the average coordination number is for clusters with @xmath1 around 150 still far from the bulk limit , whereas the average bond length has come close to the bulk limit for this size . in order to study the structural properties we analysed the eigenvalues of the matrix containing the moments of inertia and introduced so - called similarity functions .
the latter indicated that all the four isomers we have studied here are very similar in most cases , and that roughly spherical clusters were found mainly for the energetically lowest isomer but in some cases also for the second - lowest one , and that these often correspond to particularly stable structures . with the similarity functions we could identify clusters with a predominantly icosahedral structure and such with an essentially _ fcc_-like structure .
whereas the former occurred for @xmath1 around 13 , 55 , and 147 , the latter was found , e.g. , for @xmath1 around 79 .
it was interesting to see that these structural motifs were built up over a larger range of @xmath1 values and were , accordingly , not limited to those singular values of @xmath1 . by analysing the distribution of radial distances as a function of cluster size
we could identify certain size regions , e.g. , @xmath1 around 50 , 75 , 115 , and above 140 , where shell structures were clearly recognizable .
this property , as well as several others , showed , furthermore , that all four isomers for a given @xmath1 have very similar structures .
chemical - probe experiments have been used in identifying the structure of the individual clusters , and for this we found an excellent agreement .
our study shows that for clusters with @xmath1 around 5060 , a change occurs , so that above this size the number of n@xmath101 binding sites depends much weaker on @xmath1 than below this size .
this change is just on the boarder of the cluster sizes that have been studied experimentally and theoretically so far .
we used the magic numbers in analysing different models for cluster - growth processes .
generally , the cluster growth is according to multilayered - icosahedral or layer - by - layer growth .
we found that the one- , two- , and three - shell mackay icosahedra are the most stable structures for 13- , 55- , and 147-atom nickel clusters ( the latter for the first time in unbiased studies ) .
this result agrees well with the suggestion in ref . that mackay icosahedra are the most stable structures up to @xmath148 .
but the cluster - growth pattern between two closed - shell icosahedra is not simple .
it is predominantly icosahedral with islands of _ fcc _ growth for ni@xmath149 , ni@xmath14 , tetrahedral for ni@xmath150 and decahedral for ni@xmath151-n@xmath152 and ni@xmath153 .
for @xmath154 the growth is polyicosahedral .
moreover , we found a growth according to _ hcp_/anti - mackay for @xmath155 and for @xmath156 . from @xmath129 to @xmath130 ( with the exception of @xmath51 ) and from @xmath9 to @xmath119 ( except for @xmath18 ) the mic cluster growth was found .
thus , the change - over from the _ hcp_-type of cluster growth to the _ fcc _ one always happens within a transition region of cluster sizes and not immediately .
moreover , whereas the growth process only with difficulties could be identified for the smaller @xmath1 ( due to the lack of similarity between the clusters ) , for larger @xmath1 we found that not only the energetically lowest isomers could be important in describing the growth .
thus , for all ranges of @xmath1 , the growth is non - trivial .
a special case is that for @xmath1 around 75 .
more different properties suggested that in this range special things occur . analysing the total energy for different structures in this size range we found indeed that more structures were energetically close and , accordingly , would play a role for those clusters .
therefore , for these values of @xmath1 there is a rapid and dramatic change in structure as a function of @xmath1 .
we gratefully acknowledge _ fonds der chemischen industrie _ for very generous support .
this work was supported by the sfb 277 of the university of saarland .
the authors are grateful to c. l. cleveland , to z. b. gven and m. bykata , and to k. michaelian and i. l. garzn for providing us with the coordinates of some cluster structures .
we would like to thank j. p. k. doye for references to some recent studies on nickel clusters .
in this appendix we shall present the structure of some of the clusters without entering a detailed discussion of each clusters individually .
we shall use the notation @xmath157 for the ni@xmath0 cluster of the @xmath71-lowest total energy .
we add that the nuclear coordinates and total energies for the optimized clusters can be obtained from the authors upon request . 3.1 : equilateral triangle . 4.1 : tetrahedron . 5.1 trigonal bipyramid . 6.1 octahedron .
6.2 : pentagonal bipyramid without a single pentagonal vertex . 7.1 : pentagonal bipyramid . 7.2 : capped octahedron . 8.1 : bisdisphenoid . 8.2 : capped pentagonal bipyramid . 9.1 : bicapped pentagonal bipyramid . 9.2 : tricapped trigonal prism . 10.1 : tricapped pentagonal bipyramid . 10.2 : trapezoidal antiprism . 11.1 : quadricapped pentagonal bipyramid . 11.2 : two pentagonal bipyramids with a common face . 12.1 : icosahedron without one atom at one vertex . 12.2 : low symmetry . 12.3 : trapezoidal bipyramid with capped trigonal faces . 12.4 : truncated trigonal bipyramid . 13.1 : mackay icosahedron .
13.2 , 13.3 , 13.4 : similar structures as 13.1 , but with the move of a vertex atom to a face .
14.1 : the structure of 13.1 with the capping of one of the trigonal faces . 14.2 : similar , but with the 14th atom on the top of a pentagonal edge . 14.3 : bicapped hexagonal antiprism . 14.4 : tricapped pentagonal prism . 15.1 : bicapped icosahedron . 15.2 : centered bicapped hexagonal antiprism . 15.3 , 15.4 : capped pentagonal prism . 16.1 : tricapped icosahedron . 18.1 : double icosahedra without one atom in one pentagonal ring . 18.2 : double icosahedra without one atom in a capping vertex . 19.1 : double icosahedron
. 19.2 , 19.3 , 19.4 : derivatives of the double icosahedron by moving of an atom from an apex or from a pentagonal ring to a face - capped or a bridging position . 20.1 : double icosahedron with an atom added to the a pentagon in a bridging position . 21.1 , 21.2 , 21.3 , 21.4 : derivatives of the double icosahedron with two additional atoms at bridging or capping positions
. 22.1 , 22.2 : triple icosahedron without one atom either from a pentagonal ring ( 22.1 ) or from a capping position ( 22.2 ) . 23.1 : triple icosahedron . 24.1 , 24.2 : triple icosahedron with an additional atom .
25.1 : three sequentially interpenetrating icosahedra whose vertices are located on one line .
26.1 : six interpenetrating double icosahedra ( a polyicosahedron ) . 27.1 : the polyicosahedron plus an extra atom at a pentagonal cap . 28.1 : tetrahedron . in general , the structures for @xmath158 do not display any dominant morphology
there is a competition between tetrahedral , polyicosahedral , and icosahedral packings .
one can consider this size range as a transition range from _ hcp _ to _ fcc _ growth . 29.1 : distorted tetrahedron . 30.1 : a polyicosahedral morphology . 32.1 : derived from the 28.1 tetrahedron by adding four surface atoms . 35.1 : mic / mackay structure . 36.1 , 37.1 : contain icosahedral motifs . 38.1 : truncated octahedron . 38.2 , 38.3 : icosahedral fragments . 39.1 : the second mackay icosahedron ( 55.1 ) without a 16-atom cap . 39.4 : a _ fcc _ cluster . 49.1 : the second mackay icosahedron ( 55.1 ) without one face . 51.1 , 52.1 , 53.1 , 54.1 : obtained from 49.1 by adding atom by atom until 55.1 : second mackay icosahedron . 54.2 : as 55.1 but without the central atom . 56.1 : as 55.1 but with an extra atom at the center of a triangular face .
@xmath159 : tic / anti - mackay growth in accordance with the umbrella model .
z. b. gven , j. jellinek and a. f. voter , in _ physics and chemistry of finite systems : from clusters to crystals _ , edited by p. jena ,
s. n. khanna and b. k. rao ( kluwer , dordrecht , 1992 ) , vol i , p. 411 . a. f. voter and s. p. chen , in _ characterization of defects in materials _ , edited by r. w. siegal , j. r. weertman , and r. sinclair , mrs symposia proceedings no .
82 ( materials research society , pittsburgh , 1987 ) , p. 175 .
.the symmetry ( sym . ) and averaged bond length ( @xmath160 ) for smaller ni@xmath0 clusters from different experimental and theoretical studies .
eam , tbmd , fs , sc , ai , emt , @xmath8 g , and exp denotes embedded - atom calculations , tight - binding molecular - dynamics calculations , finnis - sinclair potential , sutton - chen potential , _ ab initio _ calculations , effective - medium - theory calculations , many - body gupta potential , and experiment , respectively . for some values of @xmath1 , more different meta - stable isomers have been found , and they have then been characterized by an additional number ( e.g. , 7.1 , 7.2 , 7.3 , and 7.4 , with 7.1 being the stabler one , 7.2 the second - most stable one , a.s.o . ) [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] cllllcllllcllllcllll @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 + 2 & d@xmath166 & & & & 40 & c@xmath167 & c@xmath167 & c@xmath168 & c@xmath169 & 78 & c@xmath168 & c@xmath168 & c@xmath167 & c@xmath167 & 116 & c@xmath170 & c@xmath168 & c@xmath168 & c@xmath168 + [ -1.3ex ] 6 & o@xmath172 & c@xmath62 & & & 44 & c@xmath168 & c@xmath167 & c@xmath168 & c@xmath167 & 82 & c@xmath168 & c@xmath167 & c@xmath167 & c@xmath168 & 120 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath167 + [ -1.3ex ] 7 & d@xmath174 & c@xmath175 & c@xmath169 & c@xmath175 & 45 & c@xmath167 & c@xmath167 & c@xmath168 & c@xmath168 & 83 & c@xmath62 & d@xmath176 & c@xmath169 & c@xmath167 & 121 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath168 + [ -1.3ex ] 8 & d@xmath177 & c@xmath167 & d@xmath178 & c@xmath62 & 46 & c@xmath62 & c@xmath168 & c@xmath168 & c@xmath168 & 84 & c@xmath167 & c@xmath168 & c@xmath168 & c@xmath168 & 122 & c@xmath168 & c@xmath168 & c@xmath167 & c@xmath168 + [ -1.3ex ] 9 & c@xmath62 & d@xmath171 & c@xmath62 & c@xmath167 & 47 & c@xmath168 & c@xmath167 & c@xmath167 & c@xmath167 & 85 & c@xmath168 & c@xmath167 & c@xmath168 & c@xmath168 & 123 & c@xmath167 & c@xmath167 & c@xmath168 & c@xmath168 + [ -1.3ex ] 10 & c@xmath175 & d@xmath179 & c@xmath169 & c@xmath62 & 48 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 & 86 & c@xmath176 & c@xmath167 & c@xmath168 & c@xmath168 & 124 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 + [ -1.3ex ] 11 & c@xmath62 & c@xmath169 & c@xmath62 & c@xmath169 & 49 & c@xmath175 & c@xmath168 & c@xmath167 & c@xmath167 & 87 & c@xmath168 & c@xmath168 & c@xmath167 & c@xmath167 & 125 & c@xmath167 & c@xmath168 & c@xmath168 & c@xmath168 + [ -1.3ex ] 12 & c@xmath170 & c@xmath168 & d@xmath177 & d@xmath171 & 50 & c@xmath167 & c@xmath167 & c@xmath62 & c@xmath167 & 88 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 & 126 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 + [ -1.3ex ] 13 & i@xmath172 & c@xmath167 & c@xmath167 & c@xmath167 & 51 & c@xmath62 & c@xmath167 & c@xmath167 & c@xmath169 & 89 & c@xmath175 & c@xmath167 & c@xmath168 & c@xmath167 & 127 & c@xmath62 & c@xmath167 & c@xmath168 & c@xmath168 + [ -1.3ex ] 14 & c@xmath175 & c@xmath62 & c@xmath167 & c@xmath62 & 52 & c@xmath175 & c@xmath167 & c@xmath167 & c@xmath167 & 90 & c@xmath167 & c@xmath168 & c@xmath167 & c@xmath168 & 128 & c@xmath168 & c@xmath168 & c@xmath167 & c@xmath167 + [ -1.3ex ] 15 & c@xmath62 & d@xmath180 & c@xmath62 & c@xmath62 & 53 & c@xmath62 & d@xmath181 & c@xmath62 & c@xmath167 & 91 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 & 129 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath169 + [ -1.3ex ] 16 & c@xmath167 & c@xmath167 & c@xmath169 & d@xmath171 & 54 & c@xmath170 & i@xmath172 & c@xmath62 & c@xmath167 & 92 & c@xmath175 & t & c@xmath168 & c@xmath168 & 130 & c@xmath168 & c@xmath167 & c@xmath167 & c@xmath168 + [ -1.3ex ] 17 & c@xmath169 & c@xmath167 & c@xmath167 & c@xmath175 & 55 & i@xmath172 & c@xmath167 & c@xmath167 & c@xmath167 & 93 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath168 & 131 & c@xmath62 & c@xmath168 & c@xmath168 & c@xmath168 + [ -1.3ex ] 18 & c@xmath167 & c@xmath170 & c@xmath167 & c@xmath168 & 56 & c@xmath175 & c@xmath175 & c@xmath167 & c@xmath168 & 94 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath168 & 132 & c@xmath168 & c@xmath168 & c@xmath167 & c@xmath168 + [ -1.3ex ] 19 & d@xmath174 & c@xmath168 & c@xmath167 & c@xmath168 & 57 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath167 & 95 & c@xmath168 & c@xmath176 & c@xmath168 & c@xmath168 & 133 & c@xmath168 & c@xmath167 & c@xmath167 & c@xmath168 + [ -1.3ex ] 20 & c@xmath62 & d@xmath178 & d@xmath169 & d@xmath169 & 58 & c@xmath175 & c@xmath167 & c@xmath168 & c@xmath168 & 96 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath62 & 134 & c@xmath175 & c@xmath167 & c@xmath168 & c@xmath168 + [ -1.3ex ] 21 & c@xmath168 & c@xmath62 & c@xmath62 & c@xmath167 & 59 & c@xmath62 & c@xmath168 & c@xmath168 & c@xmath62 & 97 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath62 & 135 & c@xmath167 & c@xmath168 & c@xmath167 & c@xmath167 + [ -1.3ex ] 22 & c@xmath167 & c@xmath167 & c@xmath168 & c@xmath167 & 60 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 & 98 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath167 & 136 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 + [ -1.3ex ] 23 & d@xmath171 & c@xmath169 & c@xmath167 & c@xmath168 & 61 & c@xmath62 & c@xmath167 & c@xmath167 & c@xmath168 & 99 & c@xmath62 & c@xmath168 & c@xmath167 & c@xmath167 & 137 & c@xmath175 & c@xmath62 & c@xmath62 & c@xmath168 + [ -1.3ex ] 24 & c@xmath62 & c@xmath167 & d@xmath176 & c@xmath167 & 62 & c@xmath62 & c@xmath167 & c@xmath168 & c@xmath168 & 100 & c@xmath167 & c@xmath167 & c@xmath170 & c@xmath168 & 138 & c@xmath175 & c@xmath167 & c@xmath168 & c@xmath167 + [ -1.3ex ] 25 & c@xmath62 & c@xmath176 & d@xmath181 & c@xmath167 & 63 & c@xmath168 & c@xmath167 & c@xmath168 & c@xmath167 & 101 & c@xmath62 & d@xmath174 & c@xmath167 & c@xmath168 & 139 & c@xmath62 & c@xmath167 & c@xmath167 & c@xmath167 + [ -1.3ex ] 26 & t@xmath173 & c@xmath167 & c@xmath168 & c@xmath167 & 64 & c@xmath167 & c@xmath168 & c@xmath168 & c@xmath62 & 102 & c@xmath167 & c@xmath168 & c@xmath168 & c@xmath62 & 140 & c@xmath167 & c@xmath167 & c@xmath168 & c@xmath167 + [ -1.3ex ] 27 & c@xmath62 & c@xmath167 & c@xmath167 & c@xmath167 & 65 & c@xmath168 & c@xmath169 & c@xmath168 & c@xmath168 & 103 & c@xmath168 & c@xmath167 & c@xmath168 & c@xmath168 & 141 & c@xmath170 & c@xmath175 & c@xmath169 & c@xmath168 + [ -1.3ex ] 28 & t & d@xmath169 & c@xmath175 & c@xmath167 & 66 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath168 & 104 & c@xmath62 & c@xmath168 & c@xmath167 & c@xmath168 & 142 & c@xmath167 & c@xmath167 & c@xmath170 & c@xmath167 + [ -1.3ex ] 29 & c@xmath176 & d@xmath171 & c@xmath62 & c@xmath167 & 67 & c@xmath169 & c@xmath167 & c@xmath62 & c@xmath167 & 105 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath167 & 143 & c@xmath62 & c@xmath167 & c@xmath167 & c@xmath167 + [ -1.3ex ] 30 & c@xmath62 & c@xmath167 & c@xmath168 & c@xmath168 & 68 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath169 & 106 & c@xmath168 & c@xmath168 & c@xmath168 & c@xmath168 & 144 & c@xmath175 & c@xmath167 & c@xmath167 & c@xmath167 + [ -1.3ex ] 31 & c@xmath167 & c@xmath176 & c@xmath167 & c@xmath168 & 69 & c@xmath168 & c@xmath168 & c@xmath170 & c@xmath168 & 107 & c@xmath167 & c@xmath168 & c@xmath168 & c@xmath167 & 145 & c@xmath62 & c@xmath62 & d@xmath181 & c@xmath167 + [ -1.3ex ] 32 & d@xmath176 & c@xmath62 & c@xmath62 & c@xmath167 & 70 & c@xmath170 & c@xmath168 & c@xmath182 & c@xmath182 & 108 & c@xmath167 & c@xmath168 & c@xmath168 & c@xmath168 & 146 & c@xmath170 & c@xmath167 & c@xmath167 & c@xmath167 + [ -1.3ex ] 33 & c@xmath169 & c@xmath167 & c@xmath167 & c@xmath167 & 71 & c@xmath182 & c@xmath170 & c@xmath62 & c@xmath167 & 109 & c@xmath168 & c@xmath167 & c@xmath168 & c@xmath168 & 147 & i@xmath172 & c@xmath168 & c@xmath168 & c@xmath168 + [ -1.3ex ] 34 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 & 72 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 & 110 & c@xmath167 & c@xmath167 & c@xmath168 & c@xmath168 & 148 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 + [ -1.3ex ] 35 & c@xmath62 & d@xmath176 & d@xmath176 & c@xmath167 & 73 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 & 111 & c@xmath167 & c@xmath167 & c@xmath168 & c@xmath168 & 149 & c@xmath167 & c@xmath167 & c@xmath167 & c@xmath168 + [ -1.3ex ] 36 & c@xmath167 & c@xmath168 & c@xmath168 & c@xmath168 & 74 & c@xmath170 & c@xmath167 & c@xmath168 & c@xmath168 & 112 & c@xmath167 & c@xmath168 & c@xmath167 & c@xmath168 & 150 & c@xmath175 & c@xmath175 & c@xmath167 & c@xmath168 + [ -1.3ex ] | the four most stable structures of ni@xmath0 clusters with @xmath1 from 2 to 150 have been determined using a combination of the embedded - atom method in the version of daw , baskes and foiles , the _ variable metric / quasi - newton _ method , and our own _ aufbau / abbau _ method .
a systematic study of energetics , structure , growth , and stability of also larger clusters has been carried through without more or less severe assumptions on the initial geometries in the structure optimization , on the symmetry , or on bond lengths .
it is shown that cluster growth is predominantly icosahedral with @xmath2 of _ fcc _ , _ tetrahedral _ and _ decahedral _ growth .
for the first time in unbiased computations it is found that ni@xmath3 is the multilayer ( third mackay ) icosahedron .
further , we point to an enhanced ability of _ fcc _ clusters to compete with the icosahedral and decahedral structures in the vicinity of @xmath4 .
in addition , it is shown that conversion from the _ hcp_/anti - mackay kind of icosahedral growth to the _
fcc_/mackay one occurs within a transition layer including several cluster sizes .
moreover , we present and apply different analytical tools in studying structural and energetic properties of such a large class of clusters . these include means for identifying the overall shape , the occurrence of atomic shells , the similarity of the clusters with , e.g. , fragments of the _ fcc _ crystal or of a large icosahedral cluster , and a way of analysing whether the @xmath1-atom cluster can be considered constructed from the @xmath5-atom one by adding an extra atom . in addition
, we compare in detail with results from chemical - probe experiment .
maybe the most central result is that first for clusters with @xmath1 above 80 general trends can be identified . |
a fraction of the galaxies selected in the near - infrared show very red colors ( e.g. @xmath2 ) .
the first cases were serendipitously discovered by @xcite , and even more extreme cases were found by @xcite .
such galaxies are known as extremely red objects ( eros ) , and the most recent surveys demonstrated that they form a substantial field population @xcite ( @xmath30.5 eros arcmin@xmath4 for @xmath2 and @xmath5 ) , whereas eros are often found in excess around high-@xmath6 radio - loud agn @xcite . because of the `` age dust '' degeneracy , the red colors are consistent with eros being @xmath0 old passively evolving ellipticals as well as star - forming galaxies and agn strongly reddened by dust extinction .
because of their faintness , the observation of eros is very challenging ( sometimes unfeasible ) even with 10m - class telescopes .
however , recent spectroscopy demonstrated that ellipticals and starbursts are indeed present in the ero population @xcite .
the key question is then to derive the relative fractions of both galaxy types in order to exploit the stringent clues that eros can place on the formation and evolution of elliptical galaxies and on the abundance of dust obscured systems at high-@xmath6 .
the question on the origin of the present - day massive spheroidals is one of the most debated issues of structure formation and galaxy evolution @xcite .
the fundamental question is to understand if ellipticals formed at early cosmological epochs ( e.g. @xmath7 ) through a relatively short episode of intense star formation followed by a passive evolution to nowadays , or if they built up through the merging of pre - existing disk galaxies taking place mostly at @xmath8 , thus making massive spheroidals at @xmath0 rare objects @xcite . a direct way to test such galaxy formation scenarios is then to search for massive field ellipticals at @xmath0 and to compare their number with the above model predictions .
since near - ir light is a good tracer of the galaxy stellar mass @xcite ( e.g. @xmath9 m@xmath10 correspond to @xmath11 for @xmath12 @xcite ) , ero surveys are capable to select massive passively evolving elliptical candidates .
for instance , a color of @xmath13 is expected for @xmath0 , @xmath14 , @xmath15=50 km s@xmath16 mpc@xmath16 , @xmath17 , @xmath18 , salpeter imf ( from bruzual & charlot 1997 models ) .
searches based on this approach gave very discrepant results and claimed either a strong deficit or a constant comoving density of @xmath0 ellipticals @xcite . since such results were based on small field surveys ( typically 1 - 60 arcmin@xmath19 ) , wide field searches for eros have been recently made in order to overcome the possible effects of the field - to - field variations . in one of the widest field survey to date @xcite , 700 arcmin@xmath19 were observed to @xmath5 , covering a sky area about ten times larger than the previous surveys and providing a complete sample of about 400 eros with @xmath2 .
the main results of such a survey were the detection of strong angular clustering of eros ( an order of magnitude larger than that of field galaxies ) , and the solid estimate of their surface density .
such results were recently confirmed by @xcite .
the observed clustering can easily explain the previous discrepant results as due to strong `` cosmic variance '' effects , and it suggests that most eros with @xmath5 are ellipticals rather than dust reddened starbursts ( see @xcite for more details ) . if this is the case , the spatial correlation length @xmath20 of @xmath21 ellipticals is in the range of 6 - 10@xmath22 mpc comoving @xcite . concerning the comparison with galaxy evolution scenarios , even in a conservative case where 70% of eros are ellipticals , their observed surface density is in good agreement with the predictions of passive evolution ( fig .
1 ) , suggesting that most field ellipticals were fully assembled at least by @xmath23 @xcite .
this result does not imply that the formation of massive spheroidals occurred necessarily through an old fashioned `` monolithic collapse '' @xcite , but it simply constrains the epoch when the formation took place , and it implies that , if ellipticals formed through merging , this occurred mostly at @xmath24 .
( corresponding to select passively evolving ellipticals at @xmath0 , see text ) derived from the wide field surveys of @xcite .
the error bars take into account the non - poissonian fluctuations due to the strong angular clustering discovered by @xcite .
the curves show the predicted densities in case of pure luminosity evolution ( ple ) for different formation redshifts adopting the marzke et al .
( 1994 ) local luminosity function of ellipticals and the bruzual & charlot ( 1997 ) spectral synthesis models with salpeter imf and solar metallicity ( see @xcite for more details ) .
the starred symbols show the observed density in case 70% of eros are passively evolving elliptical galaxies .
the good agreement between the observed and the predicted surface densities suggests that ple models can not be discarded and that most ellipticals formed at @xmath24 . , scaledwidth=110.0%,height=529 ] follow - up observations are needed to verify that most eros are ellipticals . despite the observational difficulties due to their faintness , the 10-m class telescopes are spectroscopically confirming that a substantial fraction of eros are @xmath0 passively evolving ellipticals with old ages ( @xmath25 gyr ) consistent with being formed at remote cosmological epochs @xcite .
the preliminary results of the _ k20 _
survey @xcite show that emission lines were detected in only @xmath320% of the eros with @xmath26 and @xmath2 observed so far , whereas the remaining fraction is made by spectroscopically identified ellipticals at @xmath0 and by unidentified objects with no emission lines which could be either ellipticals at higher @xmath6 or dusty star - forming galaxies at @xmath27 ( i.e. with emission lines falling out of the observed optical spectral range ) .
in addition to spectroscopy , recent hst imaging showed that most eros have morphologies and surface brightness profiles characteristic of dynamically relaxed spheroidals , whereas only 15 - 30% have irregular or disk - like morphologies @xcite ( see also @xcite ) .
the possibility that eros were dust reddened starbursts with high far - ir luminosity was first explored by @xcite who detected submm and mm continuum emission from hr10 ( @xmath28 ; @xcite ) and showed that such a galaxy is an ultraluminous infrared galaxy ( ulig ; @xmath29 l@xmath10 ) .
strong co emission was then detected by @xcite implying a hydrogen molecular gas mass of @xmath30 m@xmath10 .
such results demonstrated that a fraction of eros are dust obscured uligs with star formation rates ( sfrs ) @xmath31 m@xmath10yr@xmath16 ( see also @xcite ) .
besides hr10 , other eros have been detected in the submm , and all of them displayed properties consistent with being uligs at even higher redshifts ( @xmath32 ) and with extreme sfrs up to @xmath31000 m@xmath10yr@xmath16 @xcite .
preliminary scuba results suggest that eros with significant submm emission may be segregated among the objects with the most extreme colors @xcite ( e.g. @xmath33 , @xmath34 ) .
although preliminary , this seems in broad agreement with the hst results showing that eros with irregular / merging - like morphologies ( i.e. consistent with being dusty starbursts ) have the tendency to have the reddest colors @xcite . if confirmed , this would allow to disentangle dusty eros from ellipticals on the basis of their colors . if a fraction of massive ellipticals formed at early cosmological epochs through a short - lived starburst phenomenon , such a phase may have occured in a dusty environment where the uv light from ob stars is absorbed by the dust grains and re - emitted in the rest - frame far - infrared , making them weak in the observed - frame optical and bright in the submillimeter ( @xcite and references therein ) .
it is then intriguing to speculate whether the dusty eros which have properties consistent with being uligs at @xmath1 represent proto - ellipticals seen during their formation episode .
a circumstantial evidence in favour of this comes from the similarity between the estimated comoving density of submm - selected galaxies with @xmath35 mjy ( which often have ero counterparts ) and that of present - day ellipticals with @xmath36 @xcite .
the most recent results suggest that @xmath37 ellipticals are probably the dominant population among the eros selected at @xmath3818 - 20 , whereas a minor , although significant , fraction is made by dust enshrouded uligs over a wide range of redshifts ( @xmath39 ) .
multiwavelength observations of complete samples are needed to establish the relative fractions of the galaxies forming the ero population in order to constrain the evolution of massive galaxies and of their clustering , and to investigate the link between eros and obscured agn suggested by recent x - ray observations @xcite .
elston r. , rieke g.h .
, rieke m.j .
1988 , apj , 331 , l77 hu , e. m. ; ridgway , s. e. 1994 , aj , 107 , 1303 thompson d. et al . 1999 ,
apj , 523 , 100 daddi e. et al .
2000 , a&a , 361 , 535 cimatti a. et al .
2000 , mnras , 318 , 453 chapman , s. c. ; mccarthy , p. j. ; persson , s. e. 2000 , aj , 120 , 1612 graham j.r .
, dey a. 1996 , apj , 471 , 720 spinrad h. et al .
1997 , apj , 484 , 581 cimatti a. et al . 1999 ,
a&a , 352 , l45 soifer b.t . 1999 ,
aj , 118 , 2065 liu m.c .
2000 , aj , 119 , 2556 renzini a. , cimatti a. 2000 , in the hy - redshift universe : galaxy formation and evolution at high redshift " , ed .
bunker & w.j.m .
van breugel , a.s.p .
series vol .
193 , in press ( astro - ph/9910162 ) kauffmann g. 1996 , mnras , 281 , 487 baugh c.m . , cole s. , frenk c.s .
1996 , mnras , 283 , 1361 gavazzi g. , pierini d. , boselli a. 1996 , a&a , 312 , 397 kauffmann g. , charlot s. 1998 , mnras , 297 , l23 zepf s.e . 1997 , nature , 390 , 377 franceschini a. et al .
1998 , apj , 506 , 600 barger a.j .
et al . 1999 ,
aj , 117 , 102 totani t. , yoshii j. 1997 , apj , 501 , l177 benitez n. et al . 1999 ,
apj , 515 , l65 broadhurst t.j .
, bouwens r.j .
1999 , apj , 530 , l53 schade d. et al . 1999 ,
apj , 525 , 31 scodeggio m. , silva d.r .
2000 , a&a , 359 , 953 treu t. , stiavelli m. 1999 , apj , 524 , l27 daddi e. et al . , these proceedings mccarthy p.j .
these proceedings daddi e. , cimatti a. , renzini a. 2000 , a&a , 362 , l45 eggen , o. j. ; lynden - bell , d. ; sandage , a. r. 1962 , apj , 136 , 748 cohen j.g .
et al . 1999 ,
apj , 512 , 30 stern d. et al . , these proceedings cimatti et al . , in preparation ( see http://www .
arcetri.astro.it/@xmath3k20/ ) moriondo g. , cimatti a. , daddi e. 2000 , a&a , in press stiavelli m. , treu t. 2000 , astro - ph/0010100 corbin m.r .
2000 , aj , 120 , 1209 cimatti a. , andreani p. , rttgering h. , tilanus r. 1998 , nature , 392 , 895 andreani p. , cimatti a. , loinard l. , rttgering h.j.a . 2000 ,
a&a , 354 , l1 dey a. et al . 1999 ,
apj , 519 , 610 smail i. et al .
1999 , mnras 308 , 1061 gear w.k .
2000 , mnras , in press ( astroph/0007054 ) ivison r.j .
2000 , apj , 542 , 271 dunlop j.s .
2000 , astroph/0011077 cimatti et al .
, in preparation granato g.l .
2000 , astro - ph/0006408 hasinger et al . , these proceedings | extremely red objects ( eros ) provide the important possibility to shed light on the formation and evolution of the present - day massive ellipticals .
on one hand , they allow to select @xmath0 old passively evolving spheroidals and to compare their abundance with the predictions of galaxy formation scenarios . on the other hand
, they provide the possibility to find dust obscured starbursts , a fraction of which may trace the formation of proto - ellipticals at @xmath1 . in this paper , the most recent results on eros are reviewed and the main implications discussed . |
inspired by the success of iterative decoding of low - density parity - check ( ldpc ) codes , originally introduced by gallager @xcite and later rediscovered in the mid 1990 s by mackay and neal @xcite , on a wide variety of communication channels , the idea of iterative , soft - decision decoding has recently been applied to classical algebraically constructed codes in order to achieve low - complexity belief propagation decoding @xcite .
also , the classical idea of using the automorphism group of the code , @xmath0 , to permute the code , @xmath1 , during decoding ( known as _ permutation decoding _ ( pd ) @xcite ) has been successfully modified to enhance the sum - product algorithm ( spa ) in @xcite .
we will denote this algorithm by spa - pd .
furthermore , good results have been achieved by running such algorithms on several structurally distinct representations of @xmath1 @xcite .
both reed - solomon and bose - chaudhuri - hocquenghem ( bch ) codes have been considered in this context .
certain algebraically constructed codes are known to exhibit large minimum distance and a non - trivial @xmath0 .
however , additional properties come into play in modern , graph - based coding theory , for instance , sparsity , girth , and trapping sets @xcite
. structural weaknesses of graphical codes are inherent to the particular parity - check matrix , @xmath2 , used to implement @xmath1 in the decoder .
this matrix is a non - unique @xmath3-dimensional basis for the null space of @xmath1 , which , in turn , is a @xmath4-dimensional subspace of @xmath5 .
although any basis ( for the dual code , @xmath6 ) is a parity - check matrix for @xmath1 , their performance in decoders is not uniform .
@xmath2 is said to be in standard form if the matrix has @xmath7 weight-1 columns .
the weight of @xmath2 is the number of non - zero entries , and the minimum weight is lower - bounded by @xmath8 , where @xmath9 denotes the minimum distance of @xmath6 .
it is well - known that @xmath2 can be mapped into a bipartite ( tanner ) graph , @xmath10 , which has an edge connecting nodes @xmath11 and @xmath12 _ iff _ @xmath13 . here ,
@xmath14 , refers to the bit nodes ( columns of @xmath2 ) , and @xmath15 , refers to the check nodes ( rows of @xmath2 ) .
the local neighborhood of a node , @xmath16 , is the set of nodes adjacent to @xmath16 , and is denoted by @xmath17 .
the terms standard form and weight extend trivially to @xmath10 . in the following , we use bold face notation for vectors , and the transpose of @xmath2
is written @xmath18 .
this paper is a continuation of our previous work on edge local complementation ( elc ) and iterative decoding , in which selective use of elc ( with preprocessing and memory overhead ) equals spa - pd @xcite . in this work ,
we use elc in a truly random , online fashion , thus simplifying both the description and application of the proposed decoder .
the key difference from our previous work is that we do not take measures to preserve graph isomorphism , and explore the benefits of going outside the automorphism group of the code .
this means that we alleviate the preprocessing of suitable elc locations ( edges ) , as well as the memory overhead of storing and sampling from such a set during decoding .
our proposed decoding algorithm can be thought of as a combination of spa - pd @xcite and multiple bases belief propagation @xcite .
we also discuss the modification of the powerful technique of damping to a graph - local perspective .
the operation of elc @xcite , also known as pivot , is a local operation on a simple graph ( undirected with no loops ) , @xmath19 , which has been shown to be useful both for code equivalence and classification @xcite , and for decoding purposes @xcite .
it has recently been identified as a useful local unitary primitive to be applied to _
graph states _
@xcite a proposed paradigm for quantum computation @xcite .
[ elc1 ] shows @xmath20 , the local subgraph of a bipartite graph induced by nodes @xmath21 , and their disjoint neighborhoods which we denote by @xmath22 and @xmath23 , respectively .
elc on a bipartite graph is described as the complementation of edges between these two sets ; @xmath24 and @xmath25 , check whether edge @xmath26 , in which case it is deleted , otherwise it is created .
finally , the edges adjacent to @xmath27 and @xmath16 are swapped
see fig .
elc on @xmath19 extends easily to elc on @xmath10 when @xmath2 is in standard form @xcite .
given a bipartite graph with bipartition @xmath28 , we then have a one - to - one mapping to a tanner graph , with check nodes from the set @xmath29 and bit nodes from @xmath30 .
[ pivot ] shows an example , where the bipartition is fixed according to the sets @xmath31 and @xmath29 . in fig .
[ gb ] , the left and right nodes correspond to @xmath31 and @xmath29 , respectively , for the simple graph @xmath19 .
@xmath10 may be obtained by replacing grey nodes by a check node singly connected to a bit node , as illustrated in fig .
[ ga ] . figs .
[ gb ] and [ ge ] show an example of elc on the edge @xmath32 . although the bipartition changes ( edges adjacent to 0 and 5 are swapped ) , figs .
[ ga ] and [ gd ] show how the map to tanner graphs , in fact , preserves the code . by complementing the edges of a local neighborhood of @xmath10
, elc has the effect of row additions on @xmath2 .
the complexity of elc on @xmath33 is @xmath34 .
the set of vertex - labeled graphs generated by elc on @xmath10 ( or , equivalently , @xmath19 ) is here called the _ elc - orbit _ of @xmath35 .
each information set for @xmath35 corresponds to a unique graph in the elc - orbit @xcite . note that this is a code property , which , as such , is independent of the initial parity - check matrix , @xmath2 .
the set of structurally distinct ( unlabeled ) graphs generated by elc is here called the _ s - orbit _ of @xmath35 , and is a subset of the elc - orbit .
graphs are structurally distinct ( i.e. , non - isomorphic ) if the corresponding parity - check matrices are not row or column permutations of each other .
each structure in the s - orbit has a set of @xmath36 isomorphic graphs , comprising an _ iso - orbit _ @xcite . in the following
, we will refer to elc directly on @xmath10 , keeping fig .
[ pivot ] in mind .
the spa is an inherently local algorithm on @xmath10 , where the global problem of decoding is partitioned into a system of simpler subproblems @xcite .
each node and its adjacent edges can be considered as a small constituent code , and essentially performs maximum - likelihood decoding ( mld ) based on local information .
the key to a successful decoder lies in this partitioning how these constituent codes are interconnected .
the summed information contained in a bit node , @xmath11 , is the _ a posteriori _ probability ( app ) , @xmath37 , at codeword position @xmath38 .
the vector @xmath39 constitutes a tentative decoding of the received channel vector , @xmath40 .
the decoder input is the log - likelihood ratio ( llr ) vector @xmath41 , where @xmath42 is the channel noise standard deviation on an additive white gaussian noise ( awgn ) channel .
subtracting the input from the app leaves the extrinsic information , @xmath43 , which is produced by the decoder .
the message on the edge from node @xmath16 to @xmath27 , in the direction of @xmath27 , @xmath44 , is computed according to the spa rule on node @xmath16 .
the spa computation of all check nodes , followed by all bit nodes , is referred to as one _ flooding _ iteration .
classical codes , for which strong code properties are known , are typically not very suitable for iterative decoding mainly due to the high weight of their parity - check matrices , which gives many short cycles in the corresponding tanner graphs .
a few recent proposals in the literature have attempted to enhance iterative decoding by dynamically modifying @xmath10 during decoding , so as to achieve diversity and avoid fixed points ( local optima ) in the spa convergence process .
efforts to improve decoding may , roughly , be divided into two categories .
the first approach is to employ several structurally distinct matrices , and use these in a parallel , or sequential , fashion @xcite
. these matrices may be either preprocessed , or found dynamically by changing the graph during decoding .
however , this incurs an overhead either in terms of memory ( keeping a list of matrices , as well as state data ) , or complexity ( adapting the matrix , e.g. , by gaussian elimination @xcite ) . the other approach is to choose a code with a non - trivial @xmath0 , such that diversity may be achieved by permuting the code coordinates @xcite . //
input : @xmath45 . //
output : @xmath46 . @xmath47 . @xmath48 and @xmath49 //
. @xmath50 .
do @xmath51 flooding iterations , @xmath52 .
take the hard decision of @xmath53 into @xmath54 , * stop * if @xmath55 .
@xmath56 . draw random permutation @xmath57 @xcite . @xmath58 and @xmath59 .
@xmath60 .
an example is spa - pd , listed in algorithm [ rrd ] , where @xmath0 is represented by a small set of generators , and uniformly sampled using an algorithm due to celler _
these permutations tend to involve all , or most , of the code coordinates , making it a global operation .
note that line 7 in algorithm [ rrd ] is to compensate for the fact that permutations are applied to @xmath61 in line 12 , rather than to the columns of @xmath2 , after which the messages on the edges no longer ` point to ' their intended recipients .
this is yet another global stage .
the extrinsic information is damped by a coefficient @xmath62 , @xmath63 , in line 10 before being used to re - initialize the decoder .
each time @xmath62 is incremented , the decoder re - starts from the channel vector , @xmath40 .
our proposed local algorithm is a two - stage iterative decoder , interleaving the spa with random elc operations .
we call this spa - elc , and say that it realizes a local diversity decoding of the received codeword .
our algorithm is listed in algorithm [ spaelc ] .
both spa - pd and spa - elc perform a maximum of @xmath64 iterations .
spa update rules ensure that extrinsic information remains summed in bit nodes , such that an edge may be removed from @xmath10 without loss of information .
new edges , @xmath65 , should be initialized according to line 13 in algorithm [ spaelc ] . although neutral ( i.e. , llr 0 ) messages will always be consistent with the convergence process , our experiments clearly indicate that this has the effect of ` diluting ' the information , resulting in an increased decoding time and worse error rate performance . //
input : @xmath66 .
// output : @xmath53 . @xmath47 .
do @xmath51 flooding iterations , @xmath52 .
take the hard decision of @xmath53 into @xmath54 , * stop * if @xmath55 .
select random edge @xmath67 .
@xmath69 , + @xmath70 , @xmath71 . @xmath60 .
the simple spa - elc decoder requires no preprocessing or any complex heuristic or rule to decide when or where to apply elc .
as elc generates the s - orbit of @xmath1 , as well as the iso - orbit of each structure , diversity of structure can be achieved even for random codes , for which @xmath36 is likely to be 1 while the size of the s - orbit is generally very large . however , going outside the iso - orbit means that we change the properties of @xmath2 , most importantly in terms of density and number of short cycles . ideally , the spa - elc decoder operates on a set of structurally distinct parity - check matrices , which are all of minimum weight . with the exception of codes with very strong structure , such as the extended hamming code ,
the elc - orbit of a code will contain structures of weight greater than the minimum .
spa - elc should take measures against the negative impact of increased weight . in this paper
, we adapt the technique of damping to our graph - local perspective .
damping with the standard spa , where @xmath10 is fixed , does not work , so we only want to damp the parts of the graph which change .
as opposed to spa - pd , only a subgraph of @xmath10 is affected by elc , so we restrict damping to new edges in line 13 . note that spa - elc simplifies to a version without damping , denoted by spa - elc@xmath72 ,
when @xmath73 , @xmath74 , and @xmath75 .
this is , simply , flooding iterations interspersed with random elc operations , where new edges are initialized with the adjacent app ( line 13 ) .
currently , the spa stopping criterion ( i.e. , the parameters used to flag when decoding should stop ) is still implemented globally . however
, a reasonable local solution would be to remove the syndrome check ( @xmath55 ) from the stopping criterion , and simply stop after @xmath76 spa - elc iterations , where @xmath76 can be empirically determined . however , this has obvious implications for complexity and latency . in some scenarios
a stopping criterion can be dispensed with anyway for instance when using the decoder as some form of distributed process controller , or for a pipelined implementation in which the iterations are rolled out .
we have compared spa - elc against standard spa , and spa - pd . extended quadratic residue ( eqr )
codes were chosen for the comparison , mainly due to the fact that for some of these codes , @xmath0 can be generated by @xmath77 generators @xcite .
in fact , our experiments have shown that eqr codes have tanner graphs well - suited to spa - elc , at least for short blocklengths .
the codes considered have parameters @xmath78 $ ] ( the extended golay code ) , @xmath79 $ ] ( eqr48 ) , and @xmath80 $ ] ( eqr104 ) .
parity - check matrices for the codes were preprocessed by heuristics to minimize the weight and the number of @xmath81-cycles .
the results are listed in table [ codes ] , where columns marked ` w ' and ` c ' show the weight and the number of @xmath81-cycles , respectively .
columns marked ` initial ' show the weight and the number of @xmath81-cycles of the initial tanner graph constructions . `
reduced ' and ` reduced ip ' refer to optimized tanner graphs , where the latter is restricted to tanner graphs in standard form .
entries marked by an asterisk correspond to minimum weight parity - check matrices .
.optimization of codes used in simulations [ cols= " < , > , > , > , > , > , > " , ] [ codes ] in figs
. [ beery_golay]-[qr104 ] , we show the frame error rate ( fer ) performance and the average number of spa messages of spa , spa - pd , and spa - elc for the extended golay code , the eqr48 code , and the eqr104 code , respectively , on the awgn channel versus the signal - to - noise ratio , @xmath82 . the specific parameters used are indicated in the figure legends .
for the extended golay code and the eqr48 code , we set a maximum at @xmath83 iterations , which we increased to @xmath84 to accommodate the larger eqr104 code .
for spa - elc we have also included results without damping . since spa - elc changes the weight of @xmath10 , we can not compare complexity by simply counting iterations .
since the complexity of one elc operation is much smaller than the complexity of a spa iteration , the total number of spa messages may serve as a common measure for the complexity of the decoders .
we have no initial syndrome check , so the number of iterations approaches @xmath85 at high @xmath82 .
in the same way , the complexity approaches the average weight of the matrices encountered during decoding .
each fer point was simulated until at least @xmath86 frame errors were observed . from the figures , we observe that the spa - elc decoder outperforms standard spa decoding , both in terms of fer and decoding complexity .
the extended golay code is a perfect example for demonstrating the benefits of spa - elc .
the s - orbit of this code contains only two structures , where one is of minimum weight ( weight @xmath87 ) and the other only slightly more dense ( weight @xmath88 ) , while the iso - orbit of the code is very large .
thus , we can extend spa - pd with multiple tanner graphs ( two structures ) while keeping the density low .
not surprisingly , spa - elc achieves the fer performance of spa - pd , albeit with some complexity penalty .
note that the simple spa - elc decoder , without damping , approaches closely the complexity of spa - pd at the cost of a slight loss in fer .
for the larger codes , the sizes of the s - orbits are very large , and many structures are less suited for spa - pd
. still , the same tradeoff between fer performance and complexity holds , based on whether or not we use damping . for the eqr48 code , we have observed a rich subset of the s - orbit containing minimum weight structures ( weight @xmath89 ) . the optimum value of @xmath90 ( see line 10 in algorithm [ spaelc ] ) was determined empirically .
we have described a local diversity decoder , based on the spa and the elc operation .
the spa - elc algorithm outperforms the standard spa both in terms of error rate performance and complexity , and compares well against spa - pd , despite the fact that spa - pd uses global operations .
ongoing efforts are devoted to further improvements , and include ; selective application of elc , rather than random ; devise techniques such that diversity may be restricted to sparse structures in the s - orbit ; identify a code construction suited to spa - elc , for which the s - orbit contains several desirable structures even for large blocklengths .
the authors wish to thank alban goupil for providing the mld curve for the eqr48 code .
d. j. c. mackay and r. m. neal , `` good codes based on very sparse matrices , '' in _ proc .
5th i m a conf .
cryptography and coding ( lecture notes in computer science ) _ , vol . 1025 , royal agricultural college , cirencester , uk , dec .
1995 , pp . 100111 .
j. jiang and k. r. narayanan , `` iterative soft - input soft - output decoding of reed - solomon codes by adapting the parity - check matrix , '' _ ieee trans .
inform . theory _
52 , no . 8 , pp . 37463756 , aug .
2006 .
t. hehn , j. b. huber , s. laendner , and o. milenkovic , `` multiple - bases belief - propagation for decoding of short block codes , '' in _ proc .
inform . theory _
, nice , france , jun .
2007 , pp . 311315 .
j. g. knudsen , c. riera , m. g. parker , and e. rosnes , `` adaptive soft - decision decoding using edge local complementation , '' in _ proc .
castle meeting on coding theory and appl .
( 2icmcta ) ( lecture notes in computer science ) _ , vol .
5228 , castillo de la mota , medina del campo , spain , sep .
2008 , pp . 8294 .
m. hein , w. dr , j. eisert , r. raussendorf , m. van den nest , and h .- j .
briegel , `` entanglement in graph states and its applications , '' in _ proc .
school of physics `` enrico fermi '' on `` quantum computers , algorithms and chaos '' _ , varenna , italy , jul .
[ online ] .
available : http://arxiv.org/abs/quant-ph/0602096 | in this paper , we propose to enhance the performance of the sum - product algorithm ( spa ) by interleaving spa iterations with a random local graph update rule .
this rule is known as edge local complementation ( elc ) , and has the effect of modifying the tanner graph while preserving the code .
we have previously shown how the elc operation can be used to implement an iterative permutation group decoder ( spa - pd)one of the most successful iterative soft - decision decoding strategies at small blocklengths . in this work , we exploit the fact that elc can also give structurally distinct parity - check matrices for the same code . our aim is to describe a simple iterative decoder , running spa - pd on distinct structures , based entirely on random usage of the elc operation . this is called spa - elc , and we focus on small blocklength codes with strong algebraic structure . in particular ,
we look at the extended golay code and two extended quadratic residue codes . both error rate performance and average decoding complexity , measured by the average total number of messages required in the decoding ,
significantly outperform those of the standard spa , and compares well with spa - pd .
however , in contrast to spa - pd , which requires a global action on the tanner graph , we obtain a performance improvement via local action alone .
such localized algorithms are of mathematical interest in their own right , but are also suited to parallel / distributed realizations . |
since its launch on 2004 november 20 , _ swift _ @xcite has provided detailed measurements of numerous gamma ray bursts ( grbs ) and their afterglows with unprecedented reaction times . by detecting burst afterglows promptly , and with high sensitivity ,
the properties of the early afterglow and extended prompt emission can be studied in detail for the first time .
this also facilitates studies of the transition between the prompt emission and the afterglow .
the rapid response of the pointed x - ray telescope ( xrt ) instrument @xcite on _ swift _ has led to the discovery that large x - ray flares are common in grbs and occur at times well after the initial prompt emission .
while there are still many unknown factors related to the mechanisms that produce grb emission , the most commonly accepted model is that of a relativistically expanding fireball with associated internal and external shocks @xcite . in this model ,
internal shocks produce the prompt grb emission .
observationally , this emission typically has a timescale of @xmath5 s for long bursts and @xmath00.3 s for short bursts @xcite .
the expanding fireball then shocks the ambient material to produce a broadband afterglow that decays quickly ( typically as @xmath6 ) .
when the doppler boosting angle of this decelerating fireball exceeds the opening angle of the jet into which it is expanding , then a steepening of the light curve ( jet break ) is also predicted @xcite . for a description of the theoretical models of grb emission and associated observational properties , see @xcite . with the advent of recent _
swift_-xrt observations of many large flares at various times after the burst , it is clear that a new constraint on grb models is available .
two pre-_swift _ observations of relatively small flux increases @xcite did not provide a complete picture of the x - ray flaring activity , and where not initially interpreted as flares ; instead they were interpreted as afterglow onsets .
recent observations by xrt indicate that flares are common , that they can have a fluence comparable to the initial prompt emission , and that they have various timescales , spectra , and relative flux increase factors @xcite . by studying the properties of these flares
, we may elucidate the nature of both the x - ray flares and the grb internal engine .
these grb flare studies can take two forms ; individual grbs with flares can be studied in detail or a large sample of flares can be studied to look at global properties of flares .
the former approach has been taken by many authors @xcite .
the latter method of looking at large samples is used in four recent papers @xcite .
grb 050502b is a prime example of a grb with at least one large flare at late times after the cessation of the initial prompt emission detected by bat @xcite .
the x - ray light curve is shown in figure [ fig:050502b_050904_lc ] . a giant flare , with a flux increase by a factor of @xmath0500 , was observed .
the fluence during the giant flare , ( 1.2 @xmath7 0.05 ) @xmath8 erg @xmath9 in the 0.2 - 10 kev band , was greater than or equal to the prompt fluence .
the flare rises to a sharp peak at 743 @xmath7 10 s. in the hard band ( 1 - 10 kev ) , there is significant short time structure within the peak of the giant flare itself
. the spectrum can be fit best by an absorbed cutoff power law ( or band function ) @xcite , rather than a simple absorbed power law , which fits the underlying afterglow nicely . for details , see @xcite .
the spectral index hardens significantly during the flare ( with a cutoff energy of @xmath02.5 kev in the xrt band ) before returning back to a softer and more typical afterglow spectrum following the flare .
before and after the flare , the temporal decay of the underlying afterglow can be fit well with a single power law @xmath10 . at much later times ,
between ( 1.9 @xmath7 0.3 ) @xmath11 s and ( 1.1 @xmath7 0.1 ) @xmath12 s , there are two broad bumps ( or possibly one broad bump with some structure ) .
grb 060526 has a flare that is remarkably similar to the giant flare of grb 050502b , with a flux increase of about @xmath13 relative to the underlying afterglow . in this case , the giant flare clearly consists of two separate , overlapping flares , peaking at about 220 s and 300 s. these giant flares dominate the light curve and resemble the prompt grb , although at lower peak energies and usually longer time - scales . from studies of a few individual grbs with several strong flares ,
temporal evolution of spectral properties has been explored .
spectral evolution in individual bright flares has been seen in grb 050406 @xcite , grb 050502b @xcite , grb 050607 @xcite , grb 060714 @xcite , grb 050822 @xcite , as well as several others . in all cases ,
spectral hardening was observed at the onset of the flare , followed by spectral softening as the flare peaked and decayed .
temporal evolution of spectra from flare to flare has also been seen , as in grb 060714 which shows a decrease in @xmath14 as a function of flare onset time @xcite . @xcite have observed the same trend in their study of several bright x - ray flares .
the exceptional short bursts , grb 050724 and grb 070724 , exhibited significant flaring detected by xrt @xcite .
it is also possible that grb 051227 , which has a significant x - ray flare peaking at @xmath15 s , is a short burst @xcite .
however , there is some ambiguity in its characterization as short or long .
it is certainly clear that the short burst and the long burst mechanisms must both be capable of producing x - ray flares .
grb 050904 , at a redshift of 6.29 , is the most distant grb detected to date .
this burst has a very interesting x - ray light curve ( see figure [ fig:050502b_050904_lc ] ) with many flares superimposed on top of the underlying temporal decay , and on top of one another @xcite . even after a transformation of the light curve into the rest frame of the grb , there is significant flaring at times as late as @xmath05000 s.
while the detailed study of individual flares is important , it is equally important to look at the properties of the flares in a more general sense to look for trends and overall mean properties of the flares and to compare these properties to those of the typical prompt and afterglow emission .
one such flare sample was chosen by looking at all swift light curves , between launch and 2006 january 24 , and eliminating the ones that did not show any hint of deviation from a power law decay with typical breaks .
only the flares with s / n@xmath16 were retained in the sample .
this analysis results in 77 flares in 33 grbs .
although some grbs had @xmath177 flares , the average was 2 flares per grb .
the temporal and spectral properties of this flare sample have been studied by @xcite and @xcite , and some results are briefly described below .
the spectral parameters of the flares were fit in a way that accounted for the contribution from the underlying light curve photons , which can have a significant impact .
this is , of course , made difficult by the fact that the canonical light curve has a variety of breaks @xcite and may be contaminated by additional flares . in general , simple absorbed power law spectra provide a reasonable fit to the flare data from the sample , but more complex models such as band functions typically provide a better fit along with more self consistent parameters ( particularly n@xmath18 , as discussed in @xcite ) .
the overall distribution of flare fluences are shown in figure [ fig : fluence_temporal ] .
this fluence value does not include the contribution from the power law component of the spectral model that was used to approximate the underlying afterglow contribution to the flare spectrum .
the frequently used practice of quoting the total fluence , without subtracting the underlying afterglow , is misleading since the underlying afterglow light curve sometimes contributes a large fraction of the total fluence .
the mean 0.2 - 10.0 kev fluence ( unabsorbed ) of our sample of flares , derived using band function fits , is @xmath19 erg @xmath9 , with a standard deviation of @xmath20 erg @xmath9 .
each flare light curve in the sample of 77 flares was fit with a model consisting of an underlying power law decay with a flare superimposed on it .
the rising portion of the flare and its decay were fit with power laws .
we then define @xmath21 as @xmath22 , where @xmath23 is the peak time ( the time when the rising power law intersects the decaying power law ) , and @xmath24 is the time when the rising power law intersects the underlying afterglow s power law decay .
similarly , the decay time is defined as @xmath25 .
figure [ fig : fluence_temporal ] shows the distributions of rise and decay times for our sample .
the distributions peak near @xmath26 , and there are several flares with @xmath27 . since external shock processes are typically much slower than this @xcite , a faster process such as internal shocks is a more likely explanation for the flares ( see @xcite for alternative view ) .
the temporal properties of the sample is described in more detail by @xcite .
based on a sample drawn from the first 110 _ swift _ grbs , it is clear that significant x - ray flares are produced frequently and at late times .
for some grbs , there have been more than 7 significant flares measured . for some flares , the fluence in the 0.2 - 10.0
kev x - ray band is in excess of the prompt emission fluence in the 15 - 150 kev band .
the rise and decay times of flares are generally very fast relative to the peak times of the flares .
frequently , it is clear that a nominal afterglow light curve decay has begun prior to a flare and that it resumes following the flare .
the spectra during flares typically shows an increase in the hardness ratio that gradually declines back to the pre - flare value as the flare flux decays .
the spectra of flares are frequently fit better with band functions ( normally used to describe prompt emission from grbs ) than with simple power laws ( normally used for afterglows ) .
based on this list of properties , it is likely that the origin of flares is more akin to the origin of prompt grb emission than it is to afterglow emission . within the context of the standard model , this would require late time internal shocks .
the initial flare sample contained 14 grbs with a measured redshift , and the average redshift did not differ from the average redshift for all _ swift _ grbs , including those without flares .
this implies that late flares can not be merely explained as redshifted normal prompt emission .
this result is also supported by individual burst analysis that are corrected for redshift , such as that of @xcite in which a high redshift burst is corrected by ( 1+z ) and still has very late flares .
all of this implies that the large fluence values for flares reported in this paper ( sometimes comparable to the nominal prompt emission ) must be produced at very late times with peak energies in the x - ray band .
grb progenitor models must be capable of producing this late emission within the same energy budget that was previously applied to only the nominal prompt emission .
the fallback of material onto the central black hole after a stellar collapse could last for long time periods @xcite and lead to late internal engine activity , but the reduced luminosity expected from this model at late times eliminates it as an explanation for all flares .
several models for continued activity of the central engine have been proposed ( e.g. @xcite ) .
these models , and others , must be evaluated within the context of the energy budget and the measured spectral parameters . during x - ray flares ,
the excess flux of x - ray photons at late times may provide a previously unknown seed population of photons that can be inverse compton scattered .
if these are observed at high energies , the lorentz factor could be constrained by comparing the x - ray synchrotron peak and the inverse compton peak .
therefore , x - ray flares may provide a new mechanism by which glast and very high energy gamma - ray instruments can study grbs . | _ swift_-xrt observations of the x - ray emission from gamma ray bursts ( grbs ) and during the grb afterglow have led to many new results during the past two years .
one of these exciting results is that @xmath0@xmath1 of grbs contain detectable x - ray flares .
the mean fluence of the x - ray flares is @xmath010@xmath2 less than that of the initial prompt emission , but in some cases the flare is as energetic as the prompt emission itself .
the flares display fast rises and decays , and they sometimes occur at very late times relative to the prompt emission ( sometimes as late as 10@xmath3 s after t@xmath4 ) with very high peak fluxes relative to the underlying afterglow decay that has clearly begun prior to some flares . the temporal and spectral properties of the flares are found to favor models in which flares arise due to the same grb internal engine processes that spawned the prompt grb emission . therefore , both long and short grb internal engine models must be capable of producing high fluences in the x - ray band at very late times .
address = pennsylvania state university , 525 davey lab , university park , pa 16802 address = pennsylvania state university , 525 davey lab , university park , pa 16802 address = pennsylvania state university , 525 davey lab , university park , pa 16802 address = inaf osservatorio astronomico di brera , merate , italy , altaddress = universit degli studi di milano - bicocca , dipartimento di fisica , milano , italy address = inaf osservatorio astronomico di brera , merate , italy address = inaf osservatorio astronomico di brera , merate , italy address = pennsylvania state university , 525 davey lab , university park , pa 16802 address = pennsylvania state university , 525 davey lab , university park , pa 16802 address = pennsylvania state university , 525 davey lab , university park , pa 16802 address = pennsylvania state university , 525 davey lab , university park , pa 16802 address = inaf osservatorio astronomico di brera , merate , italy address = inaf osservatorio astronomico di brera , merate , italy address = inaf osservatorio astronomico di brera , merate , italy address = nasa / goddard space flight center , greenbelt , md 20771 |
strongly correlated systems continue to be a popular subject due to the importance of these systems in physical processes in general .
electron systems form an important subset of these systems .
the nonlinearity that appears in this type of problem complicates greatly the calculation of any physical properties of the system that strongly depend on the correlation between the individual particles . in this paper
we are mainly concerned with a particular aspect of this area , that is , of calculating the ground state equilibrium energy of a large fermi gas in an external potential , at both zero and at nonzero temperature . at zero temperature
the random phase approximation ( rpa ) is a well known method which addresses this problem . here
we present a way to improve on it , using an effective action method , which includes higher order interactions in a consistent way .
this new method is also applicable at finite temperature . at low temperature ,
the new method shows clearly the importance of the so called `` anomalous diagrams '' in cancelling the non - physical @xmath0 term in the specific heat @xcite .
another important aspect of this new approach is that it is applicable to nonhomogeneous fermi systems .
the currently popular method for treating such systems is the density functional method ( dft)@xcite , @xcite . in dft ,
the density of the system plays a central role and the energy is a functional of it .
both the current method and density functional method are extensions of the thomas - fermi method . in dft
the density is found by solving self - consistently a one - particle schrodinger equation .
correlation effects are taken into account by using the homogeneous result locally .
the kinetic energy term is taken to be that of a non - interacting system having the same density as the real system under consideration . in spite of all these approximations ,
density functional methods have proved to give better values for binding energies than hartree - fock in atoms , molecules and solids in general .
however the method works best only for almost homogeneous systems ; surface effects are handled poorly by this method .
so far the attempt to include a gradient of the density in the energy functional makes the result worse .
the success of the method is therefore limited .
here we propose a different method , that in principle can treat nonhomogeneity in a better fashion .
however , the problem becomes harder to manage numerically .
the method proposed here is a path integral method known as the effective action method . in one dimension
it reduces to the well - known wkb method to first approximation .
this work is a first step toward finding a new practical scheme for the evaluation of correlation energies and other physical properties of strongly correlated electron systems using this method . in the homogeneous case ,
the effectiveness of the method was demonstrated in ref.@xcite .
a closely related approach also appears in ref .
@xcite where the gap equation in a superconductor has been derived using effective action ideas . in field theory this method also received some attention recently in treating a scalar field with a quartic interaction @xcite . here
we carry out the calculations of the energy in the non - homogeneous case as far as possible analytically .
it will be seen that the method provides a very compact expression for the energy that goes beyond the rpa method in a natural way .
we also apply our result for the energy to the homogeneous case .
we explicitly calculate the effect of taking exchange into account in the ring diagrams .
our results agree well with results obtained using quantum monte carlo methods . in the following , we will present a way to compute the free energy of the electron system .
zero temperature results follow easily from this calculation .
the system is assumed to be in an external static potential @xmath1 .
the result that we get is very general and could be applied to many different systems .
the paper is laid out as follows : in the next section , section 2 , we introduce the thermodynamic potential @xmath2 for a system subject to external sources . @xmath2 therefore will be a functional of these sources . the functional @xmath2 is the generator of connected green s functions at finite temperature . usually , one point source functions are used . here
, we will instead introduce two - point functions @xcite , @xcite , @xcite,@xcite .
the introduction of two - point external sources enables us to take into account the higher order corrections to the stationary phase approximation of the partition function in a better way . by including merely two diagrams ( expressed in terms of the variables which are conjugate to the external sources ) in the effective action
, we are able to obtain the contribution of all the second order diagrams ( and beyond ) of the coulomb interaction to the energy in a very compact way ( see fig . [ expansion ] ) .
this is the crucial advantage of this version of the effective action .
the diagrams which are included here are two - particle irreducible , in contrast to the one - particle irreducible diagrams that we get when only one - point functions are used in the effective action .
however , before doing this , we introduce a new field to replace the quartic coulomb interaction by the well known stratanovich - hubbard transformation , so that all the interactions become local .
this new field is simply the hartree potential as in the rpa method @xcite .
however the treatment presented here is somewhat different from the one in the aforementioned reference . here
, we explicitly introduce a term that describes correlation of the hartree field at two locations .
the sources can also be taken to be instantaneous so we can get a time - translation invariant solution . in section 3 , we introduce the finite temperature effective action @xmath3 .
we get this action by a legendre transformation from @xmath2 .
therefore , we introduce new variables conjugate to the sources .
this is exactly the same type of transformation as is used in thermodynamics . at the tree level ,
the @xmath3 functional is simply the action of the system and is related to the free energy of the system .
for homogeneous systems , @xmath3 is known as the effective potential . in our case @xmath3
will be a functional of three variables , the hartree potential @xmath4 , the correlation function @xmath5 of the hartree potential , and the green s function @xmath6 of the fermi field .
if the external sources are taken to be instantaneous , the points @xmath7 and @xmath8 , correspond to the same temperature , @xmath9 , and the function @xmath10 simply becomes the density matrix of the system . @xmath11 and @xmath12 are spin indices . even though it is possible to express the free energy in terms of the density matrix , it makes the manipulations more cumbersome . in section 4 , we solve for @xmath13 and @xmath5 in terms of @xmath14 .
we are able to do this simply because in reality @xmath10 is the only independent variable . in this case
we get an expression for the effective action @xmath3 in terms of the density only . therefore getting an expression for @xmath3 in terms of @xmath14
is the essential result of this work from which other important calculations can be made . in section 5
, we give an explicit expression for the free energy in the homogeneous case in a neutral background . here
we go beyond the rpa results .
we show that new diagrams appear at finite temperature .
the famous ` anomalous ' diagram is just one of them ( see fig . [ anomalous ] ) .
an important outcome of this calculation is the disappearance of the @xmath15 behavior in the specific heat that is due to exchange @xcite . in section 6
, we treat the zero temperature case .
we show explicitly that the anomalous diagrams disappear as they should . here again we go beyond rpa to include exchange effects on the ring diagrams .
we also show that some of the ladder terms are included in our approximation . in section 7
, we continue the treatment started in the previous section and calculate the contribution of the ring diagrams if they include exchange .
in the conclusion , we reexamine the non - homogeneous problem .
we show how this method is related to the dft formalism .
in fact it is shown that a statement like the hohenberg - kohn theorem is trivially realized in this formalism .
similarly the question of v - representability can be given an affirmative answer within perturbation theory .
in this section , we obtain an expression for the functional @xmath2 to ` order ' @xmath17 .
this functional is the logarithm of the partition functional @xmath18 of the system in the presence of external sources .
the @xmath18 functional is written as an integral over all possible allowable paths of the different fields with a weighting factor that depends on the value of the action of the given path .
the functional @xmath2 is simply the generating functional of connected green s functions at finite temperature of the fields involved .
however because the coulomb interaction , quartic in the fermion field , is hard to integrate , we introduce an auxiliary field by means of the well known hubbard - stratanovich transformation so that we can avoid the quartic interaction and its nonlocal behavior . in other words ,
the procedure consists of transforming the problem into one in which the electrons interact locally with a hartree - type potential . in this
treatment , exchange energy terms will show up directly because of the anti - commutation of the fermi fields , which is built into the calculation as will be described below . the correlation terms are due to the quantum fluctuations around the hartree potential
. in the following we give an explicit derivation of the terms to order @xmath17 in @xmath2 and no counting will be necessary .
what we mean here is that we are including effects of second order in the _ full _ hartree field of the theory .
the equation of motion satisfied by @xmath19 goes therefore beyond the hartree - fock approximation and involves second order exchange effects . a non - relativistic interacting electron gas in an external potential @xmath1
has the following hamiltonian in the second quantized form : @xmath20 \nonumber\\ & + \frac{1}{2}\int d^{3}xd^{3}y\frac{1}{|\vec{x}-\vec{y}|}\psi_{\alpha } ^{\dagger}(x)\psi_{\nu}^{\dagger}(y)\psi_{\nu}(y)\psi_{\alpha}(x).\end{aligned}\ ] ] here , we use units such that @xmath21 and @xmath22 .
@xmath23 is a two - component temperature - dependent electron field . @xmath11 and @xmath24 are spin indices , i.e. , @xmath25 for spin up and @xmath26 for spin down .
summation is implicit for repeated indices . in the following @xmath27
will always mean a 3-d space vector .
the system is constrained by the condition @xmath28 where n is the electron number operator which is constant .
therefore the density operator is simply @xmath29 because of this constraint , we prefer to work instead with the following hamiltonian @xmath30 where @xmath31 is a lagrangian multiplier .
the electron operator @xmath23 can be expanded in terms of a complete orthonormal set of one - particle functions @xmath32 so we write @xmath33 @xmath34 and @xmath35 are annihilation and creation operators .
the subscript @xmath36 includes momentum and spin .
the ground state and the excited states can be represented as slater determinants formed by the wave functions @xmath37 . in our case , it will be advantageous to take these functions to be self - consistent hartree eigenfunctions . in the homogeneous case they reduce to plane waves . since we are going to use a path integral formulation , we give the lagrangian associated with the hamiltonian h. @xmath38 \nonumber\\ & -\frac{1}{2}\int d^{3}xd^{3}y\frac{1}{|\vec{x}-\vec{y}|}\psi^{\dagger } ( x)\psi^{\dagger}(y)\psi(y)\psi(x ) .\end{aligned}\ ] ] so for a system at thermal equilibrium , we have to replace the time component t in the hamiltonian by @xmath39 .
the partition function @xmath18 is as usual defined as a sum over all possible states @xmath40 . in the absence of external sources , we have @xmath41 & = \sum_{\phi } \left\langle { \phi } \
, \vert\ , e^{-\beta h}\ , \vert\ , { \phi}\right\rangle \nonumber\\ & = n\int d\psi d{\psi}^{\dagger } \exp\left ( { -s[\psi,\psi^{\dagger } ] } \right)\end{aligned}\ ] ] where n is a normalization constant and the integration measure of the fermi fields is only defined for fields that satisfy : @xmath42 @xmath43 $ ] is the action of the system @xmath44 & = \int_{0}^{\beta}d\tau\,l\nonumber\\ & = \int_{0}^{\beta}d\tau\int d^{3}x\;\left \ { \psi^{\dagger}(x)\left [ \frac { -\partial}{\partial\tau}+\frac{1}{2}\nabla^{2}+\mu - v(x)\right ] \psi(x ) \right \ } \nonumber\\ & -\frac{1}{2}\int_{0}^{\beta}d\tau\int d^{3}xd^{3}y\;\frac{1}{|\vec{x}% -\vec{y}|}\psi^{\dagger}(x)\psi^{\dagger}(y)\psi(y)\psi(x).\end{aligned}\ ] ] + the functional integral for the partition function becomes
@xmath41 & = & n\int d\psi d\psi^{\dagger}\exp \left \ { -\int d\tau_{x } d^{3}x \left [ \psi^{\dagger}(x)\left ( \frac{\partial}{\partial\tau}-\frac { 1}{2}\nabla^{2}-\mu+v(\vec{x})\right ) \psi(x ) \right .
\nonumber\\ & & \mbox { } \left
+ \frac{1}{2 } \int_{0}^{\beta}d\tau\;d\tau_{y}\int d^{3}y \frac{\psi^{\dagger}(x)\psi(x)\psi^{\dagger}(y ) \psi(y)}{|\vec{x}-\vec{y}|}\delta(\tau-\tau_{y } ) \right ] \right \}.\end{aligned}\ ] ] in writing the last term we made use of the fact that the fermi fields satisfy a grassmann algebra .
an infinite self - energy term has been dropped from the above expression .
such a term cancels at the end and has no effect on the evaluation of the eigenfunctions or related physical quantities .
now , we set for notational convenience : @xmath45 and @xmath46 before proceeding further , we have to get rid of the quartic term in the lagrangian .
this is done by introducing an auxiliary boson field.we write @xmath41 & = n^{\prime}\int d\psi d\psi^{\dagger}d\varphi\;\exp \left \{-\int d^{4}xd^{4}y\;\psi_{\alpha}^{\dagger}(x)g^{-1}(x , y ) \psi_{\alpha}(y ) \right .\nonumber\\ & \left .
-\frac{1}{2}\int d^{4}xd^{4}y\;\varphi(x)a^{-1}(x , y)\varphi(y)+\int d^{4}x\;\varphi(x)\psi_{\alpha}^{\dagger}(x)\psi(x ) \right \}.\end{aligned}\ ] ] the fourth component is the time component integrated over the proper range .
it is easy to see that by using the following formula , @xmath47 and integrating over @xmath48 we get back the original expression for @xmath49 .
the prefactor @xmath50 is a new normalization constant .
the operator @xmath51 is clearly invertible , @xmath52 the non - local character of the coulomb interaction has been removed by introducing the new bosonic field @xmath53 it can be shown that @xmath13 is the hartree potential by using the new equations of motion derived from the new action of the problem .
we also couple the fields to local and non - local sources @xmath54 and @xmath55 .
now we introduce the functional @xmath2 , a generator of connected green functions .
this is introduced through the normalized partition functional which is now a functional of the external sources .
it is given by : @xmath56 = \exp\left \ { -\beta\;\omega[j , b , q ] \right \}\ ] ] such that @xmath57\right ) = } \nonumber\\ & & \int d\psi d\psi^{\dagger}d\varphi\exp \left \ { -s[\psi,\psi^{\dagger } , \varphi]+\int d^{4}xd^{4}y\psi^{\dagger}(x)q(x , y)\psi(y ) \right .
\nonumber \\ & & \mbox { } + \left .
\frac{1}{2}\int d^{4}xd^{4}y\varphi(x)b(x , y)\varphi(y ) + \int d^{4}x\varphi(x)j(x ) \right \}\end{aligned}\ ] ] where we have set @xmath58 .
@xmath59 $ ] is simply the action of the transformed problem , @xmath60 & = \int d^{4}xd^{4}y\;\psi^{\dagger}% ( x)g^{-1}(x , y)\psi(y)+\frac{1}{2}\varphi(x)a^{-1}(x , y)\varphi(y)\nonumber\\ & -\int d^{4}x\varphi(x)\psi^{\dagger}(x)\psi(x)\end{aligned}\ ] ] now we define three new variables : @xmath61}{\delta j(x)}\vert_{j = b = q=0}=\left\langle \varphi(x)\right\rangle \equiv\varphi_{c}(x)\ ] ] @xmath62}{\delta b(x , y)}|_{j = b = q=0 } = \frac{1 } { 2}\left\langle \varphi(x)\varphi(y ) \right\rangle \equiv\frac{1}{2 } \left [ \varphi_{c}(x)\varphi_{c}(y)+c(x , y ) \right ] \ ] ] @xmath62}{\delta q(x , y)}|_{j = b = q=0}=\left\langle \psi_{\alpha}^{\dagger}(x)\psi_{\beta}(y)\right\rangle \equiv\rho_{\alpha \beta}(x , y)\ ] ] @xmath63 is the expectation value of the field @xmath13 in the ground state .
@xmath5 is a correlation function of the field @xmath53 @xmath10 is the green function of the fermi field .
@xmath64 is therefore the density of the system . here and below
, the `` time '' ordering operator is not written explicitly
. therefore terms like @xmath65 are defined by setting @xmath66 note that @xmath5 measures the departure from quasi - independence due to the correlation between the values of the potential at two different locations .
the expectation value of the fermi field @xmath23 is zero . in the following
, we will obtain an explicit expression for @xmath67 $ ] to ` order ' @xmath17 .
then by solving for @xmath68 and @xmath19 in terms of @xmath69 and @xmath70 we get an expression for the effective action @xmath3 which is a functional of the new variables .
this will be done in the next section.now we turn to determining @xmath2 .
first we expand the exponent in the above integral around @xmath71 and @xmath72 where @xmath73 is the configuration of @xmath48 that extremizes the action s. therefore , we have @xmath74 it is understood from the above that there is an integration over space and time on the l.h.s . of this expression .
we choose from now on not to write integrals explicitly unless there might be some confusion .
now we expand around @xmath73 , so we write @xmath75 then assuming that the main contribution to the integral comes from the saddle point , we get the following expression for the partition functional : @xmath76)\;\frac{\det[\widetilde { g}^{-1}+q]\;\det[a^{-1}+b]^{-\frac{1}{2}}}{\det[g^{-1}]\;\det[a^{-1 } ] ^{-\frac{1}{2}}}\,\sigma^{-1 } \nonumber \\ & & \mbox { } \times\,\exp \left [ \int d\psi d\psi^{\dagger}d\varphi\;\;\xi\ ; \left \{\frac{1}{2}e^{2}(\varphi\psi^{\dagger}\psi)^{2}\,\;+\ldots \right \ } \ , \right ] \end{aligned}\ ] ] where we have defined @xmath77 and @xmath78 to be @xmath79 \\ \sigma & = \int d\psi d\psi^{\dagger}d\varphi\;\exp \left [ -\psi^{\dagger } ( g^{-1}+q)\psi-\frac{1}{2}\varphi(a^{-1}+b)\varphi \right ] \end{aligned}\ ] ] in the above , we have used the fact that @xmath80 for grassmann numbers a and a@xmath81 , and @xmath82 for c - numbers x. the argument of the first term on the right is @xmath83= \frac{1}{2}\varphi_{0}a^{-1}\varphi_{0 } + \frac{1}{2 } \varphi_{0}b\varphi_{0 } + j\varphi_{0}%\ ] ] and @xmath84 now using the fact that @xmath85 we can get an explicit expression for @xmath86 $ ] : @xmath87 & = s(\varphi_{0})-\;tr\ln[1+g\,\varphi_{0}+gq]+\frac { 1}{2}\;tr\ln[1+a\,b]\nonumber\\ & -\frac{1}{2}\!\int\!d\psi d\psi^{\dagger}d\varphi\;(\varphi\psi^{\dagger } \psi)^{2}\;\frac{\xi}{\sigma}+\ldots\end{aligned}\ ] ] after finding @xmath2 , we now calculate the effective action @xmath3 .
we have to solve for the external sources in terms of the physical variables @xmath88 , @xmath89 and @xmath90 .
this we do in the next section where we find the effective action at finite temperature .
however , this treatment applies to zero temperatures too . in this case too , we get an expansion in @xmath91 . again
@xmath91 is not a true expansion parameter in the true sense of the word .
it is only used as a bookkeeping method for the diagrams included in the expansion of @xmath2 .
the effective action is a functional of @xmath63 , @xmath5 , and @xmath19 , which is obtained by a triple legendre transformation from @xmath67 $ ] @xmath93 & = \omega\lbrack j , b , q]-\int d^{4}% x\;\varphi_{c}(x)j(x)-\frac{1}{2}\int d\text { } [ xy]\;\varphi_{c}% ( x)b(x , y)\varphi_{c}(y)\nonumber\\ & -\frac{1}{2}\int d[xy]\;c(x , y)b(x , y)-\int d\text { } [ xy]\rho(x , y)q(x , y)\;.\end{aligned}\ ] ] where we have written @xmath94 @xmath95 $ ] for @xmath96 to simplify the notation .
it is understood that the integration over time is carried for @xmath97 in @xmath98 $ ] for the nonzero temperature case and over all time for the zero temperature case . in this section
we will assume we are in the zero temperature regime .
however this discussion applies equally well to the nonzero temperature case .
it is easily verified that : @xmath99}{\delta\varphi_{c}(x ) } = -j(x)-\int d^{4}y\ ; b(x , y)\varphi_{c}(y)\ ] ] @xmath100}{\delta c(x , y)}=-\frac{1}% { 2}b(x , y),\ ] ] and @xmath100}{\delta\rho(x , y)}=-q(x , y)\text { } .\ ] ] when we turn off the external sources , the above equations give the values of @xmath101 and @xmath88 that minimize the effective action .
@xmath63 and @xmath5 are really dependent variables since they depend on the auxiliary field @xmath13 , so we should be able in principle to express them in terms of @xmath19 which gives us the density .
now we have to express @xmath102 and @xmath103 in terms of @xmath90 , @xmath89 and @xmath104 .
first we note that when @xmath105 we have @xmath106 hence , we can write @xmath107 where @xmath108 is of order @xmath91 .
similarly , we write @xmath109 now we seek an expression for @xmath3 in the form @xmath110 it then follows that @xmath111 now , we find approximate expressions for @xmath112 and @xmath103 in terms of @xmath101 and @xmath88 . using eq.(31 ) and eq.(35 ) , we get the following relation : @xmath113 the expression for @xmath112 is more involved . again using eq.(35 ) and treating @xmath73 as a functional of @xmath114 and @xmath112 , we get : @xmath115 inserting back all these expressions in @xmath3 , we get for @xmath116 @xmath117 the terms of order @xmath17 are more involved to treat but again straightforward .
the steps are similar to those in the case of the effective action with one - point sources only @xcite .
actually , it can be shown that the next terms in @xmath118 are the sum of two - particle irreducible diagrams of the theory @xcite,@xcite . here , we sum only the first two diagrams in this series expansion .
[ expansion ] shows the first four diagrams in this expansion .
the fifth diagram is not part of the asymptotic expansion of @xmath3 .
these first two diagrams enable us to include first and second order exchange effects .
diagram e is not part of @xmath118 since it is two - particle reducible .
however it is one - particle irreducible and it is part of the usual effective action @xcite , @xmath119\;\rho(x , y)\rho ( y , x)c(x , y)\nonumber\\ & -\frac{1}{4}e^{4}\int d[xyuv]\;\rho(x , y)\rho(y , u)\rho(u , v)\rho ( v , x)c(x , u)c(y , v).\end{aligned}\ ] ] hence the reducible graphs do not appear in the expansion .
it is the term @xmath78 that appeared in eq.(23 ) that is responsible for excluding such a graph from the expansion .
@xmath118 and higher order terms represent vacuum graphs where the propagators are @xmath19 and @xmath5 of the fermi field and the boson field respectively . therefore to order @xmath17 , the effective action has the following expression @xmath93 & = -s(\varphi_{c})+\frac{1}{2}tr\;\ln ac^{-1}+a^{-1}c\\ & -tr\;\ln\rho^{-1}g+\widetilde{g}^{-1}(\varphi_{c})\rho+\gamma_{2}% [ c,\rho],\nonumber\end{aligned}\ ] ] where @xmath120 the term @xmath121 @xmath122 cancels the term @xmath123 in the hamiltonian h , and therefore it can be handled without difficulty .
however , we have to deal with the @xmath31 that appears in the term @xmath124 .
this could be removed by a convenient choice of the path of integration in the complex @xmath125-plane .
the case @xmath126 occurs when there are no charges present .
we are dealing with negatively charged electrons bound by a static potential @xmath127 . therefore bound states appear , with negative energies bounded from below for any physically sensible @xmath127 . in fig .
[ path ] we show how by going from the @xmath128 path , which corresponds to nonzero @xmath31 , to the @xmath129 path we pick up contributions from the bound states between the two paths , since energies of bound states are poles of the propagator @xmath130 in the complex @xmath125 plane . note that the eigenvalues @xmath131 of these states do not take into account correlations due the coulomb field .
because of the logarithmic operator , there is a cut along the real positive axis . by going back to real time and then fourier transforming the time dependence , we get @xmath132 where @xmath133 applies only to spatial variables .
the @xmath131 s are the eigenvalues of the following equation @xmath134 the single particle functions @xmath37 of the potential @xmath127 are not supposed to be used alone as the starting point for any numerical calculations since they do not take account of the repulsion between electrons .
any numerical calculations will have to start by solving eq.(56 ) .
however , it is clearly advantageous to separate the effect of the external potential @xmath127 .
this separation also appears in the usual treatments .
the ground state energy e@xmath135 is given in terms of the effective action per unit time .
using the above results , and the fact that the time independent values of @xmath136 and @xmath63 are defined by setting @xmath137 , the expression for @xmath138 is @xmath139\;\varphi_{c}(x)a^{-1}(x , y)\varphi _
{ c}(y)\;+\frac{1}{2}tr\;\ln c\;a^{-1}\nonumber\\ & -\frac{1}{2}\int d[xy]\;a^{-1}(x , y)c(y , x)\;-\sum_{\mu<\omega_{k}<0}% \omega_{k}\int dt\nonumber\\ & -tr\;\ln\rho g^{-1}+\int d^{4}x\left ( \frac{\partial}{\partial\tau_{x}% } -\frac{1}{2}\nabla^{2}+v(x)-e\;\varphi_{c}(x)\right ) \rho(x , y)|_{\tau _
{ x}=\tau_{y}}^{x = y}\nonumber\\ & + \frac{1}{4}e^{4}\int d[xyuv]\;\rho(x , y)\rho(y , u)\rho(u , v)\rho ( v , x)c(x , u)c(y , v)\nonumber\\ & + \frac{1}{2}e^{2}\int d[xy]\;c(x , y)\rho(x , y)\rho(y , x)\end{aligned}\ ] ] where we have dropped terms of order higher than @xmath140 in @xmath118 for simplicity .
this is the full expression for the ground state energy of the system . here
@xmath130 has @xmath31 set to zero . to be able to use this expression for @xmath138 ,
more approximations are required .
in later sections , we use this expression to find the energy of a homogeneous electron gas both at zero and nonzero temperature .
since @xmath90 is a dependent field , then in principle we should be able to solve for @xmath90 and @xmath89 in terms of @xmath88 only .
this we do next but only approximately to keep the calculations manageable .
in this section , we find an expression for @xmath3 solely in terms of the green s function @xmath142 this is done by finding the expressions for @xmath63 and @xmath5 , that minimize @xmath3 , in terms of the density @xmath142 for @xmath63 we get the following expression : @xmath143 similarly , minimizing @xmath3 with respect to @xmath19 , we get @xmath144 \rho(x , z)+e^{2}\int d^{4}y\,\rho ( x , y)c(x , y)\rho(y , z)\ ] ] @xmath145\rho(v , u)\rho(u , x)\rho(y , v)\rho(z , y)c(u , y)c(v , x)\ : .\ ] ] finally , minimizing @xmath3 with respect to c , we get @xmath146 c(y , z)\ ] ] @xmath147c(u , v)c(y , z)\rho(x , u)\rho(u , y)\rho(y , v)\rho(v , x)\ : .\ ] ] to be able to solve for @xmath5 we have to linearize the theory . in the following ,
we keep terms only to ` order ' @xmath140 .
hence , we find that the correlation @xmath5 is given by @xmath148\;a(x , x_{1})d(x_{1},x_{2})a(x_{2},z)\ ] ] @xmath149\;a(x , x_{1}^{\prime } ) d(x_{1}^{\prime},x_{2}^{\prime})a(x_{2}^{\prime},x_{1})d(x_{1},x_{2}% ) a(x_{2},z)\ ] ] @xmath149\;\rho(x_{1},x_{1}% ^{\prime})\rho(x_{1}^{\prime},x_{2})\rho(x_{2},x_{2}^{\prime})\rho ( x_{2}^{\prime},x_{1})a(x_{2},x_{1})a(x_{2}^{\prime},z)a(x , x_{1}^{\prime } ) \ ] ] + where we have set @xmath150 the above equation for @xmath151 can be represented diagrammatically as shown in fig .
the first term is the bare coulomb potential .
the second and the third are the direct and exchange term with one bubble . so far the fermion propagators are the true propagators in this expansion . in the following
we will ignore the corrections with two bubbles and higher . to order @xmath140 . ]
clearly @xmath5 is the full screened coulomb potential if all orders of @xmath152 are included .
similarly , to order @xmath140 , the equation satisfied by @xmath153 is @xmath154 \rho(x , z)=\delta(x , z)\ ] ] @xmath155\;\rho ( x , y)\rho(y , z)a(x , x_{1})d(x_{1},x_{2})a(x_{2},y)\ ] ] @xmath156\;\rho(x_{2},x_{1})\rho(x_{1},y)\rho(x , x_{2}% ) \rho(y , z)a(x_{1},x)a(x_{2},y)\ : .\ ] ] this is obviously the known equation satisfied by the fermi green s functions @xcite .
the term on the r.h.s . of order @xmath152 is an exchange term or fock term due to the statistics of the electrons .
the last two terms of order @xmath140 take into account collisions and are equivalent to the usual born approximation for two - particle green s function in scattering theory , fig . [ ladder ] . the most likely way to solve these integro - differential equations is by iteration . using the above equations , we can write an explicit expression for the energy in terms of the full propagator @xmath153 of the theory . however , such an expression suffers from the problem of over - counting some of the states of the system .
this is mainly due to the fact that @xmath152 is not a good expansion parameter .
hence physical arguments are used to discard some terms at this level of approximation and instead include them at higher orders .
an expression for the electron propagator in terms of the free propagator is given in fig .
[ rho - approx ] .
the corresponding analytical expression can be easily written following usual rules @xcite . here
we keep only the first three terms .
the second term vanishes in the homogeneous case , the case we are interested in this paper .
we expect the expression for the energy , which has not been linearized , to be a good one if we believe that a stationary phase approximation is viable . without including @xmath118 ,
the expression obtained for the energy is correct with @xmath19 the hartree propagator and @xmath63 the hartree potential .
this indicates that a stationary phase approximation is possible and that the higher order correction @xmath118 should be a small perturbation to the hartree solution .
the linearization of the problem must take account of this .
hence , we can justify the validity of the expansion of @xmath3 in @xmath91 , @xmath157 + by the result we get in the end . to second order in @xmath91 , eq.(54 )
becomes @xmath158 @xmath159\ : \rho(x , u)\rho(u , y)\rho(y , v)\rho ( v , x)a(u , v)a(y , z)\ ] ] @xmath160\rho(x , u)\rho(u , v)\rho(v , y)\rho(y , x)a(u , v)a(y , z)\ ] ] where @xmath19 is the free electron propagator .
we have used only a first order approximation to the true propagator to find this equation for @xmath5 . using this equation
, we get an expression for the energy to order @xmath17 with proper symmetry factors for the diagrams involved in the expansion . setting @xmath161
, we obtain for an interval of time t , @xmath162=\frac{e^{2}}{2}\int dx\,dy\;\frac{\rho_{h}(x , x)\rho_{h } ( y , y)}{|\vec{x}-\vec{y}|}\,+\,e\int dx\,v(x)\rho_{h}(x , x)\ ] ] @xmath163 @xmath164 @xmath165 @xmath166 @xmath167\ ] ] @xmath168 @xmath169 @xmath170\ ] ] @xmath171 @xmath172 @xmath173 the trace , @xmath174 , acts on x and z. so now we have obtained an expression for the energy in terms of the green s functions @xmath153 only .
this is a major result of the work . as we mentioned above
, we solve for @xmath19 to order @xmath140 by iteration of eqs.(52 - 53 ) .
the expression for the energy can be solely written in terms of @xmath122 .
however as we stated earlier this is not advantageous and it is better to keep working in terms of @xmath19 .
we see that the classical coulomb term and the interaction with the external potential appear naturally in this approximation . they can be easily separated from the full expression for the energy .
another important point is that a gradient of the density also appears naturally within the above expressions .
therefore the kinetic term is treated more accurately in this method compared to the density functional method . by expanding the first logarithm
, we immediately obtain the hartree - fock exchange term , i.e. , @xmath175 + the usual expression for the exchange energy follows trivially from the previous expression upon using the commutation relation of the fermi field .
the infinity arising from this cancels the self - interaction term that was dropped in eq.(9 ) .
there is a similar exchange term that comes from the second logarithmic term .
this term is canceled by another similar term outside the logarithm . the usual rpa term is contained in the first logarithm .
the new , extra term in the same logarithm provides among other things exchange corrections to the ring diagrams .
the higher order exchange diagrams are also included in this approximation .
we will say more on this when we treat the homogeneous case in the following sections.before we go on to the next section , we should point out that an expression for the energy in terms of the density only is possible in this formalism @xcite .
the expression is complicated however and does not seem to be useful .
we were also able to show that a derivative expansion in terms of the hartree field is more suitable in our case rather than seeking one in terms of the density .
in fact a derivative expansion in terms of the hartree field seems more appropriate since it is better behaved than the density for problems where the density changes sharply .
we hope to address these questions in a later communication .
in this section we use the result found previously to improve on calculations of the exchange energy and ring diagrams made by isihara and coworkers and others and calculate the specific heat at low temperature @xcite .
a short review of these calculations appears in ref.@xcite and offers a good comparison of all these methods up to the year 1975 .
the authors of this paper end with an interesting conjecture which we prove here to be true .
the appearance of temperature dependent logarithmic terms in the internal energy found through various calculations is unacceptable .
these terms will reappear again in the specific heat and will spoil the observed linear behavior at low temperature . in the past these difficulties were not given close attention .
they were either hidden in an effective mass or the expansion was not completely correct as we argue below .
the history of this problem dates back to the 30s when it was first pointed out by bardeen in ref.@xcite . a preliminary treatment of this work appeared in ref.@xcite .
it was argued in ref.@xcite that a correct approximation to the internal energy of a many - particle system up to exchange , i.e. , a hartree - fock approximation , will automatically include summing an infinite set of diagrams generated by the known `` anomalous '' diagram , fig .
[ anomalous ] . at low temperature ,
the ring diagrams treated in ref.@xcite are of order @xmath140 .
the exchange energy is of order @xmath152 .
however at nonzero temperature , we have another contribution to order @xmath152 that comes from summing an infinite set of diagrams that were never treated before
. therefore the new set can not be ignored in any finite temperature treatment .
we show below that taking these diagrams into account is exactly what is needed to solve the anomaly in the exchange energy at finite temperature .
the grand partition function @xmath18 of the system is given in this case by the following expression : @xmath176 \right ) \text { } .\ ] ] the thermodynamical potential is a functional of @xmath177 , @xmath112 and @xmath103 but it is a function of the variables @xmath178 @xmath179 and @xmath180 . to get the free energy @xmath181
$ ] we make the following triple legendre transformation @xmath182-\mu\,n=\omega\lbrack j , b , q]-j\varphi_{c}-\frac { 1}{2}\varphi_{c}b\varphi_{c}-\rho q-\delta b\ ] ] where @xmath88 , @xmath90 and @xmath183 are defined as before , eqs.(18 - 20 ) .
we have seen that @xmath184 ( @xmath185 at @xmath186 ) can be calculated perturbatively by expanding the exponent around the classical hartree potential in a neutral background , i.e. , with no external potential .
the coefficients of this expansion are expressible in terms of feynman diagrams . here in this section
we keep only the first two terms of the expansion , in other words we only include the first diagram in fig .
[ expansion ] .
the equation of motion satisfied by @xmath187 is given by : @xmath188 where @xmath189 is the bare coulomb potential to first approximation .
it must be pointed out again that a full solution to @xmath189 gives the expected shielded potential .
an approximate solution to the full linearized equation for @xmath189 was first given in ref.@xcite .
we must stress however that in what follows we use only the zero order solution to @xmath190 in this case the above equation is the hartree - fock approximation to the equation of motion for the one - particle green s function @xmath191 at finite temperature @xcite . the final expression that we are
after is that for the thermodynamic potential @xmath192 .
hence within the above approximation , the expression for the thermodynamic potential has the following form : @xmath193 -e^{2}\re(x - y ) \right \ } \nonumber\\ & -tr \left \ { \ln \left [ \delta(x - y)-e^{2}\im(x - y ) \right ] + e^{2 } \im(x - y ) \right \}\end{aligned}\ ] ] where the functions @xmath194 and @xmath195 are given by @xmath196 and @xmath197 the term @xmath198 is the free contribution , @xmath199 is the usual first order exchange term , the third term represents the usual ring diagrams and the last term represents a new series of diagrams shown in fig . [ anomalous ] .
the function @xmath200 is the bare coulomb potential and @xmath191 is the free one - particle green s function at finite temperature . it should be stressed from the above treatment that the new diagrams are necessary for the consistency of our approximations .
the first diagram of the new set of diagrams was treated previously by isihara and kojima in ref.@xcite in a calculation of the specific heat . here
we do the same calculation again but taking into account the full sequence of the new diagrams .
the method used by these authors differs from ours , however they are both based on path integrals and both start from a partition function . in this sense
our methods are more exact than those that start from a zero temperature treatment like gell - mann s treatment in ref.@xcite or those based on a phenomenological theory like landau s .
the authors of ref.@xcite claim at the end of their paper that screening is probably not needed to have a reasonable answer for the specific heat . here we show that indeed this claim is true . it is well known that the exchange energy has an odd @xmath0 behavior that ends up contributing a @xmath201 term to the specific heat .
this is obviously not reasonable for low temperatures .
the exchange energy at low temperature is given by @xcite,@xcite : @xmath202 now we give the contribution due to the ring diagrams at finite temperature @xcite .
this contribution is given by @xmath203 -e^{2}\re(x - y ) \right \ } .\ ] ] in momentum representation , this is given by @xmath204\ ] ] where p represents momentum and @xmath205 , @xmath206 .
the function @xmath207 is given by : @xmath208 the function @xmath209 is the fermi - dirac function , @xmath210 } .\ ] ] + therefore the finite temperature contribution of the ring diagrams to leading order in @xmath211 , @xmath212 , and for high density is given by @xmath213 here @xmath214 is the total number of particles , which is a constant .
all the other variables are defined as follows : @xmath215 to obtain this answer , we have used explicitly the fact that @xmath216 is small to simplify the calculation .
it is however important to stress that this approximation does not affect our final result regarding the @xmath0 terms in the specific heat .
obviously our result will not apply even to @xmath217 .
the contribution of the new set of diagrams , fig .
[ anomalous ] , is given by : @xmath218 + e^{2}\im(x - y ) \right \ } \nonumber\\ & = \frac{2v}{\beta}{\sum_{n=-\infty}^{\infty}}\frac{4\pi k_{f}^{3}}{{\ ( 2\pi ) } ^{3}}\int_{0}^{\infty}x^{2}dx\left [ \ : \ln \left ( 1-e^{2}% \im(x,\omega_{n } ) \right ) + e^{2}\im(x,\omega_{n})\right ] .\end{aligned}\ ] ] here @xmath219 is the normalized momentum and @xmath220 for @xmath221 .
the function @xmath222 is given by : @xmath223 with @xmath224 and @xmath212 .
the mass of the electron is taken to be m=1 .
the calculation is carried out for large @xmath225 , i.e. for low temperature . again here we used explicitly that @xmath216 is a small parameter . after some lengthy calculations , we get the following explicit expression for @xmath226 to order @xmath227 : @xmath228 + where @xmath229 is a constant .
there are two things to note .
first there is _ no _ zero temperature contribution due to these diagrams , which is as it is supposed to be .
second , the sum of the new diagrams gives an answer which is of order @xmath230 , that is it has the same order as the first order exchange energy even though the terms in the sum are of order at least @xmath140 .
this is the reason we believe that an expansion in @xmath152 did not pick up this new infinite set of diagrams .
it is also of interest to note that the ring diagrams add up to a term of order @xmath140 . now
a close comparison of the three terms shows that the @xmath0 term in the exchange energy gets cancelled by the one found from the new set of diagrams .
the ring diagrams contribute a temperature dependent term of order @xmath140 and hence they are not needed to solve the @xmath0-term in the exchange .
in other words screening does not play any role in this odd behavior of the exchange energy or for that matter of the specific heat as we show next .
the specific heat formula is : @xmath231 where the free energy @xmath232 is given by , @xmath233 . using standard techniques @xcite
, we calculate the contribution of all of the above terms to the specific heat . to first order in @xmath234
, we have @xcite : @xmath235 where @xmath236 is the free energy of a free electron gas . therefore , if @xmath237 is the specific heat for the ideal fermi gas , we get for the specific heat at low temperatures and high densities the expression , @xmath238 the constants @xmath229 and @xmath239 are approximately equal to 8.0 and 2.0 , respectively .
now we turn again to the zero temperature case and calculate the correction due to the inclusion of diagram @xmath240 , fig .
[ expansion ] , in the expansion of the energy .
in this section we apply the main result of section 4 , the expression for the energy , to the homogeneous case at zero temperature .
we assume that there is a background of positive charge of equal magnitude to the average density of the electron gas .
hence the system is neutral .
the literature on this problem of the calculation of the correlation energy is huge .
the most complete treatment , so far as we are aware , was given by bishop and luhrmann and it was restricted to zero temperature effects @xcite . since the system is homogeneous , the final expression for the energy will be given in momentum space .
the green s function that will be used in the following is the solution to the first iteration of the nonlinear equation satisfied by @xmath19 .
since there is no external potential , the input green s function is that of a free electron gas .
the energy expression is given in imaginary time , hence we use the following expression for the free green s function @xmath241 in this section and the next , the electron green s function definition differs by a factor of i from the one given before . in terms of @xmath242 , the ground state energy is given by the following , omitting the subscript 0 for simplicity , @xmath243 @xmath244 @xmath245 \text { \ } \ ] ] @xmath246 \text { \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ } \ ] ] @xmath247 where @xmath248@xmath249@xmath250@xmath251@xmath252@xmath253@xmath254 the above equation , eq.(82 ) , is simply eq.(59 ) taking into account of the fact that @xmath122 gets canceled by an equal and opposite charge .
because of the transitional invariance of the system , we can write a simple expression for the correlation energy in the momentum representation , @xmath255-e^{2}\mathcal{h}(p)-e^{4}\mathcal{m}(p)+2e^{4}\pi(p)\right \ } \nonumber\\ & -v\int\frac{d^{3}p}{(2\pi)^{3}}\frac{d\omega}{2\pi}\left\ { \;\ln \left [ 1-e^{2}\mathcal{k}(p)-e^{4}\sigma(p ) \right ] + e^{2}\mathcal{k } ( p)+e^{4}\sigma(p)\;\right\ } \;.\label{eq : ec}\end{aligned}\ ] ] since the ring diagrams are part of @xmath256 , this will be the first test to see if our expansion gives the correct leading result .
first we show how this equation follows from eq.(82 ) .
we start by letting @xmath257 or , since the system is homogeneous , we can write instead @xmath258 next , we rewrite @xmath259 and @xmath260 in terms of their corresponding fourier transforms , i.e. , @xmath261d(k)\ ] ] and @xmath262a(q)\text { } .\ ] ] hence , by convolution we have , @xmath263 however since @xmath264 it follows then that , @xmath265 after using the expressions for the propagators , we end up with the following expression for @xmath266 @xmath267 with @xmath268 . integrating first over @xmath269 in the complex plane , we find that @xmath270 where we added a factor of @xmath271 to account for spin .
this term @xmath272 will prove to be all that is needed to reproduce the known rpa result .
all other terms that appear in the energy expression are new additions to the correlation . before showing this explicitly , we give the expressions for the remaining terms @xmath273 , @xmath274 , @xmath275 , @xmath276 , @xmath277 and @xmath278 .
@xmath279@xmath280@xmath281@xmath282@xmath283 and @xmath284 \text { } .\ ] ] the fourier transform @xmath285 is given as usual by @xmath286 the terms in eq .
( [ eq : ec ] ) have the following meaning after we expand the @xmath287-terms .
the first term represents the second order exchange term , fig .
[ exch2 ] .
the term @xmath266 is responsible for generating the ring diagrams , fig .
[ ring ] .
the term @xmath274 is responsible for generating some of the ring diagrams with exchange .
the cross terms give the remaining ring diagrams with exchange , fig .
[ firstlog ] . ) . ]
the @xmath275 term generates ring diagrams with a self - energy insertion in each ring , fig .
[ ringse ] .
hence the terms in the first @xmath288 should contribute the most .
some of the terms that appear upon expansion of the second @xmath287 term are shown in fig .
[ secondlog ] . ) . ]
the term @xmath273 generates the so called ` anomalous ' diagrams , fig .
[ anomalous ] .
the term @xmath276 generates ladder terms like those in figs .
[ sigma1 ] and [ ladder3 ] .
@xmath277 generates terms like those in fig .
[ sigma2 ] . cross terms of the last two terms
are shown in fig .
[ crossterm ] .
diagrams that appear in fig .
[ coul ] are important in a nonhomogeneous medium and must be accounted for since they no longer vanish .
they arise whenever we have a @xmath122 term in the expression for energy , eq.(59 ) . normalizing the momentum with respect to @xmath289 , the fermi momentum , the explicit expression for the correlation energy per particle in rydbergs ( ryd . )
has the following form , @xmath290 \right .
\nonumber\\ & & \left .
-\mathcal{h}(x , y)- \mathcal{m}(x , y)+2\pi(x , y ) \right \ } \nonumber\\ & & -\frac{3}{2\pi\alpha^{2}r_{s}^{2 } } \!\int_{-\infty}^{\infty}\!dx\int _ { 0}^{\infty}\!dy\:y^{2 } \left \ { \ln \left [ 1+\mathcal{k } ( x , y)+\sigma(x , y)\right ] -\mathcal{k}(x , y)-\sigma(x , y ) \right \ } , \label{eq : ecp}\end{aligned}\ ] ] where we have set @xmath291 and @xmath292.for convenience , we set @xmath293 where @xmath294 , @xmath295 , and @xmath296 are the first , the second and the third integrals , respectively .
expressions for few of the functions that appear above are given below .
the functions @xmath297 , @xmath298 and @xmath299 are given by , @xmath300 @xmath301 or more explicitly , we have @xmath302\end{aligned}\ ] ] and @xmath303 where @xmath304 @xmath305 is the step function . @xmath306 , @xmath307 and @xmath11 are given by @xmath308@xmath309 and @xmath310 the gell - mann - brueckner term can be isolated from the full expression for the correlation , eq .
( [ eq : ecp ] ) , by rewriting the second integral , @xmath311 , in the following way : @xmath312 where @xmath313 -\mathcal{h}(x , y ) \right \ } \ : \label{eq : ring}\\ i_{2}^{exch } & = \frac{3}{4\pi}\frac{1}{r_{s}^{2}\alpha^{2}}\int_{-\infty } ^{\infty}dx\int_{0}^{\infty}dy\ : y^{2}\left \ { \ln \left [ 1+\frac { \mathcal{m}(x , y)}{1+\mathcal{h}(x , y ) } \right ] -\mathcal{m}(x , y ) \right\}. \label{eq : exch}%\end{aligned}\ ] ] @xmath314 is what is left of @xmath315 .
the term @xmath316 is indeed the full expression for the rpa term @xmath317 by taking the limit @xmath318 and @xmath319 , it reduces to the gell - mann - brueckner result @xcite : @xmath320 as we show next . going back to the expression for @xmath266 , eq .
( [ eq : ad ] ) , we have @xmath321\ ] ] with @xmath322 defined to be @xmath323 now we set , @xmath324 terms other than @xmath316 in eq .
( [ eq : ecp ] ) provide new correlations to the rpa - term . the first term in eq .
( [ eq : ecp ] ) , as we mentioned above , is the second order exchange term and has been evaluated exactly in ref.@xcite , @xmath325 to compare our expansion to others we calculate the term that is equivalent to the rpa calculation .
we start by evaluating the integral @xmath326 . after integrating the angular variables
, we have @xmath327 \text { } .\ ] ] this integral can be easily performed over p. we get after that @xmath328 \left [ \omega^{2}+k^{2}(k_{f}-\frac{k}{2})(k_{f}+\frac{k}{2 } ) \right ] \right . \nonumber\\ & - \frac{1}{2 } \ln \left [ \omega^{2}+k^{2}(-k_{f}+ \frac { k}{2 } ) ^{2 } \right ] \ , \left [ \omega^{2}-k^{2}(k_{f } + \frac{k}{2})(-k_{f}+\frac{k}{2 } ) \right ] \nonumber\\ & \left .
-k^{2 } \omega \left \lbrack \arctan\left ( \frac{k(k_{f } + \frac { k}{2})}{\omega } \right ) -\arctan \left ( \frac{k(-k_{f } + \frac { k}{2})}{\omega}\right ) \right ] + k_{f}k^{3 } \right\ } .\end{aligned}\ ] ] the second integral @xmath329 in @xmath330 is seen to be simply obtained from @xmath331 by replacing @xmath332 by @xmath333
. therefore we have @xmath334 if we set @xmath335 and @xmath336 , we get @xmath337 \right . \nonumber\\ & \mbox{}\left . -\frac{x}{2y}\left
[ \arctan\left ( \frac{2y(1+\frac { y}{2})}{x}\right ) + \arctan\left ( \frac{2y(1-\frac { y}{2}% ) } { x}\right ) \right ] \right\}\end{aligned}\ ] ] where @xmath338 .
hence the ring diagrams contribution to the energy is given by @xmath339-\mathcal{h}(x , y)\right ] , \ ] ] where @xmath214 is the total number of electrons .
now , we notice that if we make the following substitutions : @xmath340 we recover exactly the same expression for the ring diagrams as obtained by bishop and luhrmann in ref.@xcite using a totally different expansion .
this helps to validate our original expansion in @xmath91 and later in iterating the equations of motion in terms of @xmath152 .
we stress that this expansion is valid for both small and large @xmath216 . from the above analysis
, we see explicitly that the method of effective action amounts to including another infinite set of diagrams besides the usual ring diagrams .
one subset of the diagrams added is the ring diagrams that allow exchange in them , fig .
[ firstlog ] . on physical grounds these
are expected to be the next important ones that must be summed up .
it is also easily seen from the above that all second order diagrams are included in the expansion with the right symmetry factors .
this shows that our original expansion in @xmath91 is indeed meaningful .
one last thing to note about the diagrams in fig .
[ secondlog ] is that they include the `` anomalous '' diagrams @xcite .
we have already dealt with these diagrams in the previous section . from a nonzero temperature calculation
, we were able to show that these diagrams give a zero contribution at zero temperature simply because they violate fermi statistics . from eq.(107 ) , the contribution of these diagrams involves integrating @xmath341 - \mathcal{k}(x , y)$ ] over all @xmath342 in the complex plane . since @xmath297 has a simple pole , eq.(109 ) , the above integrand ends up having poles of order two and higher and hence their zero contribution at zero temperature .
however these diagrams become essential at non - zero temperature where the above constraint of fermi statistic is no longer an issue . in the following section we calculate the contribution of the function @xmath343 to the correlation energy
in this section , we give the contribution of diagrams like those shown in fig .
[ ringexchange ] to the correlation energy at zero temperature .
hence we go beyond rpa in this case .
we will show that our method provides excellent agreement with fully numerical calculations .
we also compare our results to those found in ref.@xcite where a coupled cluster formalism has been used to get the correlation energy .
our final results clearly show that our method is more transparent than other previously used methods . to get to these final results we had to make approximations along the way .
we use two different approximations : the hubbard approximation in ref.@xcite and the bishop - luhrmann approximation , ref.@xcite .
we have found that the former applies well to high values of @xmath216 while the latter applies well to low values of @xmath216 . from eq.(118 ) , this amounts to finding @xmath344 and the second order exchange term that has already been calculated by onsager et al . in ref.@xcite .
now we show how to calculate @xmath344 .
the calculation is straightforward but special care must be exercised when it comes to numerically evaluating the final result .
first we let @xmath345\ ] ] where @xmath346 and where the region of integration @xmath3 is given by : @xmath347 it is obvious from the above that we are faced with a daunting task of having to deal with a 9-d integral inside a logarithmic function which itself has to be integrated over normalized energy @xmath342 and normalized momentum @xmath348 .
most of the complications are related to the angle integrals which can not be separated . to be able to make some progress we have to make an approximation , i.e.
, we either assume that @xmath349 and average over the angle between them , i.e. , the hubbard approximation or we use a more sophisticated approximation like the one proposed by bishop and luhrmann ( bl ) .
hubbard s approximation amounts to the following : @xmath350 on the other hand , the bl - approximation is more complicated and the reader is referred to their appendix in ref.@xcite for a discussion of their approximation .
the two approximations are almost equivalent for large values of momentum @xmath348 , i.e. @xmath351 in units of fermi momentum . for values of @xmath348 less than 2 ,
both approximations are quite different ( see fig .
[ hbl_graph ] ) .
the hubbard approximation seems to apply well for moderately large values of @xmath216 while the bl - approximation applies well for low values of @xmath216 .
in fact the bl - approximation incorporates the fermi statistics of the interacting particles where it matters most , i.e. in high density situations , and this is one reason why we found it to be a good approximation to the exchange coulomb interaction for low @xmath216 .
the hubbard approximation should be expected to be a good approximation to particles near the fermi surface and where statistics is not an issue .
the case of low density should apply well to this case .
given either approximation we end up with an expression for @xmath352 of the form : @xmath353 where the function @xmath354 is our approximation to the exchange term in @xmath343 and is assumed known .
the integrations over @xmath355 and @xmath356 are easily found if we choose to write both vectors in cylindrical coordinates with @xmath357 along the z - axis , i.e. , @xmath358 hence the integration over @xmath355 becomes : @xmath359 where @xmath360 @xmath361 @xmath362 with an equivalent expression for the region of integration over @xmath363 .
the integrations are now easily carried out .
the expression that we get for @xmath343 is naturally expressed in terms of @xmath364 instead of @xmath342 .
so in the following @xmath342 refers to @xmath364 .
we only quote the final expression here : @xmath365 with @xmath366 \right \ } \nonumber\\ & \times\left\ { -1+\frac{x}{2 } \left ( 1-\frac{y^{2}}{4}+\frac{x^{2 } } { 4 } \right ) \left ( \arctan \left ( \frac{2-y}{x } \right ) - \arctan \left ( \frac{2+y}{x } \right ) \right ) + \right .
\nonumber\\ & \left .
\ln \left [ \frac { \left ( x^{2}+(2-y)^{2 } \right ) ^{\frac{1}{2 } } \left ( x^{2}+(2+y)^{2 } \right ) ^{\frac{1}{2}}}{x^{2 } } \right ] \right\ } \right ) \theta(2-y)\nonumber\\ & \mbox{}+\left ( -\frac{y^{2}}{64}(\frac{\pi y^{2}}{2\alpha r_{s } } ) ^{2}\mathcal{h}(x , y)(\mathcal{h}(x , y)-1)\right .
\nonumber\\ & \mbox{}-\frac{yx^{2}}{128}\left\ { -4-y\ln\left [ \frac{\left ( x^{2}+(2-y)^{2 } \right ) ^{\frac{1}{2 } } \left ( x^{2}+(2+y ) \right ) ^{\frac{1}{2}}}{x^{2}}\right ] \right .
\nonumber\\ & \mbox{}+\left .
\frac{4}{x } \left ( 1-\frac{y^{2}}{4}+\frac{x^{2 } } { 4})(\arctan(\frac{2+y}{x})-\arctan(\frac{-2+y}{x } ) \right ) \right\ } \nonumber\\ & \times\left\ { \frac{1}{2}\ln\left [ \frac { \left ( x^{2}+(2-y)^{2 } \right ) ^{\frac{1}{2}}\left ( x^{2}+(2+y)\right ) ^{\frac{1}{2}}}{x^{2 } } \right ] \right .
\nonumber\\ & \mbox{}+\left .
\frac{2}{x}\frac{1-\frac{y^{2}}{4 } } { y}\left ( \arctan(\frac{-2+y}{x})-\arctan(\frac{2+y}{x})\right ) \right\ } \right\ } \right ) \theta(y-2 ) \ : .\end{aligned}\ ] ] given this expression for @xmath343 , we can easily write the full expression for the correlation due to exchange effects in ring diagrams . it is important to note that we allow for more than one exchange to take place in each ring diagram .
hence our original integral @xmath367 , can be now evaluated numerically .
so within the above approximation , we were able to reduce our calculations to a two - dimensional integral : @xmath368 -\mathcal{m}(x , y)\right ) \text { .}%\ ] ] the expression for the function @xmath369 depends on which approximation we choose to use . in this formalism
it is hard to decide on which one since both give very close answers to the second order exchange diagram .
we have found that the hubbard approximation in this instance gives 0.041 @xmath370 while the bl approximation gives 0.049 @xmath371 .
clearly the latter is closer to the true value of 0.0484 @xmath370 , but we believe that this is not enough to decide against using the hubbard approximation .
in fact at low @xmath216 we expect that the high energy part of the exchange energy between particles to dominate and this explains why we found the bl - approximations to apply well in this range . at high @xmath216
this is not true anymore and in this instance we expect the bl - approximation to overestimate this exchange of energy between the particles .
this is in fact what we have found .
this casts doubts that the bl - approximation is always better than the hubbard approximation . as we said above the second order exchange term
is not a good test since it is density independent .
this explains why the hubbard approximation gives better results for higher values of @xmath216 .
below we give some numerical answers to compare with those obtained using quantum monte carlo methods ( qmc ) @xcite .
our results are obtained using both approximations . for the second order exchange diagram
we have used the true value of 0.0484 ryd .
ref.@xcite 2 . ref .
this paper [ table ] clearly our results , fig.[table ] , agree well with previous results . for large values of @xmath372
i.e. , @xmath216 @xmath373 our results do not compare well to the qmc results .
however this is expected since we need to include corrections due to the ladder diagrams which are believed to be important at low densities .
this concludes our calculations . we have shown the effectiveness of an effective action approach to electronic systems . in the next section we shall state the viability of using this formalism in the nonhonogeneous case and show how it is related to density functional theory
in this last section , we would like to go back to the inhomogeneous case and show how our method gives some of the results known in dft @xcite @xcite .
we start by briefly reviewing the basic ideas behind dft .
the hamiltonian h is written as the sum of three terms : t , v and u. t is the kinetic energy term , v is the external potential term and u is the coulomb energy term .
so we have @xmath374@xmath375@xmath376@xmath377 if @xmath378 is the ground state , then the density @xmath379 is given by @xmath380 the first important fact to note is that @xmath381 is a unique functional of @xmath379 , assuming that there is a unique ground state for the system .
now let the functionals @xmath382 $ ] and @xmath383 $ ] be defined as follows : @xmath384=\langle\psi|t+u|\psi\rangle\text { } , \]]@xmath385=\int v(r)n(r)d^{3}r+\;f[n(r)]\text { } .\ ] ] then , if we assume that there is a one - to - one correspondence between @xmath379 and @xmath381 , it can be shown that @xmath386}{\delta n(r)}|_{g.s}=0\text { } , \ ] ] with @xmath387 this result is called the hohenberg - kohn theorem .
next , the coulomb energy is isolated from the functional @xmath388 $ ] by introducing a new functional @xmath389 $ ] : @xmath390=\frac{1}{2}\int \ : d\vec{r}\ : d{\vec{r}}^{\prime } \frac{n(r)\;n(r^{\prime})}{|r - r^{\prime}|}\ : + \
: g[n]\ : , \]]@xmath391=\frac{1}{2}\int \:d\vec{r}\ : \ : \nabla_{r}\nabla_{r^{\prime } } \ : n_{1}(r , r^{\prime } ) |_{r = r^{\prime}}+\frac{1}{2}\int\frac{c_{2}(r , r^{\prime})}{|r - r^{\prime}% @xmath392 here @xmath393 is the one - particle density matrix and @xmath394 is a correlation function defined in terms of the one- and two - particle density matrices .
dft calculations are essentially centered around finding good approximations to the functional @xmath395 $ ] .
this is usually done by postulating that there is a virtual system of free electrons in an external potential with exactly the same density as the interacting system .
the energy is found by finding the eigenfunctions that correspond to this external potential self consistently .
the method that was presented here is in fact similar in many ways to the above ideas expressed in dft .
however there are major differences .
our method clearly incorporates the hohenberg - kohn theorem . from eq.(35 ) , we have the following @xmath396}{\delta \rho ( x , x ) } = - q(x , x ) \ : .
\end{aligned}\ ] ] from eq.(49 ) , the term @xmath397 can be separated from the energy @xmath398 .
hence if we set @xmath399 , we get right away the hohenberg - kohn result .
in fact we now have an explicit expression for @xmath400 $ ] within perturbation theory .
this result also shows the one - to - one correspondence between density and external potential as long as there is a unique solution around @xmath401 for the defining equations , eqs.(33 - 35 ) .
we have solved these equations only approximately .
we have made two important approximations .
the first corresponds to the number of diagrams included in @xmath3 from the outset .
however from the calculations of the homogeneous case , we see that it will probably be the case that including only two diagrams in the inhomogeneous case will give good results . including higher order corrections by taking into account diagrams like the ones in fig .
[ nextterm ] is nevertheless straightforward if more accurate results are desired .
the second approximation we have made is that in solving for @xmath19 .
we had to linearize the equations of motions to be able to solve for @xmath19 iteratively .
also from eq.(50 ) , we get the following equivalent result : @xmath402 clearly our method is a generalization of the thomas - fermi method where @xmath379 in the last equation is the _ true _ density of the system .
a very important difference from dft is that within this method we have a systematic scheme for calculating the functional @xmath395 $ ] , eq.(49 ) . from the above analysis in the homogeneous case , it is obvious that the most important contribution to correlation at zero temperature comes from calculating the following term : @xmath403 where @xmath404 to calculate the determinant of the above operators , we need a basis of wavefunctions . in practical computations ,
it is more advantageous to use a finite basis set .
if we choose as our basis set the wavefunctions @xmath405 with eigenvalues @xmath406 such that @xmath407 then the wavefunctions , @xmath408 , of the full interacting system of @xmath214 electrons can be written as a linear combination of the above wavefunctions : @xmath409 the coefficients @xmath410 are found self consistently by solving eq.(56 ) . to solve for @xmath411
, we solve eq.(53 ) self - consistently .
clearly our method is computationally more intensive than dft .
however here we are on more secure grounds and our approximations are very well controlled to any desired accuracy .
plans to apply this method to simple atoms are underway .
we hope to report on these developments elsewhere .
finally , we would like to point out that this method applies equally well to excited states other than the ground state .
the hartree field of the excited state is found again by using eq.(33 ) .
if @xmath412 is the ground state solution then @xmath413 is the excited state solution iff @xmath414 this implies that @xmath415 in the above @xmath416 is assumed small compared to @xmath412 .
this latter equation enables us to solve for @xmath416 and hence the corresponding @xmath19 .
clearly effective actions methods are in general superior to conventional methods as is well known in field theory .
we expect this method to apply equally well to atomic and molecular systems as we have shown to be the case for homogeneous systems .
b.horovitz and r.thieberger , _ physica _ * 71 * , 99 ( 1974 ) .
i. yokota , _
j. phys . soc
. japan _ * 4 * , 82 ( 1949 ) .
a. b. lidiard , _ _ proc .
phys . soc.__**a64 * * , 814 ( 1951 ) .
e. wohlfarth , _ phil . mag .
_ * 41 * , 534 ( 1950 ) . | correlated systems at both zero and nonzero temperature are treated here from a novel angle using a functional method .
this functional method is an extension of the usual effective potential method . here
, however the effective action is made to depend explicitly on the correlation effects that are inherent in the physics involved .
this will enable us to obtain new expressions for the free energy , the specific heat and the ground state energy .
the new expansion is shown to give the expected results for the homogeneous case at zero temperature .
however at non - zero temperature we are able to get new sets of diagrams that have a vanishing effect at zero temperature .
to lowest order these diagrams if summed properly will solve a @xmath0 anomaly in the specific heat of an electron gas at low temperature .
we are also able to show that this method provides a very clear way to extend the rpa approximation .
we calculate the effect of exchange on the ring diagrams at zero temperature and show how to include some of the ladder diagrams .
our results agree well with known numerical calculations .
we conclude by showing that this method is in fact a variant of the time dependent density functional method and can in principle be applied to study the effects of correlation in the nonhomogeneous case .
wisc - th-99 - 353 * * * a. rebei * + _ department of physics + university of wisconsin - madison + madison , wi 53706 _ * w.n.g . hitchon + _ materials science program + university of wisconsin - madison + madison , wi 53706 _ * * * pacs : * * 71.10.ca , 05.30.-d , 05.30.fk , 31.15.gy , 71.27.+a , 71.15.mb |
topologically non - trivial structures have been intensively studied in many areas of physics @xcite . in particular , in the context of classical field theories , these structures are described as finite - energy solutions to some non - linear models . in this case ,
such models must be endowed by a spontaneous symmetry breaking potential for the matter self - interaction , since topological solutions are formed during symmetry breaking phase transitions . in this context , the most common topological defects are kinks @xcite , vortices @xcite and magnetic monopoles @xcite : the first ones are one - dimensional structures described by a single real scalar field , while vortices and monopoles are two- and three - dimensional configurations arising as static solutions of some abelian- and a non - abelian - higgs models .
in particular , vortices are stable configurations described by radially symmetric profile .
the simplest version of such structures arises as electrically non - charged solution of a planar maxwell - higgs model endowed by a fourth - order higgs potential .
however , under special circumstances , vortices also arise as electrically charged configurations of some chern - simons- and maxwell - chern - simons - higgs ( mcs - higgs ) models @xcite . in all these cases , such structures can be found as numerical solutions to a set of first - order differential equations , named bogomolnyi - prasad - sommerfield ( bps ) ones @xcite . in this case , as finite - energy solutions , they have interesting applications , mainly concerning the superconductivity phenomena @xcite . moreover , during the last years , beyond the standard models cited above , non - usual ones have been intensively studied .
these theories , generically named _
k - field models _ , are endowed by non - usual kinetic terms , which change the dynamics of the overall model in a non - standard way . here , it is important to point out that the motivation regarding such generalization arises in a rather natural way , in the context of string theories .
in fact , k - field theories have been used as effective models mainly in cosmology , as an attempt to explain the actual accelerated inflationary phase of the universe @xcite via the so - called k - essence models @xcite .
furthermore , they have been applied in the study of strong gravitational waves @xcite , dark- @xcite and tachyon - matter @xcite , and others @xcite .
in such a non - standard scenario , topologically non - trivial configurations , named _
topological k - solutions _ , can exist even in the absence of a symmetry breaking potential @xcite , from which one notes that the existence of such structures are quite sensible to the presence of non - standard kinetic terms . on the other hand , as an attempt to study topological k - solutions via the comparison between them and their canonical counterparts , some of us have already considered the existence of such solutions in the context of symmetry breaking k - field models @xcite . moreover , interesting results concerning these models and the corresponding solutions can be found in ref.@xcite .
in general , given non - trivial kinetic terms , k - field models can be highly non - linear , and the corresponding k - solutions can be quite hard to find . in this case , the development of a consistent self - dual theoretical framework is quite useful and desirable , since it helps to find topological k - structures as solutions to some non - standard bps equations . here , as in the usual approach , such equations can be obtained via the minimization of the energy functional related to the non - standard model , the resulting self - dual k - solutions being the minimal energy configurations possible ever . in this sense , in a recent work , some of us have presented a bps theoretical framework consistent with a generalized self - dual maxwell - higgs model endowed by non - usual kinetic terms to both gauge and scalar fields n14d .
now , we introduce an extension of that work . here
, we develop a general first - order approach consistent with a non - standard self - dual mcs - higgs model . the overall model , including its self - dual potential , depends on three different functions , @xmath4 , @xmath1 and @xmath5 , which are functions of the scalar sector of the model . in this context , we look for topologically non - trivial configurations via the usual radially symmetric static _ ansatz _ , and the finite - energy generalized numerical solutions we found behave in the same general way as their standard counterparts , as expected .
this letter is outlined as follows : in the next sec.[general ] , we introduce the generalized model and develop the bps framework which allows us to get to its non - standard bps equations . also , we verify the consistence of the overall construction by using it to develop a generalization of the usual mcs - higgs case , the general model being controlled by two real parameters , @xmath6 and @xmath7
. then , in section [ numerical2 ] , we perform the numerical analysis concerning the non - standard bps equations previously presented by means of the relaxation technique . also , we depict the corresponding self - dual k - solutions for the electromagnetic sector and comment on the main features they engender . finally , in sec.[end ]
, we present our ending comments and perspectives . from now on ,
we use the natural units system and a plus - minus signature for the planar minkowski metric .
the planar lagrangian density describing the generalized maxwell - chern - simons - higgs model is given by@xmath8where @xmath9is the usual electromagnetic field strength tensor , @xmath10 is the covariant derivative of the higgs field and @xmath11 is the @xmath12dimensional levi - civita s tensor ( with @xmath13 ) .
the additional real scalar field @xmath14 provides the stabilization of the self - dual solutions arising in the presence of both maxwell and chern - simons terms . here ,
@xmath15 and @xmath16 are positive - definite dimensionless functions of the scalar fields of the model .
the higgs potential @xmath17 supporting spontaneous symmetry breaking is supposed to have the following structure@xmath18from which one notes that @xmath19 is defined in terms of @xmath20 , @xmath21 and @xmath3 , with @xmath22 .
it is worthwhile to point out that the specific form of the potential in ( [ p ] ) assures the self - duality of the generalized model ( [ m ] ) , i.e. , it is a consequence of the bps construction and can be derived from the energy density ( [ xx ] ) .
we introduce the mass scale @xmath23 of the model , and use it to perform the scale transformations : @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 and @xmath30 , where @xmath31 is the vacuum expectation value of the higgs field .
then , we get that @xmath32 and @xmath33@xmath34 , and the modified model ( m ) is now described by the dimensionless lagrange density @xmath34 , which has the same form as @xmath35 . also , for simplicity , we choose @xmath36 .
it is well - known that chern - simons theories only exhibit electrically charged static solutions .
in fact , the static gauss law related to ( [ m ] ) is ( @xmath37 runs over spatial indices only)@xmath38from which one observes that , even in the presence of non - trivial @xmath15 , @xmath39 and @xmath40 , the absence of electrically uncharged static solutions still holds , since the temporal gauge @xmath41 does not solve ( [ gl ] ) .
now , since the gauge @xmath42 can not be implemented and supposing that @xmath20 * , * @xmath43 and @xmath3 are non - trivial functions , the looking for static solutions for the second - order euler - lagrange equations related to ( [ m ] ) can be a quite hard task , even in the presence of suitable boundary conditions . in this context ,
the adequate implementation of the bps formalism is quite useful , since it helps to find finite energy configurations as solutions of a given set of first - order differential equations .
so , from now on , we focus our attention on the development of a bps framework consistent with ( [ m ] ) and ( [ p ] ) . specifically , we look for radially symmetric solutions using the standard static _ _
ansatz__@xmath44@xmath45where @xmath46 are polar coordinates and @xmath47 , @xmath48 , @xmath49 ... is the winding number ( vorticity ) of the configuration .
the fields @xmath50 , @xmath51 and @xmath52 must obey the usual boundary conditions@xmath53@xmath54where prime denotes the derivative with respect to @xmath55 . now
, we implement the bogomolnyi approach on the energy functional related to ( [ m ] ) .
given the generalized potential ( [ p ] ) , the radially symmetric non - standard energy density can be written as@xmath56where @xmath57 * , * @xmath58 and @xmath59 .
it is clear that , in order to guarantee a positive - definite energy , both @xmath60 and @xmath61 must be positive - definite and finite .
now , by using the static gauss law ( [ gl ] ) together with the constraint @xmath62the energy density ( [ ge ] ) can be rewritten in the form@xmath63 at this point , we see that the resulting total energy is minimized by the non - standard first - order equations@xmath64@xmath65equations ( [ xx1 ] ) and ( [ xx2 ] ) are the bps equations related to the generalized mcs - higgs model ( [ m ] ) . in the present case , the bps total energy @xmath66 related to the solutions of ( [ xx1 ] ) and ( [ xx2 ] ) can be explicitly evaluated@xmath67where the minimal energy density @xmath68 is defined as@xmath69also , to compute the bps energy ( [ xx3 ] ) , we have used the boundary conditions ( [ q1 ] ) and ( [ q2 ] ) together with@xmath70finally , we assume that @xmath20 and @xmath1 behave as@xmath71@xmath72where @xmath73 , @xmath74 and @xmath75 are real non - negative constants , and @xmath76 is a real positive constant . from ( [ xx3 ] ) ,
one notes that the total energy @xmath66 is quantized according the winding number @xmath77 .
furthermore , it can be related to the magnetic flux @xmath78 in the standard way .
also , we point out that the first - order framework developed here is implemented for any function @xmath15 finite and positive - definite , i.e. , given some function @xmath0 , the set formed by ( [ xx1 ] ) and ( [ xx2 ] ) gives the bps configurations of the generalized mcs - higgs model ( [ m ] ) . on the other hand , in the absence of ( [ p ] ) or ( [ dc ] ) , the energy functional ( [ ge ] ) can not be written as ( [ xx ] ) , and the development of a consistent non - standard first - order formalism can not be performed .
the prescription for the development of generalized self - dual maxwell - chern - simons - higgs models is as follows : given any function @xmath21 , positive and finite , the corresponding @xmath40 is obtained by means of eq .
( [ dc ] ) .
then , given any function @xmath79 , also positive and finite , the resulting self - dual potential @xmath80 is determined via ( [ p ] ) . in this scenario , the bps states of the generalized model ( [ m ] )
are the solutions of the equations ( [ gl ] ) , ( [ xx1 ] ) and ( [ xx2 ] ) .
the total energy is given by ( [ xx3 ] ) , and the energy density by ( xx4 ) .
in the standard mcs - higgs case , bps solutions with non - trivial topology only exist in the asymmetric vacuum of the potential@xmath81which is defined by @xmath82 and @xmath83 .
so , for simplicity , we assume that also the generalized potential ( [ p ] ) achieves its asymmetric vacuum when @xmath82 and @xmath84 . as a consequence , topologically
non - trivial self - dual solutions of ( [ m ] ) exist in the same asymmetric phase as their standard counterparts .
the standard mcs - higgs model is trivially recovered starting from our generalized framework by setting @xmath85 .
so , instead of recovering the usual model , we introduce a generalization of such theory .
the generalized model is defined by@xmath86where @xmath7 is a positive odd- number . the corresponding @xmath87 is@xmath88for simplicity , we choose @xmath20 as follows@xmath89where @xmath6 is a real non - negative number .
note that @xmath90 and @xmath91 leads us back to the standard case .
the resulting generalized self - dual potential @xmath19 ( [ p ] ) is@xmath92 in the present case , ( [ gl ] ) , ( [ xx1 ] ) and ( [ xx2 ] ) can be written , respectively , as@xmath93@xmath94@xmath95which must be solved according the finite energy boundary conditions and ( [ q2 ] ) . in the next section ,
we solve , and via and ( [ q2 ] ) for different values of @xmath6 and @xmath7 .
then , we use the numerical solutions we found for @xmath50 , @xmath96 and @xmath52 to depict the corresponding profiles for the electric field@xmath97the magnetic field@xmath98and the minimal energy density ( [ xx4 ] ) .
also , we comment on the main features the non - standard numerical solutions engender .
in this section , we focus our attention on the non - standard numerical solutions themselves .
the equations to be studied are , and , and the fields @xmath50 , @xmath51 and @xmath99 must behave according the finite energy boundary conditions and ( [ q2 ] ) . here , for simplicity , we choose @xmath100 .
then , we solve the first - order system numerically by means of the relaxation technique , for different values of the real parameters @xmath6 and @xmath7 .
the profiles of the physical fields and energy density are depicted in the figures below .
the system was solved for @xmath90 and @xmath101 ( usual case , dash - dotted black line ) , @xmath90 and @xmath102 ( dotted blue line ) , @xmath103 and @xmath101 ( dashed red line ) and @xmath103 and @xmath102 ( solid green line ) . for @xmath104 and @xmath101 ( dash - dotted black line ) , @xmath104 and @xmath102 ( dotted blue line ) , @xmath105 and @xmath101 ( dashed red line ) and @xmath106 and @xmath107 ( solid green line).,width=321 ] in figure 1 , we depict the numerical solutions for the electric field , and we see that the non - standard ones behave , in general , as their usual counterpart : starting from @xmath108 ( zero ) , the solutions reach their maximum values at some finite distance @xmath109 from the origin , and vanish as @xmath55 goes to infinity .
however , we point out that different solutions exhibit not only different amplitudes , but also slightly different characteristic lengths : in general , increments on @xmath6 and/or @xmath7 lead to decrements on amplitudes and/or characteristic length . even in this case
, such variations are expected to occur in the context of a non - standard theory as ( [ m ] ) ; see , for instance , @xcite . .
conventions as in fig .
1.,width=321 ] in figure 2 , we present the profile of the magnetic field ( [ mg ] ) .
the same manner as for the electric field , the generalized solutions behave as the standard one : the magnetic fields are lumps centered at the origin , and they decrease monotonically as @xmath55 goes to infinity . in this case , for a fixed @xmath6 and an increasing @xmath7 , the amplitudes experience a small increase and the characteristic length decreases . on the other hand , for a fixed @xmath7 and an increasing @xmath6 , the amplitude will decrease significantly , but
the characteristic lengths remain approximately fixed by a compensatory effect relating @xmath6 and @xmath7 . .
conventions as in fig . 1 . ,
width=321 ] numerical solutions for the minimal energy density @xmath68 ( xx4 ) are depicted in figure 3 , and we see that also such solutions are lumps centered at @xmath110 which decrease monotonically to @xmath111 .
the lumps profile are similar to those the magnetic field . however , in this case , for a fixed @xmath6 and an increasing @xmath7 , the amplitudes experience a large increase and , the characteristic lengths decreased slightly . on the other hand , for a fixed @xmath7 and an increasing @xmath6 , the amplitude will decrease significantly but the characteristic lengths remain almost fixed by a compensatory effect relating @xmath6 and @xmath7 . for @xmath112 .
conventions as in fig .
1.,width=321 ] to end this section , we have depicted the numerical profiles for @xmath113 and @xmath68 for @xmath112 ( see figures 4 and 5 , respectively ) . it is well known that in the standard model , for a large winding number , the profiles of both the magnetic field and the energy density go from a lump ( @xmath100 ) to a ring ( @xmath114*)*. from figures 4 and 5 , one notes that also in this limit ( increasing vorticity ) , the generalized solutions mimic the usual ones , as expected . for large @xmath77 with fixed @xmath6 and a increasing @xmath7 ,
the profiles experience a slight change .
however for @xmath7 fixed and an increment in @xmath6 , the profile s amplitudes experience a considerable decrease .
in the present letter , we have performed the development of a first - order theoretical framework in the context of a generalized mcs - higgs model given by ( [ m ] ) . in the present case , the non - standard model , including its self - dual potential ( [ p ] ) , is given in terms of three different functions , @xmath20 , @xmath43 and @xmath3 , which are functions of the scalar fields only . here , in order to avoid problems with the energy of the model , @xmath79 and @xmath115 must be positive - definite .
then , given the general structure for the self - dual potential ( [ p ] ) , the consistence of the non - standard first - order approach only holds when @xmath115 and @xmath40 are related by the differential constraint ( [ dc ] ) . on the other hand ,
one notes that there is no additional constraint to be imposed on @xmath0 .
after perform the construction of the generalized mcs - higss model , we illustrate such realization providing an explicit non - standard model specified by ( [ j1 ] ) , ( [ j2 ] ) and ( [ j3 ] ) . in this case , the non - standard model is controlled by two real parameters , @xmath6 and @xmath7 .
immediately , we have investigated radially symmetric self - dual configurations given by the usual static _ ansatz _ ( [ z1 ] ) and ( a2 ) , and considering that the fields behave according the usual finite energy boundary conditions ( [ q1 ] ) and ( [ q2 ] ) .
we have integrated the generalized bps equations by means of the relaxation technique , for different values of @xmath6 and @xmath7 .
the numerical profiles we found are depicted in figures 1 , 2 , 3 , 4 and 5 . in general
, we have seen that the non - standard solutions mimic their usual counterparts . also , as expected , we have noted variations on the amplitudes and on the characteristic lengths of the modified solutions .
furthermore , we have identified the rules controlling such variations .
this work is an extension of a recent investigation performed by some of us @xcite . in this sense ,
a very interesting issue to be considered concerns the development of a non - standard self - dual mcs - higgs model which exhibits the very same numerical solutions engendered by the standard theory . in this case
, such investigation would be a generalization of the one presented in @xcite , regarding a non - standard maxwell - higgs model .
another issue concerns the supersymmetric extension of the self - dual model ( [ m])-([p ] ) .
these issues are in advance , and we hope report interesting results in a near future .
the authors would like to thank capes , cnpq and fapema ( brazil ) and fct project cern / fp/116358/2010 ( portugal ) for partial financial support .
also , we are grateful to c. dos santos for useful discussions .
e. da hora thanks the department of mathematical sciences of durham university ( u.k . ) , for all their hospitality while doing part of this work .
99 a. vilenkin and e. p. s. shellard , _ cosmic strings and other topological defects _
( cambridge university press , cambridge , england , 1994 ) .
n. manton and p. sutcliffe , _ topological solitons _
( cambridge university press , cambridge , england , 2004 ) .
d. bazeia , e. da hora , c. dos santos and r. menezes , phys .
d * 81 * , 125014 ( 2010 ) .
d. bazeia , e. da hora , r. menezes , h. p. de oliveira and c. dos santos , phys .
d * 81 * , 125016 ( 2010 ) . c. dos santos and e. da hora , eur .
j. c * 70 * , 1145 ( 2010 ) ; eur .
phys . j. c * 71 * , 1519 ( 2011 ) . c. dos santos , phys .
d * 82 * , 125009 ( 2010 ) . c. adam , n. grandi , j. sanchez - guillen and a. wereszczynski , j. phys . a * 41 * , 212004 ( 2008 ) .
erratum - ibid .
a * 42 * , 159801 ( 2009 ) . c. adam , j. sanchez - guillen and a. wereszczynski , j. phys .
a * 40 * , 13625 ( 2007 ) .
erratum - ibid .
a * 42 * , 089801 ( 2009 ) . c. adam , n. grandi , p. klimas , j. sanchez - guillen and a. wereszczynski , j. phys .
a * 41 * , 375401 ( 2008 ) . c. adam , p. klimas , j. sanchez - guillen and a. wereszczynski , j. phys .
a * 42 * , 135401 ( 2009 ) . | we present a consistent bps framework for a generalized maxwell - chern - simons - higgs model .
the overall model , including its self - dual potential , depends on three different functions , @xmath0 , @xmath1 and @xmath2 , which are functions of the scalar fields only .
the bps energy is proportional to the magnetic flux when @xmath1 and @xmath3 are related to each other by a differential constraint .
we present an explicit non - standard model and its topologically non - trivial static configurations , which are described by the usual radially symmetric profile .
finally , we note that the non - standard results behave in a similar way as their standard counterparts , as expected , reinforcing the consistence of the overall construction . |
the _ spitzer _ survey of the small magellanic cloud ( s@xmath0mc ) carried out deep imaging of the main body of the smc in all seven irac and mips bands @xcite .
this survey identified about 400,000 compact sources in the infrared .
compilation of spectral energy distributions and cross - correlation against optical catalogs uncovered 193 point sources with 24- emission , but with optical and near - infrared colors and magnitudes all consistent with o9b3 type main - sequence stars .
photospheres of such stars are well below the detection limit of this survey , so an additional source of infrared ( ir ) emission must be present , presumably warm dust with a temperature close to 150 k. figure [ fig : only24 ] shows the @xmath2 color magnitude diagram for all objects detected at 24 in the s@xmath0mc survey , and the box represents the @xmath3 and @xmath4 cuts on the dusty ob star sample .
these sources are a puzzle , representing a few percent of all the stars in that particular color and magnitude range in the smc .
dust in close proximity to an early b / late o star could be a remnant from the accretion process , and provide important information about the final stages of accretion and the clearing up of the original disk in massive stars .
we discuss below several competing possibilities for the nature of these sources , previously presented by @xcite .
the discovery of these objects has been confirmed independently by the _
spitzer _ sage - smc survey ( * ? ? ? * who also discuss them ) and by @xcite using akari data .
if the dust is associated with the star , it must exist either in an optically thin shell surrounding the star or in a thin or transition circumstellar disk , since no appreciable reddening of the central star is detected beyond the line - of - sight extinction by our own galaxy to the smc .
these objects have @xmath5 to 10@xmath6 , indicating that the dust is intercepting only a small fraction of the star s light and re - radiating it in the ir .
this also supports the thin shell or disk scenario .
since most of the dusty ob stars show little to no excess emission at wavelengths of 8 or shorter , a large central gap would be needed to explain the absence of significant emission from hotter dust located close to the star . a clump of interstellar dust , heated by a nearby star
, however , could also reproduce these characteristics .
this is observed , for example , in the pleiades @xcite .
vega - like stars , herbig ae / be stars , and classical be stars all have disks that are known to contain dust to varying degrees , and so we consider them , as well as cirrus hot spots in the ism , as the possible causes for the excess emission seen around these o and b stars in the s@xmath0mc survey .
vega - like stars host debris disks , which are the remnants of planet formation .
they are optically thin dust disks , mostly stripped of their primordial gas and continually replenished by collisions between planetesimals .
the prototypical examples are vega , fomalhaut , and @xmath7 pictoris , all of which have disks that extend 100 to 1000 au from the star with l@xmath8/l@xmath9 @xmath10 to 10@xmath11 ( * ? ? ?
* and references therein ) .
the central region , a few au in extent , of these disks has been cleared of dust , just as in our solar system , and the mass of emitting dust is @xmath12 10@xmath11 to 10@xmath6 m@xmath13 at temperatures between 50 and 125 k. the wider sample of debris disks shows an upper limit on f@xmath8/f@xmath9 of 10@xmath6 @xcite , temperatures up to 300 k , and dust masses as large as an earth mass @xcite .
note however that the known debris - disk hosts are observed , overwhelmingly , around much less massive stars with later spectral types .
a gas - free disk around a very luminous star would be quickly cleared by radiation pressure , unless the dust grains are very large .
classical be stars are rapidly rotating main - sequence stars with an excretion disk of gaseous material @xcite .
the disk produces ir excess in the form of free free emission with approximately @xmath14 in the near- and mid - ir @xcite ; this by itself would not explain the sudden rise observed in the spectral energy distribution ( sed ) of the dusty o and b stars at 24 .
@xcite , however , found evidence for circumstellar , thermal dust emission in some be stars in the far - ir , which they attribute to remnant dust from prior evolutionary processes .
herbig ae / be stars , on the other hand , are pre - main sequence stars that are thought to be the more massive analogs to t tauri stars .
they are surrounded by gaseous accretion disks containing dust with temperatures as high as 1500 k , creating the ir excess ( see , e.g. * ? ? ?
* ; * ? ? ?
herbig ae / be stars typically show significant excess at @xmath15 from this hot dust .
the onset of the emission at longer wavelengths indicates that the central portion of the disk nearest the star has been cleared and is free of dust .
disks with such a central hole are transition disks ( e.g. , * ? ? ?
it is possible that more massive , early - b and o stars also go through this stage , although modeling suggests that this may be an exceedingly short phase as disks are rapidly cleared @xcite .
observational evidence for circumstellar disks around massive b and o stars remains hard to come by @xcite .
the galactic cirrus , comprised of wispy or filamentary patches of dust throughout the disk of the galaxy , has a typical temperature of @xmath16 k @xcite , but heating of the cirrus by embedded stars produces small regions of warmer dust , or hot spots , near the star , reaching temperatures of @xmath17 k @xcite .
it is difficult to distinguish debris disks from hot spots in the cirrus without resolved images of either the dust emission in the ir , or the scattered light in the optical .
the observed seds can be interpreted with either scenario , as illustrated by @xcite and @xcite ; martnez - galarza et al . model some of the same objects in su et al .
as ism heated by stars passing through a cloud , while su et al.model them as debris disks .
cirrus hot spots are also sometimes the cause of the far - ir emission detected around be stars @xcite . understanding the source of the excess emission in these objects is valuable because the scenarios above represent different phases in the evolution of massive stars
. detecting circumstellar material around smc stars is interesting , because it offers the opportunity to study accretion or excretion processes in a low metallicity environment , and particularly around massive stars with an unextincted line - of - sight .
detecting debris disks in the smc would be especially exciting , as they would offer an unprecedented opportunity to study planet formation in another galaxy .
such early - type stars are under - represented in galactic studies . while the photometry suggested the stellar counterparts of the ir emission to be early - type main - sequence stars , this has yet to be confirmed through spectroscopic observation .
the vast majority of the 193 objects in our sample do not have spectral types reported in the literature , although a small subsample has fiber optics spectroscopy @xcite .
@xcite compiled a catalog of ir counterparts to massive stars spectroscopically confirmed in the literature .
they identify 44 objects that are similar ( ob stars with 24 @xmath1 m emission ) but not necessarily identical to our sample .
they constitute @xmath18% of their sample of late - o and early - b stellar types , a fraction slightly lower but otherwise similar to that reported by @xcite .
the 18 spectra @xcite have in hand show nebular line emission , unlike as we shall see the majority of the sources in this sample .
note that the spectra used by @xcite are obtained using fibers @xcite , which makes sky subtraction in complex regions problematic .
long - slit spectroscopy is preferred in regions of diffuse sky emission such as found around massive stars .
clearly , further spectroscopic study is warranted .
@xcite conclude that most of their sources are not found within very young regions , and suggest their ir emission is likely dust associated with cirrus or a nearby molecular cloud , rather than disks .
@xcite , however , noted that the majority of the sources identified in their sample are in the vicinity of active star - forming regions although not in their cores , where confusion with diffuse 24 @xmath1 m emission would make their identification extremely difficult .
we obtained optical long - slit spectroscopy for 125 of the 193 objects from the s@xmath0mc survey in order to determine how many of them are emission - line ( e.g. , classical be and herbig ae / be ) stars and to obtain spectral types to quantify the stellar radiation field illuminating the dust . in section 2 of this paper , we discuss the photometric and spectroscopic observations , while in section 3 we classify the stars , and in section 4 we estimate the properties of the dust around the non - emission line objects . in section 5 ,
we discuss how well the various scenarios explain our objects and include a comparison of the dust properties to those estimated from the sample of galactic cirrus hot spots of @xcite . in a companion paper (
* hereafter paper ii ) , we discuss further infrared spectroscopic and photometric information , and comparisons to large galactic samples based on _ wide - field infrared survey explorer _
( wise ) data .
the optical spectra were obtained at the 3.5-m new technology telescope at the european southern observatory at la silla , chile in 2007 september ( programme 079.c-0485 ; pi j.th.van loon ) , using the emmi instrument @xcite in rild mode in the red arm with grism # 2 and a @xmath19 slit .
this set - up provided wavelength coverage from @xmath20 to 870 nm with @xmath21 .
biases , dome flats , and hear arc lamp frames were acquired for calibration .
the science integrations were split into three exposures , offset along the slit , in order to correct for cosmic rays and fringing effects at wavelengths beyond 750 nm ) . a filter to block second - order contamination longward of 800 nm
was originally planned , but it was dropped because of unwanted reduction of signal in the blue region of the spectrum .
the spectra were wavelength - calibrated with hear lamp exposures , using the iraf identify , reidentify , and transform tasks in the noao spectroscopic reduction package .
the 1d spectra were then extracted using the apsum task .
the extraction was at times complicated by the background emission lines from regions within which our targets were located . to remove this background ,
the aperture and background regions were set interactively , looking first at the profile of the star along the slit , away from any emission lines .
if any [ ] 5007 emission remained in the extracted spectrum , the extraction was re - done looking instead at the profile along the slit at the [ ] line to set the aperture and background regions .
the extracted spectra were summed for each object to increase the number of counts and to average out fringing effects and an electronic interference pattern that could not be removed with the bias or dark frames due to frame - to - frame variations .
lastly , the spectra were normalized to the continuum in the region between 400 and 680 nm by fitting and dividing by an order 20 polynomial .
the photometry used to create the seds for each source is a combination of _ b _ , _ v _ , and _
i _ data from the ogle ii survey @xcite and the magellanic clouds photometric survey @xcite , _ j _ , _ h _ , and _
k@xmath22 _ from the 2mass survey @xcite , and _
space telescope @xcite data from both the original s@xmath23mc @xcite survey and the combined s@xmath23mc / sage - smc catalogs @xcite , making use of the irac @xcite and mips @xcite instruments . in the sage - smc processing ,
combined s@xmath23mc / sage - smc images were created by stacking data from both surveys before performing the photometry . for the mips combined catalog , on the other hand , photometry was performed on the three different observing epochs ( one for s@xmath23mc , and two for sage - smc ) separately , after uniformly reprocessing the data .
we searched the combined catalogs for coordinate matches to the dusty ob stars within 2 .
the irac and 2mass data points in the original s@xmath23mc photometry were replaced with the points in the combined catalog if they existed .
if more than one match was found , the closest match was used .
the 24 points in the original s@xmath23mc catalog were replaced with the weighted average of the three epochs in the combined catalog for the the closest positional match .
points were replaced as long as there was a match within 2 in at least one of the epochs .
the 70 data points , where they exist , must be treated with caution .
the diffuse emission in the smc is considerably brighter at 70 than at 24 , and the difference in beam size between the two channels is substantial .
the beam size at 24 is 6 , corresponding to 2 pc at the smc distance of 61.1 kpc @xcite , while the beam size at 70 is 18 , corresponding to 5 pc .
thus there is a good chance of source confusion between the two channels , and we have not searched the combined s@xmath23mc / sage - smc catalog for 70 matches .
the photometry data are given in tables [ tab : phottab ] and [ tab : phottab2 ] .
we use our long - slit spectroscopy to classify the stars . for the classification
, we employ the scheme described in @xcite , which compares the strengths of a number of lines in the 40005000 range .
we use the equivalent widths of the lines , guided by visual inspection of the spectra , to compare the line strengths , but our low resolution leaves some of the important lines blended with h@xmath24 ( 4088 and 4116 ) or with each other ( e.g. , 4471 and 4481 ) .
note also that the original classification scheme was designed for supergiants , and so the si lines for main - sequence stars may be too weak to detect in some cases .
the classification scheme is summarized in table [ tab : class ] . a sample spectrum of object b107
is shown in figure [ fig : samp ] . since b107 shows no absorption features , it is not an o star .
it shows no evidence of 4553 , so it is not a b2 or b3 , but the 4481 appears too weak for a spectral type later than b3 .
thus we conclude that most likely the lines are blended with h@xmath24 and that the star is a b1 . of the 125 objects for which we obtained spectra , 87 appear to be normal main - sequence stars and one ( b167 )
is discussed below as a likely supergiant . of those 87
, 53 are classified via the above scheme and assigned a subtype .
the remaining 34 stars were difficult to place precisely within the scheme , but they have been marked as either o7o9 , if they show absorption , or b0b2 if they lack the lines .
b167 is unusual , however , in that it shows no significant absorption but has very strong balmer absorption .
it appears to be an a star ; given its luminosity it is probably a supergiant .
the spectra and seds of these 88 objects are shown in figures [ fig : norm1 ] through [ fig : norm6 ] , while the assigned spectral type is given in table [ tab : spec ] as well as in the figures .
the remaining 37 of the 125 objects , which are not considered in the dust analysis in the next section , fall into one of four groups : 1 ) stars that show balmer emission and sometimes / emission ( 16 stars plus b094 ) , 2 ) stars with unusual profiles ( 7 stars , including b035 which we especifically discuss in the following section ) , 3 ) objects that show strong forbidden line emission ( 6 objects ) , and 4 ) spectra that are difficult to extract ( 8 stars ) . the first group of 17 stars shows the balmer and /ii emission lines characteristic of be stars ( figure [ fig : be ] ) , while the second group of 7 stars shows only weak absorption ( figure [ fig : flat ] ) .
b094 stands out , as it is the only object among the 17 stars that shows [ ] @xmath25 6548,6583 emission .
we include it here , rather than with the strong forbidden - line emission objects , because of the lack of [ ] emission and the weakness of the balmer emission compared to the other objects in that group .
the unusually weak balmer absorption in the second group of stars may indicate that these objects are be stars as well .
@xcite modeled the profiles of be stars and showed that , for many combinations of base - disk density , density distribution , and viewing angle , the emission from the circumstellar matter does not cancel out the absorption line of the star completely .
this effect could explain the unusual profiles .
the definite be stars are noted with a spectral type of obe in table [ tab : spec ] , while stars with possible weak emission are noted as `` obe ? '' in the table .
the third group is comprised of 6 objects ( 5 shown in figure [ fig : chii ] , plus the already discussed b094 ) , and it includes b001 , b028 , and b079 , all of which show strong balmer and [ ] @xmath25 4959,5007 emission , along with weaker [ ] @xmath25 6548,6583 and @xmath25 5876,4921 , and 4471 emission .
these spectra are typical of compact regions .
b141 is dominated by balmer emission , with weak [ ] and very weak emission .
b141 also shows moderate [ ] emission , but unlike b001 , b028 , and b079 , the forbidden - line emission is weak relative to h@xmath7 .
b036 is now known to be a young stellar object ( yso ) @xcite ; it shows strong balmer emission , with weak [ ] emission , very weak [ ] emission , and little to no emission above the continuum level .
the seds of these five sources are much flatter than those of the non - emission - line objects , possibly due to the predominance of the free free emission over any stellar photospheric emission .
these objects are marked in table [ tab : spec ] as em in the spectral type column .
the fourth group , shown in figure [ fig : diff ] , includes b005 , b011 , b112 , and b175 , all of which have low signal - to - noise spectra and , except for b011 , also have a star too nearby to set the background well .
thus it is unclear whether the absorption appears weak because of noise , background contamination , or weak be star emission .
the other 4 objects in this group are b006 , b093 , b108 , and b191 , whose complicated background in makes it difficult to disentangle emission from the surrounding nebula from possible be star emission .
these objects are given no spectral type in table [ tab : spec ] .
in the `` obe ? '' group , b035 is the only star that shows 4686 emission .
the spectrum is displayed in more detail in figure [ fig : b035 ] .
the appearance of the 4686 emission indicates a very high luminosity , class ia @xcite .
this is commonly accompanied by emission of the 4640 complex , but this is weak in the spectrum of b035 .
we suggest that this is a metallicity effect due to the low nitrogen abundance in the smc @xcite
. it does appear to be present , hence we tentatively affix the `` f '' suffix to the spectral type . with regard to the temperature class , the 4471 / 4200 ratio suggests a spectral type of about o6 @xcite .
the + ii 4026 line is weaker than that of the 4200 line , which would suggest a spectral type @xmath26 o4 , but it is also weaker than that of the 4471 line , which would suggest a spectral type @xmath27 o7 ( which does not really happen , @xcite ) we attribute this to uncertainties and/or possible binarity and arrive at a final spectral classification of o6iaf .
accurate spectral classification based on high quality data was carried out by @xcite , @xcite , @xcite and @xcite for 23 of our 125 stars ( see table [ tab : spec ] ) .
our own spectral classifications are in agreement with those determinations , with only a few deviations by one or two subclasses .
the majority ( @xmath28 ) of those classifications are of luminosity class iv
v , with only 6 of luminosity class ii and none more luminous .
this suggests that most of the stars in our sample are unevolved , main - sequence stars . the sample , however , is a mixed bag of curious and seemingly very common types of objects .
the `` em '' objects include the planetary nebulae b001 = lha115-n9 and b079 = lha115-n47 , the yso b036 @xcite , and the compact region b028 = lha115-n26 @xcite .
the sample further includes the x - ray binaries b064 = sxp214 @xcite and b085 = cxouj005245.0@xmath29722844 @xcite and no less than 10 or 11 eclipsing binaries .
we attribute the large number of eclipsing binaries to the high overall fraction of close binary systems among massive stars @xcite .
in order to assess the viability of the various hypotheses for the nature of the dusty ob stars and the origin of the dust , we now use the available photometry to estimate the gross properties of the dust .
the primary quantities we can constrain are the dust temperature , dust mass , the fraction of the star s light re - radiated by the dust ( f@xmath8/f@xmath9 ) , and the distance of the dust from the star .
we also calculate the grain size below which radiation pressure will quickly remove the dust from the system in the absence of gas drag , as well as the mass of grains of that blow - out size that would be needed to account for the observed f@xmath8/f@xmath9 . to determine these properties ,
we first find the amount of expected flux from the stellar photosphere by fitting to the sed a power law of the form @xmath30 with @xmath31 in hz and @xmath32 in , and with the slope @xmath33 derived from the kurucz stellar atmosphere model @xcite of the spectral type determined .
the intercept @xmath34 was determined using the @xmath35-band flux , or @xmath36 if @xmath35 was unavailable , to scale the models .
the @xmath35 and @xmath36 fluxes were corrected for the line - of - sight extinction due to the milky way using @xmath37 and @xmath38 from @xcite and @xcite , with @xmath39 mag @xcite .
we selected models for @xmath40=-0.5 $ ] , closest among the kurucz library to the typical abundance of young stars and gas in the smc of @xmath40\sim-0.7 $ ] @xcite , and the suggested surface temperatures and surface gravities for specific stellar types from @xcite and schmidt - kaler @xcite .
all the fits used models with @xmath41 for the surface gravity , and the surface temperature and slope used for each spectral type is given in table [ tab : class ] .
the temperature scale is similar
well within 10% to that of galactic main - sequence stars determined by @xcite and that of o - type stars in the smc determined by @xcite ; the latter found that early - o type stars in the smc are hotter than galactic stars of the same spectral type , but that this difference vanishes around b0 .
a slope @xmath42 was used if the star was given a spectral type of `` o7o9 '' , while a slope @xmath43 was used for all other stars without a specific spectral type . from these fits ,
the excesses at 4.5 , 8 , and 24 were calculated ( table [ tab : excess ] ) , along with the ratio of the excess to the expected photosphere . in all but six cases , marked with a footnote in table [ tab : excess ] ,
no emission was detected at 70 , so we can not fit a model to the 8 , 24 , and 70- excesses directly .
instead we estimate the color temperature of the dust using the 8- and 24- excesses , and place limits on the color temperature using the 24 flux and the 70 flux limit of 40,000 @xmath1jy , minus the expected photospheric flux .
the color temperatures are estimated assuming a modified blackbody with emissivity @xmath44 , using the equation @xmath45 where @xmath46 is the shorter wavelength and @xmath34 the longer .
the color temperatures calculated for the ratios of the 8- to 24- excesses , for stars with a specific spectral type , are given in table [ tab : ourdust ] . because the 8- band includes a major polycyclic aromatic hydrocarbon ( pah ) feature , these color temperatures must be treated with caution .
note , however , that pah emission in the smc is considerably weaker than in the milky way making this less of a problem than for local samples @xcite .
the color temperatures calculated with the 24- to 70- ratios are very uncertain , since in most cases we have only an upper limit on the 70- flux , and thus are not included in the table .
these values are generally in the range of 4060 k , but they should be considered as lower limits as a 70- flux below the upper limit would lead to a higher derived temperature .
mid - ir spectroscopy and deeper 70-@xmath1 m photometry to determine the dust temperature directly are presented in paper ii .
we are working in the optically thin limit at long wavelengths , so the emission of the dust is given by @xmath47 where @xmath48 is the excess emission at 24 , @xmath49 is the solid angle subtended by the source , @xmath50 is the opacity due to the dust , and @xmath51 is the emission of a blackbody at the temperature of the dust .
the opacity of the dust is @xmath52 where @xmath53 is the mass of the dust , @xmath54 is the area covered by the dust , and @xmath55 is the dust mass absorption coefficient .
thus the dust mass is given by @xmath56 here we assume @xmath57 where @xmath58 has units of @xmath59 g@xmath60 and @xmath32 is given in .
this value is due to @xcite and within 20% of @xcite .
we adopt it to be easily comparable to surveys of the milky way ( for example , the gould belt survey ; * ? ? ?
the dust mass then is @xmath61 where @xmath48 is given in mjy , @xmath62 in pc , @xmath32 in , and @xmath51 in erg cm@xmath6 s@xmath60 hz@xmath60 sr@xmath60 .
the dust masses calculated using temperatures the derived from 8/24- excess ratios are listed in table [ tab : ourdust ] . to calculate the fractional luminosity f@xmath8/f@xmath9 , we first determined the infrared flux by integrating over the source function for graybody emission from 0.7 @xmath1 m to 100 @xmath1 m , rewriting eq .
[ inu ] above as @xmath64 where @xmath53 is in m@xmath65 , @xmath62 is the distance to the source in pc , @xmath32 in , and @xmath48 in erg cm@xmath6 s@xmath60 hz@xmath60 .
we then divided by the bolometric flux , which is the luminosity for the spectral type ( table [ tab : class ] ) , scaled to the distance of the smc .
the observed dust temperature and stellar luminosity constrain the location of the dust in the circumstellar environment .
the distance of the dust from the star at a given temperature also depends on the size of the dust grains and the wavelength range of the light being absorbed .
we can set limits on the dust distance by first assuming the relation for large , blackbody grains , @xmath66 where @xmath67 is the stellar radius in units of r@xmath68 and @xmath69 is the dust distance in au .
the stellar radii for the stars are given in table [ tab : class ] and are taken from a second - order polynomial fit in @xmath70 to the values given in appendix e of @xcite .
the distances calculated with this model are given in table [ tab : ourdust ] as the blackbody ( bb ) distance .
grains that instead absorb light as blackbodies , but re - radiate as a modified blackbody proportional to @xmath71 , can be modeled using the equation @xmath72 where @xmath73 is in l@xmath68 , @xmath69 is in au , and @xmath74 is in .
this equation is taken from equation 2 of @xcite , assuming @xmath75 and @xmath76 .
we have assumed @xmath77 .
note that this last parameter is poorly constrained and it introduces significant uncertainty in @xmath78 .
the distances calculated with this model are listed in table [ tab : ourdust ] .
@xcite note that in resolved debris disks , the actual distances of the particles are at most 3 times the distance predicted by the blackbody grain model , whereas in resolved cirrus hot - spots , the dust distance is an order of magnitude or more
larger than that expected from the blackbody model , due to stochastic heating of small grains @xcite .
we also estimate the blow - out size for grains due to radiation pressure in the absence of gas drag .
grains with @xmath79 will be blown out of the system by radiation pressure .
we use the equation for @xmath7 from @xcite , @xmath80 where @xmath46 is the size of the grain , @xmath24 is the density of the grain , and @xmath81 is the average radiation pressure efficiency over the stellar spectrum for a given grain size .
values for @xmath82 for the spectral types are given in table [ tab : class ] and were derived using @xmath83 where @xmath84 k from schmidt - kaler @xcite .
values for @xmath85 were derived by fitting a second - order polynomial in @xmath70 in logarithmic space to the values in appendix e of @xcite and are also given in table [ tab : class ] .
we assume here @xmath86 , and we use @xmath87 g cm@xmath11 from the @xcite model .
setting @xmath89 gives the minimum grain size in a circular orbit that can survive the radiation pressure of the host star .
grains released from parent bodies in circular orbits , such as dust created in a debris disk , however , can be ejected from the system for @xmath7 as low as 0.5 .
we then estimate the mass of dust needed to produce the observed f@xmath8/f@xmath9 , based on the estimated blow - out size of the grains .
this mass ( in units of g ) is given by @xmath90 where @xmath91 is the minimum grain size to survive radiation pressure blow - out , in cm , and @xmath69 is the blackbody distance in cm .
we estimate the uncertainty on these parameters for the 8/24 pair of excesses using the monte carlo method . for each of 1000 trials per object ,
the color temperature is calculated , and that temperature is used in the calculation of the dust mass , f@xmath8/f@xmath9 , and dust distance .
the uncertainties on the color temperature , dust mass , f@xmath8/f@xmath9 , and dust distance are then set according to the upper and lower values that bracket the central 67% of the results for each object .
uncertainties on the excess at 8 and 24 are determined via error propagation , assuming an uncertainty in slope @xmath33 of 0.05 for b stars and 0.01 for o stars .
these ranges allow for uncertainty in the spectral type . because the o stars all have a similar slope , the range is much smaller .
we apply these uncertainties to the excesses assuming a normal distribution about the value of the excess before calculating the color temperature .
the dust mass calculation also includes an uncertainty in the distance to the stars , which we model as a normal distribution centered at @xmath92 kpc , assuming that the objects are most likely to be located near the center of the smc .
the f@xmath8/f@xmath9 calculation further requires the uncertainty on the bolometric luminosity , assumed to be a uniform distribution with a range set by the value for @xmath93 spectral sub - type . for the distance calculation
, we assume uncertainties for the stellar radius and stellar temperature uniformly distributed between the values for @xmath93 spectral sub - type . table [ tab : ourdust ] gives the dust parameters for those objects in table [ tab : excess ] that have a measurable excess at 8 @xmath1 m , and to which we give a secure main sequence spectral classification .
because the majority of our sample shows no optical emission lines , be star activity is likely not the main cause of the 24- excess in these o and b stars .
it is possible , however , that a few of the objects without emission lines could still be be stars , as be stars are known to show variability in their balmer emission . in a study of 45 milky way be stars , @xcite found 23 objects with variable emission , including six that changed from strong emission to normal b - star absorption at or vice versa over the course of just two years .
of the original sample of 193 o and b stars with 24- excess identified in the s@xmath0mc survey , 12 were previously catalogued as be stars by @xcite , and for those we did not acquire new spectra .
we identify here an additional 16 stars with confirmed emission and 7 stars with suspected weak emission , for a total of 35 be stars ( 36 if we include b097 in the count ) .
the remaining 88 stars in the spectroscopic sample do not display evidence of obe activity . to investigate the possibility of variability between the two states , we look at the 4.5- excesses , seen in table [ tab : excess ] , because be stars , both classical and herbig ae / be , are expected to show near - ir emission to some degree ( see , e.g. , * ? ? ?
indeed , we do see an excess ratio @xmath94 above the photospheric emission at 4.5 for 11 of the 16 obe stars in our spectroscopic sample , as well as in 2 of the 7 `` obe ? '' stars .
this could indicate that the other 10 obe and `` obe ? '' stars were in a normal b - star phase during the irac observations , acquired in 2005 may .
these excesses , however , are weak , at less than 4 times the expected photospheric flux , suggesting that we may just be unable to detect the excesses in our remaining spectroscopically confirmed obe stars because of the uncertainties ( @xmath95 ) in their photospheric flux in our data . adding further uncertainty is our use of @xmath96band data to scale the photosphere , because be stars may exhibit excesses even at wavelengths as short as 1.2 .
this can be seen in the slopes of the dashed lines in the seds of b038 , b081 , b085 , b119 , b140 , and b190 , for example , in figures [ fig : be ] and [ fig : flat ] .
the dashed lines are fit to @xmath97 and show shallower slopes than expected for the photospheric emission .
we also find 12 stars ( b023 , b033 , b047 , b115 , b128 , b135 , b136 , b142 , b154 , b156 , b160 , and b183 ) in table [ tab : excess ] that do not show optical emission lines ( see table [ tab : spec ] ) but do have an excess ratio at 4.5 significant at the @xmath98 level .
the 4.5- excesses are also weak , with only b136 showing an excess / photospheric ratio greater than 1 .
this excess could indicate that these stars were in a be state at the time of the irac photometry but settled into a normal b - star state by the time the spectra were acquired .
if we assume that all of the 4.5- excesses are due to be - star behavior , the fraction of transient be - stars ( counting as transient those obe and `` obe ? '' optically classified stars without 4.5 excess , plus these 12 stars with 4.5 excess but no hint of h@xmath99 emission in their spectra ) is 22/47 , or about 47% of the total be - star population in the sample .
this is on par with the fraction seen in the study of galactic be - stars by @xcite .
therefore it seems unlikely that the remainder of our stars that show no be activity in the photometry and spectra are primarily dormant be - stars . because we do not know the ages of the
be stars in our sample , it is difficult to determine to which of the be - star categories they belong .
none of the be stars we identified spectroscopically have the large excesses at @xmath100 that are typical of the hot dust in herbig ae / be objects .
@xcite , however , identify in the galaxy a set of objects with near - ir profiles that resemble classical be stars , but are apparently young stars , located near reflection nebulosity .
this location is more typical of herbig ae / be stars .
@xcite suggest that these objects , which they label group iii , are actually herbig ae / be stars that have an optically thin circumstellar disk , rather than the optically thick disk more typical of the herbig ae / be class .
the weak 4.5- excesses of the be stars detected in our spectroscopic study resemble the near - ir profiles of classical be stars as well as these group iii objects .
the stars in our sample are located within regions of recent star formation , suggesting an age of @xmath101 myr .
unfortunately , 10 myr is the approximate age at which classical be stars begin to appear in clusters @xcite , and so we can not be certain whether these objects are young stars or older , established main - sequence stars without more accurate age determinations .
the bulk of the dusty b - star sample , with neither optical emission lines nor a near - ir excess , are unlikely to be herbig ae / be stars .
@xcite note that to be certain that debris disks are the source of the excess infrared emission , spatially resolved observations of the disk are needed , because hot spots in the ism can mimic the color temperatures and fractional luminosities of debris disks .
the morphology of the dust is the best clue as to whether the heated dust is physically associated with the star in a disk , or if it is simply a concentration of ism material close to the star .
unfortunately , at the distance of the smc , debris disks can not be resolved with any available facilities , and ism hot spots are below the resolution of the _ spitzer _ data , especially with a 6 resolution at 24 .
since we lack resolved images of the dust emission , we instead compare the dust properties of the smc sample to the galactic cirrus hot - spot sample of @xcite .
a broader comparison using _ wise _ data is carried out in paper ii .
the @xcite sample was selected by looking for _ iras _ excesses at 60 with color temperatures above the @xmath102 k temperature of the general galactic cirrus and peaking at the location of a star , with @xmath103 to exclude regions , and spatially extended emission that is more extended in longer wavelengths , indicating that the temperature declines with distance from the star .
we note that in cases where the reported _ iras _
25- flux is negative , the color temperatures can not be calculated .
whether this cirrus hot - spot sample in fact consists solely of cirrus hot - spots is , however , uncertain .
@xcite found that in some cases , the hot spots have 12- emission that is brighter than the 25- emission , indicating a hotter dust component than the expected @xmath17 k temperature of illuminated cirrus .
indeed , the color temperatures we find for the hot - spot sample from the 12/25 flux - ratios are all above 100 k. furthermore , @xcite used resolved optical images of nebulosity around three b stars that appear in both the @xcite and @xcite samples to model the ir emission from the objects .
the authors concluded that , in all three cases , the ism model can not explain the 12- and 25- emission , requiring a source of warmer dust .
the coronograph used to block the starlight for their optical images unfortunately obscures the central 1001000 au around the star , where a circumstellar disk would be located .
the warmer dust could therefore be provided by a population of either blackbody grains in a close - in circumstellar disk , or very small grains at larger distances that are heated stochastically to higher temperatures than equilibrium . in cases where the stellar velocity is large compared to the local ism ( such as runaway stars ) , a bow - shock could further heat the dust in addition to stellar irradiation . distinguishing between these scenarios would require imaging at higher resolution , which we discuss further in paper ii .
the 12- iras band also contains a pah emission feature , offering another possible explanation for the excess flux at that wavelength . illustrating the difficulty of distinguishing disks from hot - spots using sed information alone , even for nearby objects , @xcite note that 31 of the 34 b - type stars in the @xcite table of debris disk candidates or vega - like stars , appear in the @xcite cirrus hot - spot list .
the full hot - spot sample contains stars of spectral types o6 through b9 of all luminosity classes , but we exclude from the comparison any objects marked in the @xcite table as emission line stars or stars of luminosity class higher than v , to match the characteristics of the smc dusty ob stars .
we further limit the hot - spot sample to those objects that have excesses , if scaled to the smc distance of 61.1 kpc , that we could detect at the 5-@xmath104 level or better at 24 in the s@xmath0mc survey ( 215 @xmath1jy ; @xcite ) .
this sub - sample contains 18 stars .
we then calculated the dust parameters for this sub - set of the sample , using the 12/25- combination in place of 8/24- , and these values are listed in table [ tab : gaudust ] .
stars flagged in table [ tab : gaudust ] as `` known reflection nebula '' are noted in @xcite to have extended emission that is brighter in the blue than in the red on the palomar sky survey or eso sky survey plates .
the comparison smc sample contains a total of 31 stars ( c.f . , table [ tab : ourdust ] ) , and
it is constituted by normal main sequence stars ( no obe , `` obe ? , '' or em classification ) that have a measurable 8 emission excess , in order to determine the temperature of the dust . while the distributions in both samples over spectral types peak at a similar type b0 in the smc vs. b1 in the hot - spot sub - sample the smc sample extends more towards earlier types while the hot - spot sub - sample extends towards later types .
this is reflected in the distributions over stellar effective temperature , shown in figure [ fig : maxhist ] . in figure [ fig : maxobs ]
we show histograms of the observable parameters , f@xmath105/f@xmath106 or f@xmath107/f@xmath108 , l@xmath106 or l@xmath108 , and @xmath109/f@xmath9 or @xmath110/f@xmath9 .
the ir luminosities and fractional ir luminosities are similar ; the difference in the ir ratio is largely driven by the use of the 8- vs. 12- bands . in figure
[ fig : maxdust ] we show histograms of the color temperature , mass , and fractional luminosity , where the samples are restricted to spectral types o9b2 , since both samples have at least one star of each of these spectral types .
the histograms for the hot spots in figures [ fig : maxobs ] and [ fig : maxdust ] are created by weighting the number of objects in that sample so that the total number of hot spots in a spectral type bin equals the total number of objects in the bin for the smc sample . in this manner the weighting controls for the different spectral makeup of the two samples .
figure [ fig : vstemp ] shows the averages of each of the observable parameters and associated dust parameters against stellar temperature , without the restrictions on visibility at the smc distance that are used in the histograms .
the distributions over dust temperature are very similar for both samples .
they both have similar mean temperatures in the 125130 k range , although the hot - spot sample also displays a broader range in dust temperatures , 97226 k compared to 116164 k for the smc sample .
the hottest color temperatures in the hot - spot sample in figure [ fig : maxobs ] result from the coolest and most predominant spectral type , b2 .
the trend toward higher color temperatures in the hot - spot sample is possibly due to pah contamination .
the 12-pah emission feature remains relatively unchanged over a broad range of radiation field intensities , but the thermal emission at 25 changes dramatically ( e.g. , figure 13 in @xcite ) .
thus pahs around cooler stars would create higher color temperatures than around hotter stars ; f@xmath107/f@xmath108 would be artificially high as a result of similar amounts of pah emission at 12 but reduced thermal emission at 25 .
this is likely a lesser effect in the smc , where pah emission is weak ( e.g. * ? ? ?
the distributions over dust mass are also similar for both samples , though again the distribution appears broader ( both towards higher and lower dust masses ) in the hot - spot sample . the smc sample may be devoid of examples of lower dust mass as these would not have been detected .
the average mass at b0 and b1 , where there is the greatest number of objects in both samples , is similar for each sample ( fig .
[ fig : vstemp ] ) , but the average dust mass in the smc sample is higher at b2 , and lower for o7o9 , compared to the cirrus hot - spots .
we see a trend in the hot - spots of lower dust mass for cooler stars , which is reasonable given that a hotter , more massive star could illuminate larger regions of the ism .
we do not see this dependence on the stellar temperature for the smc objects , although the smaller temperature range of the smc sample would make such a trend perhaps difficult to detect . at first glance
, the distributions over @xmath111 do also appear similar between the two samples . on closer inspection ,
the one for the hot - spots is clearly broader .
the average @xmath112 for the hot - spot sample is smaller than that of the smc , which is @xmath113 , although this is somewhat misleading as the averages are skewed by the extremes in the sample .
the derived @xmath111 for both samples do fall within the typical @xmath114 to @xmath115 range of debris disks @xcite . while we noted minor differences in the distributions over dust temperature , dust mass and @xmath111 between the smc and hot - spots samples ,
these can be explained by a combination of small - number statistics , selection bias , and the opposite sense of skewedness of the distributions over spectral type .
indeed , figure [ fig : vstemp ] shows that the trends or absence thereof and mean values are consistent between both samples if the breakdown over spectral types is taken into account .
interestingly , despite the low dust - to - gas ratio in the smc ( about @xmath116 that of the milky way ; * ? ? ?
* ) , the smc objects show dust masses similar to the galactic hot - spots if common selection thresholds are applied .
this suggests that , if they are hot - spots , the mass of gas being illuminated is @xmath117 times larger in the smc than in the milky way . assuming similar gas densities
, this suggests that the typical sizes of hot - spots would be @xmath18 larger in the smc than in the milky way .
larger dust masses , warmer dust and thus brighter emission may arise from bow - shocks ahead of runaway stars , or stars being over - run by expanding interstellar bubbles .
stars with stronger stellar winds
i.e. , of earlier type and/or at later evolutionary stages and/or larger relative velocity with respect to the local ism sweep up more mass ; larger relative velocities also result in stronger shocks and hence greater heating .
@xcite present a sample of 12 candidate runaway stars with bowshocks in the smc .
curiously , there is no overlap whatsoever between their sample and ours .
the distribution of our sample on the sky ( figure [ fig : map ] ) seems to indicate that at least a portion of them are found outside of sites of vigorous star formation .
while the late - b stars may be old enough to have migrated from their birth - sites or for their natal clouds to have dispersed , the fact that some of the o - type stars are also found outside star - forming regions suggests they may be runaways .
one of our `` normal '' ob stars , b064 is in fact the primary in an x - ray binary , and it is likely that the binary system will have received a high peculiar velocity when the neutron star s progenitor star exploded as a core - collapse supernova see @xcite for the proto - typical example of a bow - shock accompanied runaway x - ray binary , velax-1 .
we also note that the structure of the ism in the smc differs in important ways from that in the milky way disk ; the smc is dominated by expanding bubbles @xcite while gas in the milky way disc is more strongly entrained and collected by the recurrence of a spiral density wave .
this could affect the ubiquity and properties of bow - shocks ; for instance stars in the smc might be over - run by expanding shells ( see * ? ? ? * for such mechanisms operating within the large magellanic cloud ) , resulting in large relative velocities between stars and ism without the requirement for a `` kick '' velocity of the star .
while proper motions relative to the systemic motion of the smc are too small to be measured ( 1 mas yr@xmath60 @xmath118 300 km s@xmath60 ) , radial velocities can be determined with much better accuracy .
our own spectroscopy , however , has neither the spectral resolving power nor velocity calibration accuracy required for such measurements . in table [ tab : spec ] we list the radial velocities measured for stars in our sample by @xcite , @xcite and @xcite .
we also determined the velocity of the peak in the emission around the location of the star in question , from inspection of the datacube produced by @xcite .
this differs from the intensity - weighted velocity map produced by @xcite , because we are interested in the most likely velocity difference between a star and the densest parts of the ism .
often , more than one strong peak is seen in the data ; in that case we list both . of the 21 stars with radial velocity measurements ,
most stellar kinematics coincide with that of a strong peak .
b116 , b124 ( and perhaps b119 ) fall in between peaks , suggesting they might reside in the middle of a shell expanding at @xmath1620 km s@xmath60 . but two o7 stars , b142 and especially b145 , have large velocity differences of @xmath119 and 75 km s@xmath60 with respect to the nearest ( in velocity ) peak .
no uncertainties were quoted for the stellar radial velocities for these stars , but measurements for other stars from the same work ( with the caveat that those were all later spectral type ) often agreed with kinematics , which suggests typical errors not much larger than @xmath120 km s@xmath60 .
we thus suggest that these two o7 stars may be runaways they are located just outside of the brightest region in the smc , lha115-n66 containing the massive o - star cluster ngc346 , lending further support to their large space velocities . statistics for circumstellar disk detections around stars with spectral types as early as the smc sample are difficult to come by .
few surveys exist of stars younger than @xmath121 myr , which encompasses most , if not all , of the main - sequence lifetimes of these late - o and early - b stars , and those surveys that do sample the appropriate age range still contain very few stars of these types ( see * ? ? ?
* and references therein ) .
our sample of 87 objects with normal o- or b - star main sequence spectra in the smc shows excesses at 24 that are larger than those of typical debris disks , in studies of somewhat less massive ( late - b and a - type ) stars in the milky way .
@xcite set a limit , based on the properties of their sample of b7a9 stars , for the ratio of observed flux to expected photospheric flux at 24 of @xmath122 for debris disks ( but only a ratio of @xmath123 for solar - type stars ) . any object with a larger ratio
is considered to be a primordial disk . @xcite
define objects with @xmath124 \gtrsim 5 $ ] mag as herbig ae / be stars , while objects below this limit are debris disks .
@xcite also define any object with @xmath125 \lesssim 5 $ ] mag as a massive debris disk .
the smc objects have observed flux to expected photospheric ratios @xmath126 , with only 6 objects being less than 100 and a further 7 objects being within 1 @xmath104 of 100 .
the smc objects also have @xmath124 \gtrsim 3.4 $ ] mag , with only 11 objects below the limit of 5 . of those 11 , we identify 7 as obe stars .
the studies of @xcite and @xcite only consider circumstellar matter around intermediate and low mass stars , so it is possible that these dividing limits around the more massive stars in our smc sample should be higher .
@xcite , however , specifically excluded earlier type stars from consideration because , aside from being few in number , they may show a severely reduced dust luminosity due to the loss of small grains to radiation pressure .
we find blowout grain sizes for the smc objects on the order of 1 mm , in the absence of gas drag ( see table [ tab : ourdust ] ) .
the smallest dust grains are the most efficient ( per unit mass ) at capturing and reprocessing uv and optical radiation into infrared .
this suggest that , if blowout is significant , @xmath111 started much larger than we currently measure in our sample .
@xcite suggest two scenarios to explain massive circumstellar dust disks
. they could be the result of collisions between two large planetesimals , greater than 1000 km , or they could be explained as a transition phase from herbig ae / be star to debris disks , in which the disk still contains a significant fraction of primordial dust .
the small number of these objects detected in the smc sample suggests that they could be in a short - lived phase , and perhaps constitute transition disks remnants from the stellar accretion . in the original s@xmath0mc sample ,
the 24- excess objects were 5% of the total number of objects that met the requirements for normal late - o and early - b stars in the specific magnitude and color cuts @xcite .
counting only the stars that have normal o- and b - star spectra , which is 87 out of the 193 objects in the original list , and also assuming that all of these objects are debris disks , rather than ism hot spots or quiescent be stars , gives a detection fraction of only 2.2% . including the 45 objects for which we have no spectra to judge their emission - line star status
gives a maximum of 3.4% .
if the possible transition object that @xcite detect is truly in this transition phase , that would give a detection fraction of 3% for transition objects in the 5 myr cluster in which it lies .
this detection fraction is consistent with that seen in the smc sample . finally , it is worth noting that comparison with literature classifications shows that a significant fraction of our smc sample consists of binaries .
the effect of binary systems on the formation of circumstellar disks is unclear .
one might expect the disks to be dynamically disrupted , although circumbinary planets have been recently discovered by _
kepler _ @xcite .
we obtained long - slit spectra of 125 of the 193 dusty ob stars in the s@xmath0mc survey that show large excess emission at 24 .
these new observations cover a significant fraction of the sample , make possible accurate sky removal in the complex regions where these stars live , and allow us to draw robust conclusions on the statistics of different contributors to the sample .
we use these spectra to classify the stars and look for the signatures of the obe activity .
we find 87 objects that lack emission lines and appear to be normal , main sequence late - o / early - b main - sequence stars .
for these stars , we use our spectral classification to estimate their luminosity , and calculate the dust temperature , dust mass , f@xmath8/f@xmath9 , and equilibrium dust distance from the host star .
we discuss the spectral and photometric properties of our sample in relation to several possible scenarios : classical be stars , massive analogs of herbig ae / be stars , hot - spots in the interstellar medium ( possibly with contributions of bow - shocks ) , and disks .
the results of this study are : * we identify 17 stars in the dusty o and b stars sample as obe stars because of clear emission in their spectra ( this count includes b094 , the object with forbidden nitrogen emission ) .
a further 7 stars exhibit weaker than expected absorption and are therefore classified as possible obe stars one of these is an o6iaf star showing 4686 emission .
we also identify 12 stars that show no emission in their spectra but do show 4.5 excess in the photometry that could be indicative of transient be - star activity .
if these 12 stars are dormant obe stars , along with the 11 spectroscopically identified obe stars that show no excess at 4.5 , the transient total obe star fraction would be @xmath127% .
this fraction is consistent with the study of transient be star activity by @xcite .
thus it seems unlikely that the remaining 75 stars , with neither 4.5 excesses nor emission , are obe stars that happened to be in normal ob star states during both the photometric and spectroscopic observations . * our dusty ob stars
are generally located near regions of recent star formation , similar to herbig ae / be type stars , and in that respect resemble the group iii herbig ae / be objects in @xcite .
our objects have ages @xmath101 myr .
note that classical be - star activity begins to appear in clusters around @xmath121 myr of age , once again suggesting that the majority of them can not be explained by excretion disks with a dusty component . * the smc sample exhibits broadly similar characteristics when compared to the cirrus hot - spot sample of @xcite .
the hot - spot sample shows trends for decreasing dust temperature and increasing dust mass for hotter stars , which the smc sample does not . however , these differences likely arise from the different distributions over spectral type , and perhaps stronger contamination of the broadband colors by pahs in the milky way .
we show evidence that at least some of the stars in the smc sample are runaways , including one x - ray binary ; heating in a bow - shock could add to the radiative heating of the dust .
* the smc sample of objects show higher 24 excesses than typical debris disks around intermediate mass stars in the milky way .
there are not enough debris disks known around comparably massive stars , however , to indicate what is typical for an early - b or late - o debris disk . alternatively ,
if the stars are young , we may be observing the remnants of the accretion process analogs to transition disks around massive stars .
the detection fraction of these objects when compared to stars with similar optical colors and luminosities is 2.2%3.5% , which is consistent with the expected short lifetime of a transition phase .
adams , j. j. , simon , j. d. , bolatto , a. d. , sloan , g. c. , sandstrom , k. m. , schmiedeke , a. , van loon , j. th . , oliveira , j. m. , & keller , l. d. 2013 , submitted ( paper ii ) allen , c. w. 1973 , london : university of london , athlone press alexander , r. d. , clarke , c. j. , & pringle , j. e. 2006 , , 369 , 229 aller , l. h. , et al .
1982 , landolt - bornstein : numerical data and functional relationships in science and technology andr , p. , menshchikov , a. , bontemps , s. , et al .
2010 , , 518 , l102 antoniou , v. , hatzidimitriou , d. , zezas , a. , & reig , p. 2009
, , 707 , 1080 arny , t. 1977 , , 217 , 83 backman , d. e. , & paresce , f. 1993 , protostars and planets iii , 1253 bolatto , a. d. , et al . 2007 , , 655 , 212 bonanos , a. z. , et al . 2010 , , 140 , 416 burns , j. a. , lamy , p. l. , & soter , s.1979 , icarus , 40 , 1 carpenter , j. m. , mamajek , e. e. , hillenbrand , l. a. , & meyer , m. r. 2009 , , 705 , 1646 castelli , f. , & kurucz , r. l.2003 , modelling of stellar atmospheres , 210 , 20p cesaroni , r. , galli , d. , lodato , g. , walmsley , c. m. , & zhang , q. 2007 , protostars and planets v , 197 chen , c. h. , et al .
2006 , , 166 , 351 cieza , l. a. , schreiber , m. r. , romero , g. a. , et al .
2010 , , 712 , 925 coe , m. j. , et al .
2011 , , 414 , 3281 dawson , j. r. , mcclure - griffiths , n. m. , wong , t. , dickey , j. m. , hughes , a. , fukui , y. , & kawamura , a. 2013 , , 763 , 56 dekker , h. , delabre , b. , & dodorico , s.1986 , , 627 , 339 doyle , l. r. , carter , j. a. , fabrycky , d. c. , et al .
2011 , science , 333 , 1602 draine , b. t. , & lee , h. m. 1984 , , 285 , 89 draine , b. t. , & li , a. 2007 , , 657 , 810 evans , c. j. , howarth , i. d. , irwin , m. j. , burnley , a. w. , & harries , t. j. 2004 , , 353 , 601 evans , c. j. , lennon , d. j. , smartt , s. j. , & trundle , c. 2006 , , 456 , 623 evans , c. j. , & howarth , i. d.2008 , , 386 , 826 fabregat , j. , & torrejn , j. m. 2000 , , 357 , 451 fazio , g. g. , et al .
2004 , , 154 , 10 flower , p. j. 1996 , , 469 , 355 gaustad , j. e. , & van buren , d.1993 , , 105 , 1127 gordon , k. d. , et al .
2011 , , 142 , 102 gvaramadze , v. v. , pflamm - altenburg , j. , & kroupa , p. 2011
, , 525a , 17 herbig , g. h. , & simon , t. 2001 , , 121 , 3138 hernndez , j. , briceo , c. , calvet , n. , hartmann , l. , muzerolle , j. , & quintero , a. 2006 , , 652 , 472 hildebrand , r. h. 1983 , , 24 , 267 hillenbrand , l. a. , massey , p. , strom , s. e. , & merrill , k. m. 1993 , , 106 , 1906 hillenbrand , l. a. , strom , s. e. , vrba , f. j. , & keene , j. 1992 , , 397 , 613 hunter , i. , lennon , d. j. , dufton , p. l. , trundle , c. , simn - daz , s. , smartt , s. j. , ryans , r. s. i. , & evans , c. j. 2008 , , 479 , 541 ita , y. , et al . 2010 , , 62 , 273 kalas , p. , graham , j. r. , beckwith , s. v. w. , jewitt , d. c. , & lloyd , j. p. 2002
, , 567 , 999 kaper , l. , van loon , j. th .
, augusteijn , t. , goudfrooij , p. , patat , f. , waters , l. b. f. m. , & zijlstra , a. a. 1997 , , 475 , l37 keller , s. c. , & wood , p. r. 2006 , , 642 , 834 krivov , a. v. 2010 , research in astronomy and astrophysics , 10 , 383 leroy , a. , bolatto , a. , stanimirovi , s. , mizuno , n. , israel , f. , & bot , c. 2007 , , 658 , 1027 martayan , c. , frmat , y. , hubert , a .- m . , floquet , m. , zorec , j. , & neiner , c. 2007 , , 462 , 683 martnez - galarza , j. r. , kamp , i. , su , k. y. l. , gspr , a. , rieke , g. , & mamajek , e. e. 2009 , , 694 , 165 martins , f. , schaerer , d. , & hillier , d. j. 2005 , , 436 , 1049 massey , p. , puls , j. , pauldrach , a. w. a. , bresolin , f. , kudritzki , r. p. , & snow , t. 2005 , , 627 , 477 mcswain , m. v. , huang , w. , & gies , d. r. 2009 , , 700 , 1216 meyssonnier , n. , & azzopardi , m. 1993 , , 102 , 451 miroshnichenko , a. s. , & bjorkman , k. s. 2000 , iau colloq . 175 : the be phenomenon in early - type stars , 214 , 484 nieva , m. f. 2013 , , 550 , a26 oliveira , j. m. , et al . 2013 , , 428 , 3001 orosz , j. a. , welsh , w. f. , carter , j. a. , et al .
2012 , , 758 , 87 orosz , j. a. , welsh , w. f. , carter , j. a. , et al .
2012 , science , 337 , 1511 ostlie , d. a. , & carroll , b. w.1996 , an introduction to modern stellar astrophysics pagel , b. e. j. , edmunds , m. g. , fosbury , r. a. e. , & webster , b. l. 1978 , , 184 , 569 planck collaboration , abergel , a. , ade , p. a. r. , et al .
2011 , , 536 , a24 porter , j. m. , & rivinius , t. 2003 , , 115 , 1153 rieke , g. h. , & lebofsky , m. j.1985 , , 288 , 618 rieke , g. h. , et al .
2004 , , 154 , 25 sana , h. , de koter , a. , de mink , s. e. , et al .
2013 , , 550 , a107 sana , h. , de mink , s. e. , de koter , a. , et al .
2012 , science , 337 , 444 sandstrom , k. m. , bolatto , a. d. , draine , b. t. , bot , c. , & stanimirovi , s. 2010 , , 715 , 701 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 schwamb , m. e. , orosz , j. a. , carter , j. a. , et al . 2012 , arxiv:1210.3612 silaj , j. , jones , c. e. , tycner , c. , sigut , t. a. a. , & smith , a. d. 2010 , , 187 , 228 skrutskie , m. f. , et al . 2006 , aj , 131 , 1163 sloan , g. c. , et al .
2004 , , 614 , l77 smith , r. , & wyatt , m. c. 2010 , , 515a , 95 sota , a. , maz - apellniz , j. , walborn , n. r. , alfaro , e. j. , barb , r. h. , morrell , n. i. , gamen , r. c. , arias , & j. i. 2011 , , 193 , 24 stanimirovi , s. , staveley - smith , l. , dickey , j. m. , sault , r. j. , & snowden , s. l. 1999 , , 302 , 417 stanimirovi , s. , staveley - smith , l. , & jones , p. a. 2004 , , 604 , 176 staveley - smith , l. , sault , r. j. , hatzidimitriou , d. , kesteven , m. j. , & mcconnell , d. 1997 , , 289 , 225 su , k. y. l. , et al .
2006 , , 653 , 675 takeuchi , t. , & artymowicz , p.2001 , , 557 , 990 terebey , s. , & fich , m. 1986 , , 309 , l73 testor , g. 2001 , , 372 , 667 touhami , y. , et al .
2010 , , 122 , 379 udalski , a. , szymaski , m. , kubiak , m. , pietrzyski , g. , woniak , p. , & @xmath128ebru , k. 1998 , acta astronomica , 48 , 147 van buren , d. , & mccray , r. 1988 , , 329 , l93 van loon , j. th . ,
oliveira , j. m. , gordon , k. d. , sloan , g. c. , & engelbracht , c. w. 2010 , , 139 , 1553 waters , l. b. f. m. , & marlborough , j. m. 1994 , iau symposium 162 : pulsation ; rotation ; and mass loss in early - type stars ( kluwer ) , 399 waters , l. b. f. m. , & waelkens , c. 1998 , , 36 , 233 welsh , w. f. , orosz , j. a. , carter , j. a. , et al .
2012 , , 481 , 475 werner , m. w. , et al .
2004 , , 154 , 1 white , r. e. , & bally , j. 1993 , , 409 , 234 zaritsky , d. , harris , j. , thompson , i. b. , grebel , e. k. , & massey , p. 2002
, , 123 , 855 lccccccccc b001 & [email protected] & 10.90234 & @xmath2973.04078 & 2078.39 & 1585.23 & 918.67 & 1088.65 & 564.29 & 827.17 + b003 & [email protected] & 11.29644 & @xmath2973.35332 & 1909.53 & 1424.6 & 780.47 & 387.0 & 278.2 &
+ b004 & [email protected] & 11.40574 & @xmath2973.21014 & 3743.92 & 2976.41 & 1783.03 & 1008 & 573.7 & 391.1 + b005 & [email protected] & 11.50056 & @xmath2973.39015 & 1502.89 & 1180.57 & 715.75 & & & + b006 & [email protected] & 11.54496 & @xmath2973.42656 & 2684.89 & 2484.81 & 1688.72 & 1339 & 996.1 & 852.6 + b007 & [email protected] & 11.54884 & @xmath2973.39043 & 6530.19 & 5655.79 & 3742.47 & 2295.56 & & + b009 & [email protected] & 11.73266 & @xmath2973.08136 & 3625.15 & 2824.18 & 1596.46 & 768.6 & 523.2 & 261.5 + b010 & [email protected] & 11.73367 & @xmath2973.18190 & 8190.79 & 6787.24 & 4409.19 & 2667 & 1709 & 1199 + b011 & [email protected] & 11.73879 & @xmath2973.30242 & 1575.16 & 1221.5 & 682.27 & 353.91 & &
+ b013 & [email protected] & 11.78192 & @xmath2973.50227 & 5187.12 & 4007.68 & 2255.06 & 1099.74 & 749.38 &
+ b014 & [email protected] & 11.82905 & @xmath2973.11978 & 6590.62 & 5134.44 & 2956.37 & 1613 & 881.3 & 536.9 + b015 & [email protected] & 11.86142 & @xmath2973.22115 & 3905.94 & 3351.92 & 1933.56 & 1740 & 1549 & 1125 + b016 & [email protected] & 11.86473 & @xmath2973.28497 & 1945.03 & 1775.39 & 1073.39 & 477.85 & & + b019 & [email protected] & 11.91752 & @xmath2973.26351 & 2780.52 & 2214.59 & 1273.97 & 699.7 & 403.9 & 280.5 + b021 & [email protected] & 11.92833 & @xmath2973.00271 & 8800.88 & 7009.6 & 4080.92 & 2281 & 1324 & 849.4 + b022 & [email protected] & 11.94339 & @xmath2973.30276 & 2239.37 & 1744.59 & 988.91 & 518.2 & 280.5 &
+ b023 & [email protected] & 11.94806 & @xmath2973.29104 & 2592.54 & 2333.96 & 1620.16 & 977.4 & 618.7 & 433.2 + b024 & [email protected] & 11.94971 & @xmath2973.07532 & 2757.57 & 2144.34 & 1229.02 & 580.3 & 379.8 & 317.3 + b025 & [email protected] & 11.96681 & @xmath2972.86885 & 6393.31 & 5755.63 & 3495.9 & 1899 & 1146 & 700.0 + b026 & [email protected] & 11.96750 & @xmath2973.35610 & 3806.50 & 2976.41 & 1728.06 & 947.3 & 519.4 & 275.1 + b028 & [email protected] & 12.03507 & @xmath2973.24853 & 2906.2 & 6316.74 & 3064.49 & 5376 & 5601 & 7522 + b029 & [email protected] & 12.04036 & @xmath2973.40363 & 9676.67 & 7566.47 & 4141.51 & 2310 & 1308 & 792.0 + b032 & [email protected] & 12.07537 & @xmath2973.22022 & 2196.47 & 1689.25 & 939.2 & 492.6 & 262.5 & + b033 & [email protected] & 12.07881 & @xmath2973.45425 & 6806.53 & 5139.17 & 2852.06 & 1425 & 800.0 & 547.9 + b034 & [email protected] & 12.12513 & @xmath2973.30265 & 1872.95 & 1424.6 & 900.24 & 466.11 & 360 &
+ b035 & [email protected] & 12.16846 & @xmath2973.07830 & 3018.05 & 2662.52 & 1750.49 & 1133 & 687.9 & 434.0 + b036 & [email protected] & 12.17408 & @xmath2973.43757 & 2631.03 & 2575.69 & 1516.21 & 1272 & 1047 & 987.9 + b037 & [email protected] & 12.18000 & @xmath2973.19657 & 7239.79 & 5645.38 & 3084.31 & 1520 & 935.6 & 503.0 + b038 & [email protected] & 12.22087 & @xmath2973.23620 & 4818.66 & 4011.37 & 2713.68 & 1780 & 1253 & 1029 + b040 & [email protected] & 12.24565 & @xmath2973.23434 & 2295.76 & 2012.3 & 1234.69 & 790.1 & 482.5 & + b042 & [email protected] & 12.28289 & @xmath2973.15175 & 3180.72 & 2793.14 & 1618.67 & 988.3 & &
+ b043 & [email protected] & 12.29525 & @xmath2973.12213 & 5868.49 & 4717.3 & 2555.99 & 1273 & 786.9 & 467.7 + b044 & [email protected] & 12.31962 & @xmath2973.27545 & 5661.43 & 4410.55 & 2548.93 & 1344 & 791.9 & 521.8 + b045 & [email protected] & 12.32949 & @xmath2973.21312 & 2638.31 & 2198.33 & 1320.56 & 832.7 & 515.6 & 314.4 + b047 & [email protected] & 12.37908 & @xmath2973.16469 & 2255.93 & 1881.45 & 1102.45 & 636.9 & 346.0 & + b048 & [email protected] & 12.38313 & @xmath2972.85469 & 10827.45 & 8474.09 & 4906.35 & 2619 & 1574 & 903.5 + b049 & [email protected] & 12.42801 & @xmath2973.28839 & 7470.09 & 6161.59 & 4257.54 & & & 1264.72 + b051 & [email protected] & 12.43493 & @xmath2972.85057 & 1212.62 & 979.25 & 586.62 & & &
+ b052 & [email protected] & 12.43470 & @xmath2973.45601 & 1274.46 & 1115.05 & 733.77 & 437.8 & 263.0 &
+ b055 & [email protected] & 12.46821 & @xmath2972.85878 & 2336.29 & 1838.62 & 1099.41 & & & + b057 & [email protected] & 12.49123 & @xmath2972.86522 & 8752.38 & 5766.24 & 3861.52 & 2101 & 1216 & 735.8 + b059 & [email protected] & 12.50192 & @xmath2972.92010 & 8190.79 & 6133.28 & 3670.78 & 2172 & 1295 & 815.82 + b062 & [email protected] & 12.54577 & @xmath2973.16206 & 2399.53 & 1845.41 & 1167.24 & 712.0 & 454.1 & 326.2 + b064 & [email protected] & 12.54675 & @xmath2973.00723 & 4259.18 & 3302.88 & 2129.9 & 1190 & 691.0 & 468.1 + b065 & [email protected] & 12.56816 & @xmath2972.87751 & 1845.55 & 1444.42 & 875.71 & 477.0 & 306.7 & 255.6 + b066 & [email protected] & 12.64223 & @xmath2972.89116 & 3523.11 & 2630.83 & 1473.53 & 681.2 & 413.3 & 262.7 + b068 & [email protected] & 12.70665 & @xmath2972.71110 & 10678.89 & 7850.42 & 4454.09 & 2245 & 1317 & 824.0 + b069 & [email protected] & 12.72363 & @xmath2972.85454 & 1759.24 & 1352.99 & 792.79 & 505.9 & 253.2 & 244.3 + b070 & [email protected] & 12.74222 & @xmath2972.86143 & 3009.72 & 2194.29 & 1303.64 & 663.3 & 424.9 & 248.2 + b072 & [email protected] & 12.77149 & @xmath2972.80280 & 2561.69 & 1925.28 & 1086.32 & 501.7 & 279.2 & 168.4 + b073 & [email protected] & 12.77430 & @xmath2972.68264 & 7094.57 & 5177.18 & 2878.45 & 1719 & 933.0 & 632.6 + b075 & [email protected] & 12.82818 & @xmath2973.50450 & 4881.19 & 3834.36 & 2106.49 & 1037 & 538.4 & 356.7 + b076 & [email protected] & 12.88192 & @xmath2973.25806 & 2842.67 & 2101.32 & 1298.85 & & & + b077 & [email protected] & 12.90577 & @xmath2972.95379 & 8281.82 & 6469.83 & 3670.78 & 2010 & 1195 & 751.5 + b079 & [email protected] & 12.99207 & @xmath2973.34205 & 644.65 & 945.57 & 525.72 & 1320 & 1429 & 2060 + b080 & [email protected] & 13.00888 & @xmath2972.65836 & 6038.46 & 4165.75 & 2586.77 & 1339 & 745.3 & 541.4 + b081 & [email protected] & 13.02502 & @xmath2972.86922 & 1918.34 & 1536.36 & 1067.48 & 1106 & 874.8 & 823.2 + b082 & [email protected] & 13.02685 & @xmath2972.82968 & 5298.17 & 4204.29 & 2502.41 & & 1153 & 895.2 + b083 & [email protected] & 13.04448 & @xmath2972.68867 & 3497.24 & 2408.2 & 1561.56 & 989.2 & 560.7 & 346.4 + b085 & [email protected] & 13.18771 & @xmath2972.47890 & 4366.43 & 4015.07 & 2782.01 & 1894 & 1281 & 855.7 + b087 & [email protected] & 13.25087 & @xmath2972.67440 & 9014.18 & 6093.87 & 3518.51 & 1929 & 1137 & 766.9 + b088 & [email protected] & 13.28879 & @xmath2972.89203 & 6812.81 & 4989.91 & 2771.78 & 1358 & 771.8 & 493.8 + b089 & [email protected] & 13.29413 & @xmath2973.17733 & 4652.92 & 3699.07 & 2149.6 & 1130 & 615.3 & 413.4 + b091 & [email protected] & 13.31541 & @xmath2973.19116 & 5708.55 & 4542.48 & 2656.8 & 1374 & 823.9 & 441.7 + b092 & [email protected] & 13.42664 & @xmath2973.38894 & 2401.74 & 1881.45 & 1146.99 & 658.4 & 429.2 & 239.0 + b093 & [email protected] & 13.42678 & @xmath2972.65598 & 2206.61 & 1966.5 & 1357.56 & 1379.39 & 920.24 & 654.03 + b094 & [email protected] & 13.47266 & @xmath2972.80794 & 10503.31 & 8003.74 & 5128.12 & 3023 & 1773 & 1116 + b096 & [email protected] & 13.50825 & @xmath2972.70609 & 3461.99 & 2533.34 & 1495.4 & 777.8 & 468.9 & 308.7 + b098 & [email protected] & 13.62401 & @xmath2972.43538 & 3956.64 & 3296.8 & 2151.59 & 1258.02 & 678.42 & 629.79 + b100 & [email protected] & 13.72685 & @xmath2972.53598 & 2283.11 & 1777.02 & 954.9 & 478.3 & 318.5 &
+ b101 & [email protected] & 13.72946 & @xmath2972.42702 & 6657.73 & 5087.37 & 2766.68 & 1446 & 842.4 & 489.7 + b102 & [email protected] & 13.76366 & @xmath2972.92260 & 6145.06 & 4635.46 & 2445.45 & 1205.84 & 709.09 & 487.07 + b103 & [email protected] & 13.76883 & @xmath2972.77709 & 9784.22 & 8100.15 & 4541.07 & 2484.78 & 1521.6 & 981.74 + b104 & [email protected] & 13.78901 & @xmath2972.46571 & 9966.12 & 7996.37 & 4528.54 & 2607 & 1531 & 984.3 + b105 & [email protected] & 13.79445 & @xmath2972.57244 & 10245.33 & 8528.9 & 5185.12 & 3049 & 1862 & 1130 + b107 & [email protected] & 13.92170 & @xmath2972.38893 & 5645.81 & 4648.29 & 2454.47 & 2984.61 & 1963.81 & 1216.73 + b108 & [email protected] & 13.93584 & @xmath2972.26943 & 3139.97 & 2316.82 & 1206.59 & 666.3 & 377.7 & 282.0 + b109 & [email protected] & 13.94066 & @xmath2972.63821 & 1563.6 & 1189.3 & 653.97 & 422.75 & & + b112 & [email protected] & 14.04843 & @xmath2972.30674 & 7029.53 & 5527.05 & 2878.45 & 1383 & 814.1 & 474.2 + b115 & [email protected] & 14.07210 & @xmath2972.29132 & 3716.43 & 3110.95 & 1636.66 & 928.3 & 576.4 & 357.7 + b116 & [email protected] & 14.08366 & @xmath2972.45062 & 2260.09 & 1773.75 & 1003.6 & 491.7 & 318.2 & + b117 & [email protected] & 14.09807 & @xmath2972.35670 & 4770.08 & 4402.44 & 2252.99 & 1016 & 580.6 & 353.1 + b118 & [email protected] & 14.11296 & @xmath2972.42858 & & 2638.11 & 1499.54 & 864.8 & 521.8 & 319.7 + b119 & [email protected] & 14.18448 & @xmath2972.48521 & 6328.87 & 5557.67 & 3886.5 & 2922 & 2060 & 1465 + b122 & [email protected] & 14.23481 & @xmath2972.81855 & 5486.89 & 4282.46 & 2474.9 & 1185 & 739.1 & 389.0 + b123 & [email protected] & 14.30056 & @xmath2972.65976 & 5293.29 & 4239.29 & 2438.7 & 1265 & 711.0 & 426.1 + b124 & [email protected] & 14.31341 & @xmath2972.45953 & 4423.09 & 3348.83 & 1879.14 & 992.9 & 575.3 & 363.7 + b125 & [email protected] & 14.33005 & @xmath2972.52025 & 3641.88 & 2938.28 & 1626.14 & 949.9 & 501.5 & 298.1 + b126 & [email protected] & 14.40530 & @xmath2972.36509 & 8849.65 & 6843.73 & 3763.21 & 1959 & 1171 & 639.6 + b127 & [email protected] & 14.54937 & @xmath2972.59796 & 7779.04 & 6457.93 & 3704.75 & 1941 & 1147 & 697.5 + b128 & [email protected] & 14.55161 & @xmath2972.65967 & 6242.03 & 4840.53 & 2733.75 & 1421 & 817.9 & 525.7 + b129 & [email protected] & 14.55790 & @xmath2972.15542 & 3631.83 & 2911.34 & 1606.79 & 790.1 & 470.7 & 340.4 + b132 & [email protected] & 14.61457 & @xmath2972.19831 & 2827 & 2306.18 & 1413.7 & 795.2 & 492.8 & 297.8 + b135 & [email protected] & 14.65429 & @xmath2972.23331 & 4173.75 & 3312.02 & 1967.7 & 1130 & 733.0 & 410.7 + b136 & [email protected] & 14.72359 & @xmath2972.37521 & 1838.76 & 1508.32 & 904.4 & 527.8 & 411.1 & 405.8 + b137 & [email protected] & 14.74641 & @xmath2972.74281 & 9224.14 & 7239.24 & 4114.89 & 2061 & 1244 & 807.5 + b139 & [email protected] & 14.78591 & @xmath2972.44683 & 2239.37 & 1703.31 & 942.67 & 501.7 & 308.4 &
+ b140 & [email protected] & 14.80172 & @xmath2972.27163 & 1508.43 & 1215.89 & 713.11 & 679.3 & 461.2 & 379.8 + b141 & [email protected] & 14.80780 & @xmath2972.15759 & 1675.43 & 1560.61 & 851.84 & 1060.93 & 875.59 & 530.63 + b142 & [email protected] & 14.83629 & @xmath2972.28636 & 9014.18 & 7003.14 & 3940.57 & 1990 & 1264 & 794.9 + b144 & [email protected] & 14.87009 & @xmath2972.20877 & 4480.5 & 3395.42 & 1873.95 & 877.6 & 577.4 & 280.8 + b145 & [email protected] & 14.88242 & @xmath2972.22638 & 3433.41 & 2580.44 & 1400.74 & 708.7 & 487.4 & 353.8 + b148 & [email protected] & 15.22989 & @xmath2972.13463 & 6825.37 & 5082.68 & 2784.57 & 1474 & 823.9 & 537.9 + b152 & [email protected] & 15.57407 & @xmath2971.86197 & 5027.2 & 4048.49 & 2189.57 & 1214 & 705.8 & 405.1 + b154 & [email protected] & 15.67326 & @xmath2972.01004 & 2115.08 & 1686.14 & 910.25 & 458.87 & & + b156 & [email protected] & 15.68790 & @xmath2971.93674 & 6670 & 5603.93 & 3400.62 & 1981 & 1281 & 806.0 + b159 & [email protected] & 15.75023 & @xmath2972.46061 & 5172.8 & 3816.74 & 2068.04 & 1060 & 519.9 & 373.2 + b160 & [email protected] & 15.77017 & @xmath2971.89327 & 1500.12 & 1461.82 & 830.92 & 534.2 & 294.0 & 392.64 + b161 & [email protected] & 15.80126 & @xmath2972.07935 & & 2241.26 & 1222.25 & 620.1 & 391.8 &
+ b163 & [email protected] & 15.81567 & @xmath2972.04767 & 5101.83 & 3695.67 & 1884.34 & 960.5 & 563.8 & 335.1 + b165 & [email protected] & 15.82883 & @xmath2971.90371 & 5879.31 & 5096.75 & 3036.4 & 1845 & 1167 & 790.6 + b167 & [email protected] & 15.84393 & @xmath2972.10461 & 1219.34 & 1441.76 & 876.51 & 899.7 & 632.6 & 354.8 + b169 & [email protected] & 15.90086 & @xmath2972.18898 & 3938.46 & 2938.28 & 1594.99 & 775.7 & 440.9 & 260.1 + b172 & [email protected] & 15.91347 & @xmath2972.58976 & 1749.55 & 1385.78 & 812.75 & 471.3 & 375.9 & + b175 & [email protected] & 15.93491 & @xmath2971.99816 & 1311.37 & 1123.29 & 713.77 & 452.6 & 308.4 & + b177 & [email protected] & 15.98504 & @xmath2972.68832 & 2790.78 & 2029.05 & 1045.1 & 436.6 & 274.1 & + b179 & [email protected] & 16.08913 & @xmath2971.94511 & & 5851.85 & 3271.59 & 1694 & 952.1 & 631.4 + b182 & [email protected] & 16.22062 & @xmath2971.91369 & 12397.27 & 8357.82 & 5118.69 & 2564 & 1540 & 973.5 + b183 & [email protected] & 16.42175 & @xmath2972.06122 & 5031.83 & 3675.3 & 2217.98 & 1499 & 886.1 & 634.9 + b184 & [email protected] & 16.48946 & @xmath2972.19840 & 5050.4 & 3624.88 & 1955.05 & 969.37 & 564.81 & 418.02 + b188 & [email protected] & 17.05033 & @xmath2971.97593 & 4468.13 & 3278.63 & 1694.96 & 885.7 & 583.3 & 292.6 + b190 & [email protected] & 18.12382 & @xmath2973.29162 & 8632.29 & 4492.55 & 2794.85 & 4542 & 3332 & 2509 + b191 & [email protected] & 18.45350 & @xmath2973.30140 & 7916.37 & 6651.09 & 4245.79 & 1835.22 & 1034.43 & 917.9 + b192 & [email protected] & 18.80712 & @xmath2973.33425 & 8537.41 & 6369.32 & 3413.17 & 1636 & 997.9 & 585.5 + b193 & [email protected] & 19.19872 & @xmath2973.14434 & 5230.29 & 4262.78 & 2337.54 & 1228 & 757.0 & 462.5 + lcccccccccccc b001 & & & 444.40 & 2.93 & & & & & 198102.30 & 624.29 & & + b003 & 58.53 & 2.28 & 44.12 & 2.51 & & & & & 930.76 & 26.62 & & + b004 & 159.10 & 4.02 & 99.99 & 3.94 & 98.28 & 8.64 & & & 1684.34 & 26.40 & & + b005 & 189.40 & 7.51 & 168.70 & 6.12 & 382.80 & 23.00 & 666.20 & 33.63 & 4215.01 & 42.52 & & + b006 & 263.10 & 14.05 & 238.36 & 1.74 & 287.10 & 17.45 & 659.70 & 40.85 & 22090.02 & 111.60 & & + b007 & 575.10 & 9.76 & 412.60 & 9.30 & 323.90 & 14.23 & 198.40 & 12.74 & 1647.90 & 29.82 & &
+ b009 & 142.30 & 6.08 & 96.72 & 4.76 & & & 92.00 & 9.48 & 2500.48 & 31.08 & & + b010 & 661.70 & 14.72 & 500.50 & 11.84 & 478.50 & 18.28 & 584.80 & 31.21 & 7382.77 & 55.72 & & + b011 & 51.80 & 1.96 & 39.78 & 2.78 & & & & & 2087.20 & 28.72 & & + b013 & 167.60 & 3.61 & 110.60 & 3.62 & & & & & 1149.26 & 28.52 & & + b014 & 225.30 & 4.66 & 142.00 & 4.02 & 123.76 & 9.19 & 87.58 & 11.13 & 1948.75 & 28.46 & & + b015 & 485.20 & 14.90 & 319.70 & 12.28 & 238.50 & 15.34 & 158.24 & 14.35 & 4843.77 & 46.41 & & + b016 & 114.80 & 4.17 & 83.30 & 4.40 & & & 211.60 & 14.35 & 7047.57 & 52.57 & & + b019 & 118.30 & 3.81 & 92.86 & 3.76 & & & & & 2534.84 & 54.48 & & + b021 & 331.90 & 6.01 & 214.00 & 4.20 & 155.40 & 12.08 & 104.88 & 8.54 & 1093.05 & 27.72 & & + b022 & 79.59 & 2.59 & 67.77 & 2.90 & 93.73 & 9.06 & 186.10 & 11.38 & 6160.64 & 58.06 & &
+ b023 & 178.60 & 3.62 & 162.40 & 5.56 & 229.90 & 13.47 & 790.30 & 27.31 & 39214.47 & 204.24 & & + b024 & 88.11 & 2.55 & 58.20 & 1.84 & & & 55.02 & 9.16 & 1093.05 & 29.29 & & + b025 & 282.60 & 6.98 & 180.80 & 3.83 & 138.32 & 8.68 & 80.96 & 10.03 & 835.30 & 28.19 & &
+ b026 & 140.80 & 3.77 & 90.96 & 3.72 & & & & & 1002.48 & 27.70 & & + b028 & 9138.00 & 466.30 & 14030.00 & 513.70 & 37120.00 & 1022.00 & 100300.00 & 8321.00 & 1632288.00 & 1264.81 & 11107800.00 & 36590.40 + b029 & 319.80 & 8.72 & 186.60 & 5.31 & 178.20 & 13.72 & 125.12 & 13.98 & 5955.56 & 46.19 & & + b032 & 76.24 & 2.36 & 48.01 & 2.97 & & & & & 3525.87 & 91.61 & & + b033 & 208.50 & 4.38 & 133.10 & 3.82 & 119.21 & 9.19 & & & 1591.69 & 25.90 & & + b034 & 86.66 & 2.59 & 62.21 & 2.99 & & & & & 1414.72 & 31.51 & &
+ b035 & 204.30 & 4.58 & 128.90 & 3.61 & 75.53 & 8.78 & & & 721.31 & 60.49 & & + b036 & 1836.80 & 27.55 & 983.10 & 39.52 & 3013.00 & 104.30 & 7000.00 & 432.70 & 256606.50 & 591.81 & 1687950.00 & 8396.19 + b037 & 231.70 & 4.55 & 147.90 & 4.19 & 141.96 & 8.34 & & & 1380.37 & 28.68 & &
+ b038 & 619.30 & 10.64 & 498.10 & 13.19 & 402.40 & 12.17 & 274.60 & 11.75 & 726.10 & 28.66 & & + b040 & 133.80 & 3.67 & 84.04 & 4.17 & & & & & 729.74 & 55.17 & & + b042 & 161.60 & 4.59 & 115.30 & 4.05 & 152.88 & 8.34 & 298.70 & 17.57 & 7348.42 & 47.92 & & + b043 & 188.00 & 3.99 & 116.30 & 3.88 & 106.47 & 8.22 & 90.80 & 12.88 & 3146.94 & 32.16 & & + b044 & 208.90 & 4.30 & 134.20 & 3.80 & 102.83 & 6.18 & & & 751.91 & 64.57 & &
+ b045 & 135.70 & 2.91 & 87.31 & 3.59 & & & & & 1661.44 & 30.26 & & + b047 & 118.70 & 3.30 & 103.50 & 4.81 & 114.66 & 6.97 & 99.36 & 12.24 & 3603.94 & 37.27 & & + b048 & 391.10 & 6.55 & 251.50 & 5.06 & 173.60 & 17.48 & 157.00 & 11.22 & 2900.23 & 35.98 & & + b049 & 1217.00 & 25.07 & 922.20 & 19.78 & 664.50 & 17.57 & 478.80 & 14.78 & 3100.10 & 43.77 & &
+ b051 & 47.97 & 1.97 & 34.55 & 2.44 & & & & & 563.18 & 41.64 & & + b052 & 69.03 & 3.48 & 52.24 & 2.32 & & & & & 1175.29 & 24.27 & & + b055 & 90.66 & 2.93 & 68.99 & 2.85 & & & 180.80 & 15.15 & 12804.30 & 117.53 & & + b057 & 311.40 & 6.36 & 206.00 & 5.53 & 174.72 & 9.92 & 192.20 & 12.72 & 7842.89 & 75.97 & & + b059 & 339.20 & 9.42 & 215.70 & 4.84 & 197.10 & 18.82 & 105.80 & 12.33 & 2094.49 & 29.79 & & + b062 & 150.40 & 3.46 & 112.10 & 4.10 & 88.82 & 6.36 & 139.84 & 12.33 & 2556.70 & 32.02 & & + b064 & 203.10 & 4.37 & 127.60 & 3.50 & 111.93 & 8.63 & 45.17 & 11.22 & 1308.54 & 26.49 & & + b065 & 74.88 & 2.12 & 53.21 & 3.18 & & & & & 603.16 & 34.53 & & + b066 & 108.70 & 3.69 & 65.00 & 2.60 & 90.73 & 8.52 & 122.36 & 12.24 & 4538.76 & 46.96 & & + b068 & 428.00 & 7.25 & 321.90 & 8.50 & 262.60 & 11.26 & 202.10 & 11.29 & 1424.09 & 25.89 & & + b069 & 70.29 & 2.74 & 56.89 & 2.86 & & & & & 802.61 & 58.30 & & + b070 & 102.00 & 3.16 & 73.00 & 3.23 & & & 71.12 & 12.24 & 1229.42 & 29.92 & &
+ b072 & 77.25 & 2.10 & 55.66 & 2.23 & & & 101.20 & 12.24 & 2370.36 & 29.36 & & + b073 & 218.70 & 5.25 & 133.20 & 4.86 & 114.66 & 6.50 & 97.52 & 11.59 & 5202.92 & 33.44 & & + b075 & 168.20 & 5.22 & 108.40 & 4.76 & & & & & 10368.36 & 56.60 & &
+ b076 & 96.57 & 3.35 & 66.32 & 2.62 & & & & & 593.37 & 33.31 & & + b077 & 287.90 & 8.23 & 180.10 & 4.98 & & & & & 1106.58 & 24.86 & & + b079 & 3343.00 & 83.64 & 4702.00 & 69.86 & 8568.00 & 107.40 & 23370.00 & 213.00 & 123670.80 & 436.18 & & + b080 & 199.70 & 3.73 & 129.90 & 4.72 & 109.20 & 5.93 & 156.80 & 16.59 & 6383.41 & 35.59 & & + b081 & 445.60 & 8.69 & 373.80 & 7.63 & 302.10 & 11.44 & 236.90 & 9.44 & 870.07 & 28.67 & & + b082 & 173.70 & 3.95 & 119.20 & 4.18 & 98.28 & 8.72 & 197.60 & 21.12 & 6613.47 & 69.29 & & + b083 & 135.20 & 3.08 & 96.48 & 4.87 & 107.38 & 5.76 & & & 854.04 & 34.96 & &
+ b085 & 469.30 & 8.46 & 337.70 & 13.86 & 230.20 & 11.22 & 169.70 & 28.80 & 693.51 & 27.76 & & + b087 & 308.50 & 5.90 & 199.00 & 5.07 & 160.16 & 7.50 & 120.52 & 9.20 & 3015.78 & 30.47 & & + b088 & 204.30 & 4.36 & 124.70 & 3.38 & 91.00 & 9.28 & 74.52 & 11.22 & 1408.47 & 27.68 & &
+ b089 & 179.50 & 4.20 & 115.00 & 2.91 & 84.45 & 8.91 & 50.97 & 9.08 & 768.26 & 32.27 & & + b091 & 214.90 & 5.42 & 136.60 & 3.53 & 118.30 & 9.28 & 67.90 & 10.58 & 430.97 & 33.31 & & + b092 & 115.40 & 3.08 & 72.07 & 3.09 & 79.81 & 8.00 & & & 687.06 & 49.97 & & + b093 & 417.76 & 3.89 & 281.79 & 1.27 & 535.99 & 10.56 & & & 51925.08 & 181.97 & & + b094 & 493.80 & 8.59 & 363.70 & 7.00 & 273.30 & 11.54 & 263.20 & 9.89 & 4334.72 & 46.56 & & + b096 & 116.10 & 2.82 & 76.57 & 2.27 & 68.43 & 6.29 & 58.60 & 8.72 & 796.16 & 25.82 & & + b098 & 227.20 & 5.41 & 152.40 & 4.26 & & & 94.76 & 13.34 & 582.02 & 22.55 & & + b100 & 70.68 & 2.20 & 47.07 & 2.25 & & & & & 1319.99 & 26.75 & & + b101 & 210.60 & 4.30 & 133.70 & 4.17 & 115.57 & 9.10 & & & 727.35 & 26.19 & & + b102 & 194.80 & 8.24 & 117.40 & 4.12 & & & & & 1660.39 & 25.78 & & + b103 & 486.08 & 4.87 & 284.82 & 1.16 & 264.81 & 9.74 & 159.16 & 9.38 & 870.28 & 32.01 & & + b104 & 427.40 & 8.06 & 280.60 & 6.77 & 209.60 & 11.90 & 168.90 & 11.31 & 655.83 & 32.27 & &
+ b105 & 482.60 & 8.98 & 345.80 & 7.71 & 211.60 & 12.35 & 136.60 & 13.50 & 2679.53 & 29.50 & 72527.40 & 10291.05 + b107 & 240.70 & 4.89 & 183.90 & 4.96 & 188.37 & 8.09 & 225.10 & 14.99 & 1967.49 & 29.69 & 48787.20 & 2308.68 + b108 & 99.76 & 3.96 & 65.54 & 9.10 & & & & & 7553.50 & 53.29 & & + b109 & 52.32 & 1.93 & 32.05 & 2.12 & & & & & 1467.81 & 24.17 & & + b112 & 202.80 & 3.69 & 124.30 & 5.38 & 99.19 & 7.28 & 101.20 & 16.28 & 1660.39 & 28.11 & & + b115 & 176.80 & 5.20 & 117.90 & 5.42 & 106.47 & 6.99 & 94.76 & 19.14 & 4382.61 & 36.25 & & + b116 & 86.94 & 2.53 & 56.00 & 2.42 & & & & & 638.55 & 34.38 & &
+ b117 & 146.20 & 2.59 & 86.68 & 2.69 & 66.52 & 6.03 & & & 501.76 & 34.35 & & + b118 & 148.90 & 3.87 & 101.70 & 4.08 & 91.91 & 8.64 & & & 1270.02 & 25.91 & & + b119 & 600.70 & 13.71 & 435.50 & 8.94 & 352.80 & 17.41 & 245.40 & 11.80 & 901.51 & 30.18 & & + b122 & 207.50 & 4.53 & 131.20 & 4.25 & 136.50 & 8.47 & & & 1361.63 & 23.57 & & + b123 & 195.60 & 4.59 & 132.60 & 5.01 & 127.40 & 8.65 & & & 766.28 & 24.85 & & + b124 & 135.60 & 3.29 & 87.41 & 3.56 & 74.44 & 8.04 & & & 624.60 & 34.35 & & + b125 & 143.80 & 3.85 & 98.06 & 5.69 & & & & & 1465.73 & 29.20 & & + b126 & 297.10 & 5.27 & 192.60 & 4.27 & 140.14 & 6.91 & 81.70 & 10.76 & 661.03 & 31.23 & & + b127 & 628.00 & 10.72 & 447.10 & 8.66 & 360.40 & 14.61 & 263.40 & 9.61 & 1441.78 & 27.39 & & + b128 & 206.20 & 5.18 & 135.30 & 4.07 & 144.69 & 9.83 & 107.64 & 12.97 & 2709.72 & 33.50 & & + b129 & 123.20 & 3.59 & 83.92 & 2.78 & 99.19 & 9.83 & 132.10 & 15.81 & 2466.13 & 28.99 & & + b132 & 135.20 & 5.09 & 96.28 & 4.71 & 104.65 & 9.83 & 190.60 & 20.00 & 12814.71 & 87.02 & &
+ b135 & 189.30 & 4.48 & 118.40 & 3.75 & 131.04 & 9.37 & 88.96 & 12.33 & 2993.92 & 34.81 & & + b136 & 228.80 & 10.10 & 164.80 & 7.40 & 143.78 & 7.60 & 81.42 & 11.13 & 579.84 & 31.23 & & + b137 & 319.90 & 8.71 & 206.80 & 5.76 & 158.10 & 15.93 & 87.40 & 12.33 & 754.93 & 25.00 & & + b139 & 65.09 & 2.06 & 50.48 & 2.37 & 58.24 & 6.93 & 52.26 & 9.57 & 1522.98 & 26.40 & & + b140 & 211.40 & 5.08 & 161.50 & 4.86 & 137.41 & 9.83 & 86.02 & 12.70 & 675.40 & 62.44 & &
+ b141 & & & 730.23 & 1.52 & & & 9752.00 & 49.40 & 136371.00 & 514.46 & & + b142 & 304.40 & 6.22 & 193.40 & 5.10 & 132.30 & 13.08 & 86.11 & 12.70 & 2992.88 & 31.77 & & + b144 & 145.80 & 3.42 & 87.71 & 3.58 & 82.90 & 8.60 & 102.12 & 15.18 & 2608.75 & 35.49 & & + b145 & 103.20 & 2.75 & 71.58 & 3.16 & 100.10 & 8.22 & 189.20 & 11.01 & 3907.91 & 49.22 & & + b148 & 212.30 & 3.71 & 139.20 & 3.62 & 126.49 & 9.46 & 85.38 & 14.08 & 927.84 & 27.39 & & + b152 & 174.90 & 3.71 & 110.30 & 4.02 & 68.80 & 8.79 & & & 2816.95 & 29.05 & &
+ b154 & 101.60 & 2.56 & 81.33 & 3.06 & 73.35 & 6.89 & & & 1415.76 & 29.00 & & + b156 & 342.40 & 8.35 & 210.70 & 5.29 & 182.00 & 11.38 & 124.80 & 17.24 & 2463.01 & 32.10 & &
+ b159 & 147.20 & 3.55 & 96.16 & 2.31 & 104.65 & 9.19 & 55.02 & 8.81 & 1303.33 & 27.99 & & + b160 & 87.84 & 2.79 & 67.16 & 2.60 & & & 240.90 & 12.10 & 9942.59 & 61.82 & & + b161 & 87.86 & 1.97 & 56.74 & 2.17 & 64.97 & 6.17 & & & 2092.41 & 35.31 & & + b163 & 139.50 & 2.98 & 94.13 & 2.71 & 118.30 & 7.46 & 159.70 & 10.80 & 5640.14 & 45.03 & & + b165 & 310.70 & 5.42 & 203.70 & 4.86 & 173.81 & 8.41 & 104.88 & 12.24 & 1355.38 & 27.62 & & + b167 & 183.40 & 4.12 & 118.60 & 3.13 & & & & & 1087.85 & 55.17 & & + b169 & 118.40 & 4.67 & 79.94 & 2.77 & & & & & 1088.89 & 26.06 & & + b172 & 98.63 & 4.45 & 81.56 & 4.29 & & & & & 537.16 & 33.31 & & + b175 & 88.11 & 2.95 & 65.13 & 3.51 & & & & & 1009.46 & 26.55 & 75141.00 & 3016.53 + b177 & 66.51 & 2.35 & 49.52 & 2.59 & & & & & 1630.21 & 40.73 & & + b179 & 257.40 & 4.99 & 154.90 & 4.56 & 148.33 & 8.78 & & & 2327.68 & 29.01 & 41055.30 & 2123.55 + b182 & 378.10 & 7.24 & 246.80 & 5.58 & 173.30 & 12.55 & 125.12 & 13.06 & 1099.30 & 23.78 & & + b183 & 239.40 & 5.89 & 167.00 & 6.22 & 181.09 & 10.83 & 280.40 & 13.41 & 12679.38 & 168.64 & & + b184 & 141.30 & 6.19 & 93.57 & 4.53 & & & & & 711.11 & 36.93 & & + b188 & 131.60 & 5.57 & 80.09 & 3.15 & & & & & 3050.13 & 29.75 & & + b190 & 868.70 & 15.06 & 714.40 & 11.39 & 591.50 & 17.26 & 451.40 & 12.28 & 617.21 & 26.92 & &
+ b191 & & & 872.64 & 2.21 & & & 5805.20 & 38.00 & 135121.80 & 689.04 & & + b192 & 251.90 & 5.55 & 160.90 & 4.36 & 142.87 & 12.19 & & & 960.74 & 25.23 & & + b193 & 183.90 & 4.07 & 122.80 & 3.47 & & & & & 1018.72 & 26.20 & &
+ lcrrrccl o6 & 39 & 12.2 & 30.5 & 309.0 & @xmath292.91 & @xmath293.72 & 4686 , 4541 , + 4026 @xmath12 4200 + o7 & 37 & 11.3 & 27.3 & 213.3 & @xmath292.91 & @xmath293.54 & 4686 , 4200 , 4541 @xmath12 4471 + o8.5 & 34 & 9.9 & 22.8 & 118.4 & @xmath292.91 & @xmath293.27 & 4686 , 4200 , 4541 @xmath12 4387 + o9 & 33 & 9.5 & 21.5 & 96.3 & @xmath292.91 & @xmath293.18 & 4686 , 4200 @xmath12 4143 + b0 & 30 & 8.3 & 17.8 & 50.3 & @xmath292.89 & @xmath292.98 & 4686 , 4541 , weak 4200 + b0.5 & 27 & 7.2 & 14.5 & 24.8 & @xmath292.86 & @xmath292.60 & weak 4686 , no 4200 , 4541 + b1/1.5 & 25 & 6.5 & 12.5 & 15.0 & @xmath292.84 & @xmath292.41 & 4088 , 4116 + b2/b3 & 22 & 5.6 & 10.0 & 6.6 & @xmath292.79 & @xmath292.11 & 4553 + b5/b8 & 15 & 3.8 & 5.4 & 0.6 & @xmath292.67 & @xmath291.24 & 4128/4132 , no 4553 + b9 & 10 & 2.8 & 3.2 & 0.1 & @xmath292.52 & @xmath290.30 & 4128/4132 , 4481 @xmath27 4471 + lccclcl b001 & em & & & & & planetary nebula + b003 & b0b2 & & & & & + b004 & o9 & & & & & + b005 & & & & & & eclipsing binary + b006 & & & & & & + b007 & obe & & & & & + b009 & b0b2 & & & & & + b010 & obe & & & & & emission - line star + b011 & & & & & & + b013 & b0 & & & & & + b014 & b0b2 & & & & & + b015 & obe ? & & & & & + b016 & obe & & & & & + b019 & b0b2 & & & & & + b021 & b0 & b12ii & @xmath129 & eh08 & 136 & + b022 & b0 & & & & & + b023 & o7 & & & & & + b024 & b0b2 & & & & & eclipsing binary + b025 & b0b2 & & & & & + b026 & b0b2 & & & & & + b028 & em & & & & & compact region + b029 & b0 & & & & & + b032 & b1 & & & & & + b033 & o7o9 & & & & & + b034 & b0 & & & & & + b035 & obe ? & & & & & o6iaf ( see section 3 ) + b036 & em & b0v ? & & s & & young stellar object + b037 & b1 & & & & & + b038 & obe & & & & & + b040 & b0b2 & & & & & + b042 & b0 & & & & & + b043 & o9 & & & & & + b044 & b0 & & & & & + b045 & b1 & & & & & + b047 & o7 & & & & & + b048 & b2 & b0.5iv & @xmath130 & eh08 & 119 , 141 & + b049 & obe & & & & & + b051 & b0 & & & & & + b052 & b0 & & & & & + b055 & obe & & & & & emission - line star + b057 & b0b2 & b2 & & s & & + b059 & o9 & b3 & & s & & + b062 & obe ? & & & & & + b064 & b0 & b13iii
v(e ) & & c11 & & x - ray binary + b065 & b1 & & & & & + b066 & o9 & & & & & + b068 & obe & b2 ? & & s & & emission - line star + b069 & o7o9 & & & & & + b070 & b0b2 & & & & & + b072 & o8.5 & & & & & + b073 & b1 & & & & & eclipsing binary + b075 & b0b2 & & & & & eclipsing binary + b076 & b1 & & & & & + b077 & b0b2 & & & & & + b079 & em & & & & & planetary nebula + b080 & o8.5 & & & & & emission - line star + b081 & obe & & & & & emission - line star + b082 & b1 & & & & & emission - line star + b083 & obe ? & b15iii & @xmath131 & eh08 & 125 , 159 & + b085 & obe & o9b0e & & a09 & & x - ray binary + b087 & b0 & & & & & + b088 & o7 & & & & & + b089 & b1 & & & & & + b091 & b0b2 & & & & & + b092 & b0 & & & & & eclipsing binary + b093 & & & & & & eclipsing binary + b094 & em & b2 ? & & s & & emission - line star + b096 & b0 & & & & & eclipsing binary + b098 & b1 & & & & & + b100 & b1 & b0.5v & @xmath132 & h08 , e06 & 129 , 160180 & + b101 & b5 & b15ii & @xmath133 & eh08 & 134 , 167 & + b102 & o9 & & & & & + b103 & obe & b1ii & & s & & + b104 & obe & b13ii & @xmath134 & eh08 & 128 , 163 & + b105 & b2 & b3 ? & & s & & + b107 & b1 & & & & & emission - line star + b108 & & b2iv & @xmath135 & m07 & 137 , 165 & + b109 & b0b2 & & & & & + b112 & & & & & & eclipsing binary + b115 & o7o9 & o9v & @xmath136 & m07 & 136 , 164 & + b116 & obe & b2iii(ive ? ) & 151 & h08 , s , e06 & 130 , 170 & + b117 & b0b2 & b0.5v & & e06 & & spectroscopic binary + b118 & b0 & b2ii & 130 & h08 , e06 & 130 , 166 & + b119 & obe & b0iii & @xmath137 , 159 & m07 , e06 & 129 , 167 & + b122 & o7o9 & & & & & eclipsing binary ?
+ b123 & b0b2 & & & & & + b124 & b0 & b1v , b12iii & @xmath138 & h08 , eh08 & 131 , 168 & + b125 & b1 & & & & & + b126 & b1 & b0iv , b0.5iv & @xmath139 , @xmath140 & m07 , eh08 & 131 , 160180 & + b127 & obe & & & & & + b128 & o9 & & & & & + b129 & b0b2 & b0v & 172 & h08 , e06 & 166 & + b132 & obe ? & b0.5v & 176 & h08 , e06 & 180 & emission - line star + b135 & o7o9 & & & & & + b136 & b0 & & & & & + b137 & b1 & b13ii & @xmath141 & eh08 & 130 , 146 , 177 & eclipsing binary + b139 & o9 & & & & & + b140 & obe & & & & & + b141 & em & b0v & & h08 & & emission - line star + b142 & o8.5 & o7iiin((f ) ) & 208 & h08 , e06 & 131 , 173 & + b144 & o7o9 & & & & & x - ray source within 5 + b145 & o7o9 & o7vn & 250 & h08 , e06 & 130 , 175 & + b148 & b0 & & & & & + b152 & o7o9 & & & & & + b154 & b1 & & & & & + b156 & o7o9 & b1 & & s & & + b159 & b0 & & & & & + b160 & o7o9 & & & & & emission object within 5 + b161 & b0 & & & & & + b163 & o7o9 & & & & & + b165 & obe ? & & & & & + b167 & a0 & & & & & + b169 & o7o9 & & & & & + b172 & b0b2 & & & & & + b175 & & & & & & + b177 & o7o9 & o9v & @xmath142 & eh08 & 131 , 185 & + b179 & b0 & & & & & + b182 & b0 & & & & & + b183 & o7o9 & & & & & emission object at 3 + b184 & b0b2 & b0.5v & @xmath143 & eh08 & 144 , 187 & eclipsing binary + b188 & obe ? & & & & & + b190 & obe & & & & & emission - line star at 2 + b191 & & be ? &
& s & & + b192 & b0b2 & b13ii & @xmath144 & eh08 & 144 , 179 & + b193 & b0 & & & & & + lccc@cc@cc@c b001 & 340@xmath14513 & [email protected] & & & 198098@xmath145624 & 41436@xmath1455181 & 10.86 + b003 & 7@xmath1455 & [email protected] & & & 929@xmath14527 & 547@xmath14564 & + b004 & 12@xmath1455 & [email protected] & & & 1681@xmath14526 & 467@xmath14519 & 6.49 + b005 & 137@xmath1457 & [email protected] & 655@xmath14534 & [email protected] & 4214@xmath14543 & 2935@xmath145318 & + b006 & 110@xmath14513 & [email protected] & 615@xmath14541 & [email protected] & 22084@xmath145112 & 3756@xmath145387 & 8.44 + b007 & 193@xmath14526 & [email protected] & 122@xmath14515 & [email protected] & 1638@xmath14530 & 162@xmath14518 & + b009 & 23@xmath1459 & [email protected] & 67@xmath14510 & [email protected] & 2497@xmath14531 & 740@xmath14580 & 7.36 + b010 & 246@xmath14529 & [email protected] & 496@xmath14533 & [email protected] & 7371@xmath14556 & 629@xmath14565 & 6.88 + b011 & 6@xmath1456 & [email protected] & & & 2086@xmath14529 & 1342@xmath145231 & + b013 & 12@xmath14512 & [email protected] & & & 1145@xmath14529 & 275@xmath14533 & + b014 & @xmath14616 & @xmath1470.10 & 34@xmath14512 & [email protected] & 1942@xmath14528 & 274@xmath14528 & 6.31 + b015 & 153@xmath14521 & [email protected] & 101@xmath14516 & [email protected] & 4836@xmath14546 & 633@xmath14566 & 6.49 + b016 & 38@xmath1458 & [email protected] & 196@xmath14515 & [email protected] & 7045@xmath14553 & 3357@xmath145488 & + b019 & 26@xmath1458 & [email protected] & & & 2532@xmath14554 & 824@xmath14590 & 7.30 + b021 & 10@xmath14521 & [email protected] & 36@xmath14511 & [email protected] & 1084@xmath14528 & 126@xmath14513 & 5.18 + b022 & 21@xmath1456 & [email protected] & 170@xmath14512 & [email protected] & 6159@xmath14558 & 3139@xmath145349 & + b023 & 77@xmath1457 & [email protected] & 762@xmath14527 & [email protected] & 39211@xmath145204 & 11244@xmath145481 & 9.80 + b024 & 3@xmath1456 & [email protected] & 36@xmath1459 & [email protected] & 1091@xmath14529 & 428@xmath14544 & 6.25 + b025 & @xmath14819 & @xmath1490.10 & 18@xmath14512 & [email protected] & 827@xmath14528 & 99@xmath14511 & 5.10 + b026 & 0@xmath14510 & [email protected] & & & 998@xmath14528 & 240@xmath14525 & 6.31 + b028 & 13516@xmath145517 & [email protected] & 100122@xmath1458321 & [email protected] & 1632264@xmath1451265 & 69138@xmath1457623 & 10.75 + b029 & @xmath15022 & @xmath1510.10 & 55@xmath14516 & [email protected] & 5947@xmath14546 & 680@xmath14571 & 7.10 + b032 & 1@xmath1456 & [email protected] & & & 3524@xmath14592 & 1629@xmath145184 & + b033 & 9@xmath1456 & [email protected] & & & 1587@xmath14526 & 312@xmath14511 & 6.07 + b034 & 20@xmath1457 & [email protected] & & & 1413@xmath14532 & 801@xmath145123 & + b035 & 21@xmath14513 & [email protected] & & & 716@xmath14560 & 144@xmath14521 & 5.46 + b036 & 862@xmath14542 & [email protected] & 6958@xmath145433 & [email protected] & 256601@xmath145592 & 45936@xmath1454839 & 10.95 + b037 & 3@xmath14516 & [email protected] & & & 1374@xmath14529 & 206@xmath14522 & 6.00 + b038 & 328@xmath14522 & [email protected] & 216@xmath14513 & [email protected] & 718@xmath14529 & 92@xmath14510 & 4.53 + b040 & 9@xmath1459 & [email protected] & & & 726@xmath14555 & 209@xmath14528 & + b042 & 27@xmath14511 & [email protected] & 269@xmath14518 & [email protected] & 7345@xmath14548 & 1963@xmath145236 & + b043 & 5@xmath1455 & [email protected] & 54@xmath14513 & [email protected] & 3142@xmath14532 & 692@xmath14525 & 6.98 + b044 & 14@xmath14513 & [email protected] & & & 747@xmath14565 & 147@xmath14520 & 5.31 + b045 & 8@xmath1459 & [email protected] & & & 1658@xmath14530 & 453@xmath14551 & 6.72 + b047 & 48@xmath1456 & [email protected] & 81@xmath14512 & [email protected] & 3602@xmath14537 & 1585@xmath14590 & + b048 & @xmath15228 & @xmath1530.10 & 62@xmath14515 & [email protected] & 2887@xmath14536 & 216@xmath14522 & 6.17 + b049 & 736@xmath14528 & [email protected] & 414@xmath14516 & [email protected] & 3092@xmath14544 & 362@xmath14540 & 5.88 + b051 & 11@xmath1453 & [email protected] & & & 562@xmath14542 & 567@xmath14571 & + b052 & 13@xmath1455 & [email protected] & & & 1174@xmath14524 & 708@xmath14580 &
+ b055 & 21@xmath1456 & [email protected] & 164@xmath14515 & [email protected] & 12802@xmath145118 & 5806@xmath145590 & + b057 & 5@xmath14521 & [email protected] & 123@xmath14515 & [email protected] & 7834@xmath14576 & 849@xmath14587 & 7.48 + b059 & 26@xmath1458 & [email protected] & 43@xmath14513 & [email protected] & 2087@xmath14530 & 269@xmath14511 & 5.93 + b062 & 44@xmath1458 & [email protected] & 116@xmath14513 & [email protected] & 2554@xmath14532 & 817@xmath14588 & 7.14 + b064 & 21@xmath14512 & [email protected] & 9@xmath14512 & [email protected] & 1304@xmath14526 & 289@xmath14531 & 6.02 + b065 & 8@xmath1456 & [email protected] & & & 601@xmath14535 & 287@xmath14536 & 5.84 + b066 & 6@xmath1454 & [email protected] & 103@xmath14512 & [email protected] & 4536@xmath14547 & 1866@xmath14586 & 8.00 + b068 & 107@xmath14524 & [email protected] & 128@xmath14514 & [email protected] & 1414@xmath14526 & 143@xmath14515 & 5.50 + b069 & 13@xmath1454 & [email protected] & & & 801@xmath14558 & 444@xmath14539 & 6.20 + b070 & 10@xmath1459 & [email protected] & 49@xmath14513 & [email protected] & 1227@xmath14530 & 421@xmath14560 & 6.65 + b072 & 12@xmath1453 & [email protected] & 87@xmath14512 & [email protected] & 2369@xmath14529 & 1323@xmath14567 & 7.78 + b073 & @xmath15418 & @xmath1550.09 & 41@xmath14513 & [email protected] & 5195@xmath14533 & 688@xmath14573 & 7.20 + b075 & 9@xmath14511 & [email protected] & & & 10364@xmath14557 & 2276@xmath145238 & 8.57 + b076 & 10@xmath1456 & [email protected] & & & 591@xmath14533 & 227@xmath14526 & + b077 & @xmath14620 & @xmath1530.10 & & & 1098@xmath14525 & 124@xmath14513 & 5.33 + b079 & 4576@xmath14571 & [email protected] & 23326@xmath145213 & [email protected] & 123665@xmath145436 & 21333@xmath1452220 & 9.35 + b080 & 13@xmath1456 & [email protected] & 118@xmath14517 & [email protected] & 6379@xmath14536 & 1335@xmath14544 & 7.59 + b081 & 268@xmath14513 & [email protected] & 200@xmath14510 & [email protected] & 865@xmath14529 & 178@xmath14519 & 4.97 + b082 & 10@xmath14514 & [email protected] & 160@xmath14522 & [email protected] & 6608@xmath14569 & 1317@xmath145162 & 7.08 + b083 & 2@xmath14511 & [email protected] & & & 850@xmath14535 & 196@xmath14522 & 5.89 + b085 & 157@xmath14523 & [email protected] & 107@xmath14530 & [email protected] & 685@xmath14528 & 82@xmath1459 & 4.68 + b087 & 26@xmath14518 & [email protected] & 62@xmath14511 & [email protected] & 3008@xmath14530 & 412@xmath14543 & 6.40 + b088 & 6@xmath1455 & [email protected] & 35@xmath14511 & [email protected] & 1404@xmath14528 & 290@xmath14511 & 6.05 + b089 & 7@xmath14511 & [email protected] & 14@xmath14510 & [email protected] & 763@xmath14532 & 154@xmath14517 & 5.58 + b091 & 5@xmath14514 & [email protected] & 22@xmath14512 & [email protected] & 425@xmath14533 & 70@xmath1459 & 4.88 + b092 & 13@xmath1457 & [email protected] & & & 685@xmath14550 & 275@xmath14535 & 6.06 + b093 & 150@xmath14516 & [email protected] & & & 51919@xmath145182 & 8571@xmath1451054 & 9.66 + b094 & 75@xmath14530 & [email protected] & 163@xmath14514 & [email protected] & 4321@xmath14547 & 326@xmath14534 & 6.38 + b096 & 7@xmath1458 & [email protected] & 35@xmath1459 & [email protected] & 793@xmath14526 & 269@xmath14530 & 5.94 + b098 & 32@xmath14514 & [email protected] & 53@xmath14514 & [email protected] & 576@xmath14523 & 104@xmath14512 & 4.82 + b100 & 1@xmath1456 & [email protected] & & & 1318@xmath14527 & 627@xmath14571 & + b101 & @xmath15618 & @xmath1570.08 & & & 717@xmath14526 & 68@xmath1458 & 5.34 + b102 & 12@xmath1457 & [email protected] & & & 1656@xmath14526 & 385@xmath14521 & 6.24 + b103 & 47@xmath14525 & [email protected] & 77@xmath14513 & [email protected] & 859@xmath14532 & 79@xmath1459 & 4.78 + b104 & 31@xmath14526 & [email protected] & 82@xmath14514 & [email protected] & 644@xmath14532 & 56@xmath1456 & 4.47 + b105 & 35@xmath14533 & [email protected] & 26@xmath14518 & [email protected] & 2664@xmath14530 & 172@xmath14518 & 5.85 + b107 & @xmath15830 & @xmath1590.07 & 126@xmath14518 & [email protected] & 1954@xmath14530 & 149@xmath14516 & 5.43 + b108 & 2@xmath14511 & [email protected] & & & 7551@xmath14553 & 2580@xmath145280 & 8.48 + b109 & @xmath1606 & @xmath1610.13 & & & 1466@xmath14524 & 790@xmath145117 & + b112 & @xmath16015 & @xmath1530.11 & 55@xmath14517 & [email protected] & 1654@xmath14528 & 272@xmath14530 & 6.27 + b115 & 37@xmath1456 & [email protected] & 68@xmath14519 & [email protected] & 4379@xmath14536 & 1322@xmath14551 & 7.63 + b116 & 9@xmath1456 & [email protected] & & & 636@xmath14534 & 295@xmath14536 & + b117 & @xmath16210 & @xmath1630.10 & & & 497@xmath14534 & 111@xmath14514 & 5.29 + b118 & 24@xmath1459 & [email protected] & & & 1267@xmath14526 & 387@xmath14541 & 6.41 + b119 & 156@xmath14530 & [email protected] & 149@xmath14515 & [email protected] & 889@xmath14530 & 69@xmath1458 & 4.38 + b122 & 28@xmath1456 & [email protected] & & & 1357@xmath14524 & 321@xmath14514 & 6.27 + b123 & 12@xmath14514 & [email protected] & & & 761@xmath14525 & 137@xmath14515 & 5.55 + b124 & @xmath16410 & @xmath1650.11 & & & 621@xmath14534 & 165@xmath14520 & 5.50 + b125 & 7@xmath14511 & [email protected] & & & 1462@xmath14529 & 350@xmath14537 & 6.64 + b126 & 5@xmath14520 & [email protected] & 17@xmath14513 & [email protected] & 652@xmath14531 & 76@xmath1459 & 4.94 + b127 & 262@xmath14521 & [email protected] & 199@xmath14512 & [email protected] & 1433@xmath14527 & 168@xmath14518 & 5.70 + b128 & 11@xmath1456 & [email protected] & 66@xmath14513 & [email protected] & 2705@xmath14533 & 533@xmath14518 & 6.69 + b129 & 8@xmath1458 & [email protected] & 106@xmath14516 & [email protected] & 2463@xmath14529 & 710@xmath14575 & 7.06 + b132 & 20@xmath1459 & [email protected] & 164@xmath14520 & [email protected] & 12811@xmath14587 & 3669@xmath145387 & 8.99 + b135 & 20@xmath1455 & [email protected] & 56@xmath14512 & [email protected] & 2990@xmath14535 & 742@xmath14527 & 7.07 + b136 & 118@xmath1459 & [email protected] & 65@xmath14511 & [email protected] & 578@xmath14531 & 289@xmath14535 & 5.30 + b137 & 10@xmath14521 & [email protected] & 19@xmath14514 & [email protected] & 746@xmath14525 & 82@xmath1459 & 4.84 + b139 & 7@xmath1453 & [email protected] & 38@xmath14510 & [email protected] & 1521@xmath14526 & 850@xmath14546 & + b140 & 97@xmath1459 & [email protected] & 63@xmath14513 & [email protected] & 672@xmath14562 & 225@xmath14532 & 5.53 + b141 & 629@xmath14514 & [email protected] & 9717@xmath14550 & [email protected] & 136366@xmath145514 & 29269@xmath1453996 & 10.93 + b142 & 20@xmath1457 & [email protected] & 28@xmath14513 & [email protected] & 2986@xmath14532 & 421@xmath14514 & 6.35 + b144 & 11@xmath1454 & [email protected] & 77@xmath14515 & [email protected] & 2606@xmath14535 & 832@xmath14531 & 7.33 + b145 & 10@xmath1454 & [email protected] & 169@xmath14511 & [email protected] & 3905@xmath14549 & 1544@xmath14565 & 7.52 + b148 & 7@xmath14514 & [email protected] & 41@xmath14515 & [email protected] & 922@xmath14527 & 165@xmath14518 & 5.50 + b152 & 4@xmath1455 & [email protected] & & & 2813@xmath14529 & 649@xmath14522 & 7.01 + b154 & 37@xmath1457 & [email protected] & & & 1414@xmath14529 & 702@xmath145101 & + b156 & 38@xmath1457 & [email protected] & 67@xmath14517 & [email protected] & 2456@xmath14532 & 347@xmath14511 & 6.12 + b159 & 1@xmath14510 & [email protected] & 23@xmath1459 & [email protected] & 1299@xmath14528 & 324@xmath14535 & 6.27 + b160 & 21@xmath1453 & [email protected] & 225@xmath14512 & [email protected] & 9941@xmath14562 & 5216@xmath145249 & 8.42 + b161 & 1@xmath1456 & [email protected] & & & 2090@xmath14535 & 890@xmath14596 & + b163 & 10@xmath1455 & [email protected] & 132@xmath14511 & [email protected] & 5637@xmath14545 & 1645@xmath14572 & 7.97 + b165 & 27@xmath14519 & [email protected] & 44@xmath14514 & [email protected] & 1347@xmath14528 & 166@xmath14518 & 5.49 + b167 & @xmath16615 & @xmath1670.09 & & & 1076@xmath14555 & 94@xmath14511 & 6.13 + b169 & 12@xmath1454 & [email protected] & & & 1086@xmath14526 & 392@xmath14520 & 6.46 + b172 & 37@xmath1457 & [email protected] & & & 535@xmath14533 & 259@xmath14533 & + b175 & 22@xmath1456 & [email protected] & & & 1007@xmath14527 & 507@xmath14558 &
+ b177 & 11@xmath1453 & [email protected] & & & 1629@xmath14541 & 1046@xmath14563 & + b179 & 3@xmath14516 & [email protected] & & & 2321@xmath14529 & 362@xmath14538 & 6.33 + b182 & 17@xmath14524 & [email protected] & 48@xmath14515 & [email protected] & 1090@xmath14524 & 112@xmath14512 & 5.04 + b183 & 36@xmath1458 & [email protected] & 237@xmath14514 & [email protected] & 12674@xmath145169 & 2370@xmath14599 & 8.16 + b184 & 1@xmath14512 & [email protected] & & & 707@xmath14537 & 166@xmath14521 & 5.49 + b188 & @xmath1689 & @xmath1690.11 & & & 3046@xmath14530 & 783@xmath14583 & 7.45 + b190 & 280@xmath14546 & [email protected] & 301@xmath14520 & [email protected] & 597@xmath14527 & 30@xmath1453 & 3.39 + b191 & 697@xmath14524 & [email protected] & 5744@xmath14539 & [email protected] & 135114@xmath145689 & 16765@xmath1452269 & 10.33 + b192 & 5@xmath14517 & [email protected] & & & 954@xmath14525 & 133@xmath14514 & 5.45 + b193 & 13@xmath14512 & [email protected] & & & 1014@xmath14526 & 218@xmath14523 & 5.77 + lcccccccc b021 & b0 & 135@xmath170 & 0.17@xmath171 & 0.42@xmath172 & 954@xmath173 & 2.3 & 0.5 & 13 + b022 & b0 & 132@xmath174 & 1.08@xmath175 & 2.30@xmath176 & 994@xmath177 & 2.5 & 0.5 & 80 + b023 & o7 & 127@xmath178 & 8.25@xmath179 & 3.28@xmath180 & 2211@xmath181 & 5.8 & 1.3 & 1689 + b029 & b0 & 118@xmath182 & 1.82@xmath183 & 1.95@xmath184 & 1248@xmath185 & 3.5 & 0.5 & 134 + b042 & b0 & 136@xmath174 & 1.12@xmath175 & 2.88@xmath186 & 933@xmath187 & 2.3 & 0.5 & 82 + b043 & o9 & 125@xmath188 & 0.71@xmath189 & 0.57@xmath190 & 1526@xmath191 & 4.0 & 0.8 & 83 + b047 & o7 & 129@xmath192 & 0.70@xmath193 & 0.31@xmath194 & 2144@xmath195 & 5.5 & 1.3 & 144 + b048 & b2 & 128@xmath188 & 0.58@xmath196 & 7.94@xmath197 & 380@xmath198 & 1.0 & 0.1 & 10 + b059 & o9 & 128@xmath199 & 0.43@xmath200 & 0.39@xmath201 & 1470@xmath202 & 3.8 & 0.8 & 50 + b064 & b0 & 115@xmath203 & 0.46@xmath204 & 0.42@xmath205 & 1316@xmath206 & 3.8 & 0.5 & 34 + b066 & o9 & 129@xmath192 & 0.88@xmath207 & 0.86@xmath208 & 1438@xmath209 & 3.7 & 0.8 & 104 + b072 & o8.5 & 136@xmath192 & 0.36@xmath210 & 0.39@xmath211 & 1432@xmath212 & 3.5 & 0.9 & 49 + b073 & b1 & 116@xmath188 & 1.73@xmath213 & 5.63@xmath214 & 705@xmath215 & 2.0 & 0.2 & 54 + b080 & o8.5 & 127@xmath192 & 1.38@xmath216 & 0.95@xmath217 & 1665@xmath218 & 4.4 & 0.9 & 187 + b082 & b1 & 130@xmath192 & 1.25@xmath219 & 8.13@xmath220 & 559@xmath221 & 1.4 & 0.2 & 39 + b087 & b0 & 128@xmath222 & 0.61@xmath223 & 1.08@xmath224 & 1059@xmath225 & 2.7 & 0.5 & 45 + b088 & o7 & 131@xmath226 & 0.26@xmath227 & 0.12@xmath228 & 2095@xmath229 & 5.3 & 1.3 & 53 + b089 & b1 & 126@xmath230 & 0.17@xmath231 & 0.90@xmath232 & 598@xmath233 & 1.6 & 0.2 & 5 + b096 & b0 & 139@xmath170 & 0.11@xmath234 & 0.32@xmath235 & 893@xmath236 & 2.1 & 0.5 & 8 + b098 & b1 & 153@xmath237 & 0.05@xmath238 & 0.93@xmath239 & 407@xmath240 & 0.9 & 0.2 & 2 + b105 & b2 & 118@xmath241 & 0.80@xmath242 & 6.69@xmath243 & 448@xmath244 & 1.3 & 0.1 & 14 + b107 & b1 & 146@xmath222 & 0.22@xmath171 & 2.88@xmath245 & 445@xmath246 & 1.0 & 0.2 & 7 + b126 & b1 & 131@xmath247 & 0.12@xmath248 & 0.81@xmath249 & 552@xmath250 & 1.4 & 0.2 & 4 + b128 & o9 & 130@xmath222 & 0.51@xmath251 & 0.52@xmath252 & 1413@xmath253 & 3.6 & 0.8 & 59 + b136 & b0 & 157@xmath188 & 0.05@xmath254 & 0.29@xmath255 & 707@xmath256 & 1.5 & 0.5 & 4 + b137 & b1 & 131@xmath247 & 0.14@xmath257 & 0.92@xmath258 & 552@xmath259 & 1.4 & 0.2 & 4 + b139 & o9 & 131@xmath188 & 0.28@xmath260 & 0.29@xmath261 & 1410@xmath262 & 3.6 & 0.8 & 33 + b142 & o8.5 & 118@xmath263 & 0.90@xmath264 & 0.42@xmath265 & 1909@xmath266 & 5.4 & 0.9 & 123 + b148 & b0 & 139@xmath237 & 0.13@xmath267 & 0.37@xmath268 & 893@xmath269 & 2.1 & 0.5 & 9 + b159 & b0 & 126@xmath237 & 0.29@xmath270 & 0.46@xmath271 & 1095@xmath272 & 2.9 & 0.5 & 21 + b182 & b0 & 139@xmath273 & 0.15@xmath274 & 0.44@xmath275 & 895@xmath276 & 2.1 & 0.5 & 11 + lcccccccc hr 533 & b1 & 126 & 0.20 & 1.06 & 593 & 1.4 & 0.2 & 16 + hr 1191 & b1 & 139 & 0.06 & 0.58 & 488 & 1.3 & 0.2 & 6
+ hr 1763 & b1 & 120 & 5.68 & 23.17 & 653 & 2.0 & 0.2 & 426 + hr 1798 & b2 & 124 & 0.07 & 0.83 & 408 & 1.0 & 0.1 & 3 + hr 1855 & b0 & 107 & 0.16 & 0.10 & 1522 & 4.6 & 0.5 & 23 + hr 1868 & b1 & 124 & 0.48 & 2.38 & 614 & 1.8 & 0.2 & 39 + hr 1931 & o9 & 103 & 2.39 & 0.58 & 2278 & 7.1 & 0.8 & 487 + hr 2058 & b1 & 108 & 0.18 & 0.39 & 814 & 2.2 & 0.2 & 11 + hr 2734 & b0 & 162 & 0.08 & 0.61 & 664 & 1.2 & 0.5 & 27 + hr 5993 & b1 & 137 & 0.06 & 0.57 & 506 & 1.4 & 0.2 & 6
+ hr 6028 & b2 & 121 & 0.10 & 0.92 & 428 & 1.1 & 0.1 & 4 + hr 6165 & b0 & 145 & 1.66 & 6.01 & 831 & 1.9 & 0.5 & 422 + hr 6601 & b1 & 130 & 0.08 & 0.52 & 558 & 1.3 & 0.2 & 7
+ hr 6788 & b1 & 115 & 0.08 & 0.27 & 711 & 2.2 & 0.2 & 6 + hr 7993 & b0 & 146 & 0.04 & 0.16 & 814 & 1.7 & 0.5 & 11 + hr 8047 & b1 & 150 & 0.02 & 0.33 & 423 & 1.0 & 0.2 & 3 + hr 8854 & b0 & 121 & 0.77 & 0.95 & 1188 & 3.2 & 0.5 & 137 + | we present the results of optical spectroscopic follow - up of 125 candidate main sequence ob stars in the small magellanic cloud ( smc ) that were originally identified in the s@xmath0mc infrared imaging survey as showing an excess of emission at 24 indicative of warm dust , such as that associated with a transitional or debris disks .
we use these long - slit spectra to investigate the origin of the 24 @xmath1 m emission and the nature of these stars .
a possible explanation for the observed 24 excess , that these are emission line stars with dusty excretion disks , is disproven for the majority of our sources .
we find that 88 of these objects are normal stars without line emission , with spectral types mostly ranging from late - o to early - b ; luminosity classes from the literature for a sub - set of our sample indicate that most are main - sequence stars .
we further identify 17 emission - line stars , 7 possible emission - line stars , and 5 other objects with forbidden - line emission in our sample .
we discover a new o6iaf star ; it exhibits strong 4686 emission but relatively weak 4640 emission that we attribute to the lower nitrogen abundance in the smc .
two other objects are identified with planetary nebulae , one with a young stellar object , and two with x - ray binaries . to shed additional light on the nature of the observed 24 excess we use optical and infrared photometry to estimate the dust properties of the objects with normal o and b star spectra and
compare these properties to those of a sample of hot spots in the galactic interstellar medium ( ism ) .
we find that the dust properties of the dusty ob star sample resemble the properties of the galactic sample of hot spots
. some may be runaway systems with bow - shocks resulting from a large velocity difference between star and ism .
we further investigate the nature of these dusty ob stars in a companion paper presenting mid - infrared spectroscopy and additional imaging . |
[ ls def ] we denote by @xmath5 the symmetric difference between two sets @xmath6 , @xmath7 , and by @xmath8 the symmetric difference distance between these sets .
( all the sets considered in this chapter are finite . )
the symbols @xmath9 and @xmath10 stand for the disjoint union and the proper inclusion of sets respectively .
a _ ( knowledge ) structure _ is a pair @xmath11 where @xmath1 is a non empty set and @xmath2 is a family of subsets of @xmath1 containing @xmath12 and @xmath13 .
the latter is called the _ domain _ of @xmath11 .
the elements of @xmath1 are called _ items _ and the sets in @xmath2 are _ ( knowledge ) states_. since @xmath14 , the set @xmath1 is implicitly defined by @xmath2 and we can without ambiguity call @xmath2 a knowledge structure .
a knowledge structure @xmath2 is _ well graded _ if for any two distinct states @xmath6 and @xmath7 with @xmath15 there exists a sequence @xmath16 such that @xmath17 and @xmath18 for @xmath19 .
we call such a sequence @xmath20 a _ tight path _ from @xmath6 to @xmath7 .
we say that @xmath2 is a _ knowledge space _ if it is closed under union , or _
@xmath21-closed_. a knowledge structure @xmath11 is a _ learning space _ ( cf . * ? ? ?
* ) if it satisfies the following two conditions : learning smoothness . for any two @xmath22 , with @xmath23 and @xmath24 ,
there is a chain @xmath25 such that , for @xmath19 , we have @xmath26 for some @xmath27 . learning consistency . if @xmath23 are two sets in @xmath2 such that @xmath28 for some @xmath29 , then @xmath30 .
a learning space is also known in the combinatorics literature as an ` antimatroid ' , a structure introduced independently by @xcite with slightly different , but equivalent axioms ( cf .
also * ? ? ?
* ; * ? ? ?
another name is ` well
graded knowledge space ' @xcite ; see our lemma [ echan ] .
a family @xmath31 of subsets of a set @xmath1 is a
_ partial knowledge structure _ if it contains the set @xmath32 .
we do not assume that @xmath33 .
we also call ` states ' the sets in @xmath31 .
a partial knowledge structure @xmath31 is a _ partial learning space _ if it satisfies axioms [ l1 ] and [ l2 ] .
note that @xmath3 is vacuously well - graded and vacuously @xmath21-closed , with @xmath34 .
thus , it a partial knowledge structure and a partial learning space ( cf .
lemma [ partialechan ] ) .
the following preparatory result will be helpful in shortening some proofs .
[ toprovewg ] a @xmath21-closed family of set @xmath2 is well
graded if , for any two sets @xmath35 , there is a tight path from @xmath6 to @xmath7 .
suppose that the condition holds .
for any two distinct sets @xmath6 and @xmath7 , there exists a tight path @xmath36 and another tight path @xmath37 .
these two tight paths can be concatenated .
reversing the order of the sets in the latter tight path and redefining @xmath38 we get the tight path @xmath39 , with @xmath40 .
as mentioned in our introduction , some knowledge structures may be so large that a splitting is required , for convenient storage in a computer s memory for example .
also , in some situations , only a representative part of a large knowledge structure may be needed .
the concept of a projection is of critical importance in this respect .
we introduce a tool for its construction .
[ def sim ] let @xmath11 be a partial knowledge structure with @xmath41 and let @xmath0 be any proper non empty subset of @xmath1 . define a relation @xmath42 on @xmath2 by @xmath43 thus , @xmath42 is an equivalence relation on @xmath2 .
when the context specifies the subset @xmath0 , we sometimes use the shorthand @xmath44 for @xmath42 in the sequel .
the equivalence between the right hand sides of ( [ def sim eq ] ) and ( [ def sim eq 2 ] ) is easily verified .
we denote by @xmath45 $ ] the equivalence class of @xmath44 containing @xmath6 , and by @xmath46\st k\in\kkk\}$ ] the partition of @xmath2 induced by @xmath44 .
we may say for short that such a partition is induced by the set @xmath0 . in the sequel
we always assume that @xmath41 , so that @xmath47 . [ def projection ]
let @xmath48 be a partial knowledge structure and take any non empty proper subset @xmath0 of @xmath1 .
the family @xmath49 is called the _ projection _ of @xmath2 on @xmath0 .
we have thus @xmath50 as shown by example [ examp projection ] , the sets in @xmath4 may not be states of @xmath2 . for any state @xmath6 in @xmath2 and with @xmath45
$ ] as in definition [ def sim ] , we define the family @xmath51}= \{m\st m= l\setminus \cap[k]\text { for some } l\sim k\}.\ ] ] ( if @xmath52 , we have thus @xmath53}=[\es]$ ] . )
the family @xmath54 is called a _
@xmath0-child _ , or simply a _ child _ of @xmath2 ( _ induced by @xmath0 _ ) . as shown by the example below
, a child of @xmath2 may take the form of the singleton @xmath3 and we may have @xmath55 } = \kkk_{[l]}$ ] even when @xmath56 .
the set @xmath3 is called the _
trivial _ child .
[ examp projection ] equation ( [ lin_space_ex ] ) defines a learning space @xmath31 on the domain @xmath57 : @xmath58 -4.5 cm the inclusion graph of this learning space is pictured by the grey parts of the diagram of figure [ projection_graph ] .
the sets marked in black in the eight ovals of the figure represents the states of the projection @xmath59 of @xmath31 on the set @xmath60 .
it is clear that @xmath59 is a learning space in this particular case .
this property does hold in general . ] .
each of these ovals also surrounds the inclusion subgraph corresponding to an equivalence class of the partition @xmath61 .
this is consistent with lemma [ proj def k struct = > ]
( ii ) according to which there is a 1 - 1 correspondence between @xmath61 and @xmath59 . in this example , the ` learning space ' property is hereditary in sense that not only is @xmath59 a learning space , but also any child of @xmath31 is a learning space or a partial learning space .
indeed , we have @xmath62}= \fff_{[\{a , b , c , e\ } ] } = \{\{b\},\{c\},\{b , c\},\{b , c , e\}\},\\ \fff_{[\ { b , c , d , e\}]}= \fff_{[\ { b , c , d , e , f\ } ] } = \fff_{[\{a , b , c , d , e\ } ] } = \{\es , \{c\},\{c , e\}\},\\ \fff_{[\ { a , b , c , d , e , f , g\ } ] } = \{\es , \{c\},\{c , e\},\{c , e , g\}\}\\ \fff_{[\ { b , c , e , f\ } ] } = \fff_{[\{a , b , c , e , f\ } ] } = \{\es\}.\end{gathered}\ ] ] these four children are represented in the four black rectangles in figure [ projection_graph ] .
theorem [ projectiontheorem1 ] shows that this hereditary property is general in the sense that the children of a partial learning space are always partial learning spaces . in the particular case of this example , just adding the set @xmath3 to the child not containing it already , that is , to the child @xmath63}=\fff_{[\{a , b , c , e\}]}$ ] , would result in having all the children being learning spaces or trivial .
this is * not * generally true .
the situation is clarified by theorem [ projectiontheorem2 ] .
[ rem project ] the concept of projection for learning spaces is closely related to the concept bearing the same name for media introduced by @xcite .
the projection theorems [ projectiontheorem1 ] and [ projectiontheorem2 ] , the main results of this chapter , could be derived via similar results concerning the projections of media ( cf .
theorem 2.11.6 in * ? ? ?
this would be a detour , however .
the route followed here is direct . in the next two lemmas , we derive a few consequences of definition [ def projection ] .
[ proj def k struct = > ] the following two statements are true for any partial knowledge structure @xmath11 .
the projection @xmath4 , with @xmath64 , is a partial knowledge structure .
if @xmath11 is a knowledge structure , then so is @xmath4 .
the function @xmath65\mapsto k\cap q'$ ] is a well defined bijection of @xmath66 onto @xmath4 .
\(i ) both statements stem from the observations that @xmath67 and @xmath68 .
\(ii ) that @xmath69 is a well defined function is due to ( [ def sim eq ] ) .
it is clear that @xmath70 by the definitions of @xmath69 and @xmath4 .
suppose that , for some @xmath45,[l]\in\kkk_{\sim}$ ] , we have @xmath71 ) = k \cap q ' = h([l ] ) = l\cap q ' = x$ ] .
whether or not @xmath72 , this entails @xmath73 and so @xmath45=[l]$ ] .
[ proj def k space = > ] if @xmath2 is a @xmath21-closed family , then the following three statements are true . @xmath74 $ ] for any @xmath75 .
@xmath4 is a @xmath21-closed family . if @xmath2 is a knowledge space , so is @xmath4 .
the children of @xmath2 are also @xmath21-closed . in ( ii ) and ( iii ) , the converse implications are not true . for knowledge spaces , lemma [ proj def k space = > ] ( ii )
was obtained by ( * ? ? ?
* theorem 1.16 on p. 25 ) where the concept of a projection was referred to as a ` substructure . '
their proof applies here .
we include it for completeness .
\(i ) as @xmath76 $ ] is the union of states of @xmath2 , we get @xmath76\in\kkk$ ] .
we must have @xmath77)\cap q'$ ] because @xmath78 for all @xmath79 $ ] ; so @xmath80 $ ] .
\(ii ) since @xmath81 by hypothesis , @xmath2 is a knowledge structure .
lemma [ proj def k struct = > ] ( i ) , implies that @xmath4 is a partial knowledge structure .
any subfamily @xmath82 is associated to the family @xmath83 .
as @xmath2 is a partial knowledge space , we get @xmath84 , yielding @xmath85 , with @xmath86 thus @xmath4 is a partial knowledge space .
this argument is valid for knowledge spaces .
\(iii ) take @xmath87 arbitrarily .
we must show that @xmath54 is @xmath21-closed . if @xmath88 , this is vacuously true .
otherwise , for any @xmath89 we define the associated family @xmath90 \in \hhh\}.\ ] ] so , @xmath91 $ ] , which gives @xmath92 for any @xmath93 .
we get thus @xmath94 . since @xmath2 is @xmath21-closed , we have @xmath95 .
the @xmath21-closure of @xmath54 follows from the string of equalities @xmath96 ) = \cup_{h'\in\hhh'}(h'\cap ( \overline{\cap[k ] } ) = ( \cup_{h'\in\hhh'}h')\setminus \cap[k]\end{aligned}\ ] ] which gives @xmath97 because @xmath98 .
example [ onehd ] shows that the reverse implications in ( ii ) and ( iii ) do not hold .
[ onehd ] consider the projection of the knowledge structure @xmath99 on the subset @xmath100 .
we have thus the two equivalence classes @xmath101 $ ] and @xmath102 $ ] , with the projection @xmath103 . the two @xmath100-children are @xmath104}= \big\{\es,\{a\ } , \{b\},\{a , b\}\big\}$ ] and @xmath105 } = \big\{\ , \es,\ , \{a\ } , \{a , b\ } \,\big\}$ ] .
both @xmath104}$ ] and @xmath105}$ ] are well - graded and @xmath21-closed , and so is @xmath106 , but @xmath107 is not since @xmath108 is not a state .
we omit the proof of the next result which is due to @xcite .
[ echan ] a knowledge structure @xmath11 is a learning space if and only if it a well graded knowledge space .
as indicated by the next lemma , the equivalence ceases to hold in the case of partial spaces .
[ partialechan ] any well - graded @xmath21-closed family is a partial learning space .
the converse implication is false .
let @xmath2 be a well - graded @xmath21-closed family .
axiom [ l1 ] is a special case of the well - gradedness condition .
if @xmath35 for two sets @xmath6 and @xmath7 in @xmath2 and @xmath109 is in @xmath2 , then the set @xmath110 is in @xmath2 by @xmath21closure , and so [ l2 ] holds .
the example below disproves the converse .
[ l - space(not=>)u - closed ] the family of sets @xmath111 is a partial learning space since it is the union of the two chains @xmath112 with the only common set @xmath113 .
however , @xmath114 is neither @xmath21-closed nor well - graded .
we state the first of our two projection theorems .
[ projectiontheorem1 ] let @xmath2 be a learning space ( resp . a well - graded @xmath21-closed family ) on a domain @xmath1 with @xmath115 .
the following two properties hold for any proper non empty subset @xmath0 of @xmath1 : the projection @xmath116 of @xmath2 on @xmath0 is a learning space ( resp . a well - graded @xmath21-closed family ) ; in either case , the children of @xmath2 are well - graded and @xmath21-closed families .
note that we may have @xmath88 in ( ii ) ( cf .
example [ examp projection ] ) .
\(i ) if @xmath2 is a learning space , then @xmath4 is a knowledge structure by lemma [ proj def k struct = > ] ( i ) .
since @xmath2 is @xmath21-closed by lemma [ echan ] , so is @xmath54 by lemma [ proj def k space = > ] ( ii ) .
it remains to show that @xmath54 is well - graded .
( by lemma [ echan ] again , this will imply that @xmath4 is a learning space . )
we use use lemma [ toprovewg ] for this purpose .
take any two states @xmath117 in @xmath4 with @xmath118 for some positive integer @xmath119 . by lemma [ proj def k struct = > ] ( ii )
, we have thus @xmath120 and @xmath121 for some @xmath122 . as @xmath2 well - graded by lemma [ echan ] , there exists a tight path @xmath123 from @xmath6 to @xmath7 , with either @xmath124 or @xmath125 for some @xmath126 and @xmath127 .
let @xmath128 be the first index such that @xmath129 .
we have then necessarily @xmath130 and @xmath124 .
this yields @xmath131 with @xmath132 . defining @xmath133 and @xmath134
, we have @xmath135 , with @xmath136 a state of @xmath4 . by induction
, we conclude that @xmath4 a wellgraded @xmath21 closed knowledge structure .
since @xmath4 is @xmath21-closed , it must be a learning space by lemma [ echan ] .
suppose now that @xmath2 is a well - graded @xmath21-closed family ( rather than learning space ) .
in such a case , there is no need to invoke lemma [ echan ] and the above argument can be used to prove that @xmath4 a wellgraded @xmath21 closed family .
\(ii ) take any child @xmath54 of @xmath2 . by lemma [ proj def k space = > ] ( iii )
, @xmath54 is a @xmath21-closed family .
we use lemma [ toprovewg ] to prove that @xmath54 is also well - graded .
take any two states @xmath137 in @xmath54 .
we have thus @xmath138)$ ] and @xmath139)$ ] for some @xmath140 and @xmath141 in @xmath45 $ ] , with @xmath142\sb l'\subset m'.\ ] ] since @xmath2 is well - graded , there is a tight path @xmath143 with all its states in @xmath45 $ ] .
indeed , @xmath144 and @xmath145 imply @xmath146 for any index @xmath147 .
we now define the sequence @xmath148 $ ] , @xmath149 .
it is clear that ( [ capksbl ] ) and ( [ tightpathchain ] ) imply @xmath150 and it is easily verified that @xmath151 is a tight path from @xmath7 to @xmath152 . applying lemma [ toprovewg ] , we conclude that @xmath54 is well - graded .
[ assinges ] in example [ examp projection ] , we had a situation in which the non trivial children of a learning space were either themselves learning spaces , or would become so by the addition of the set @xmath3 .
this can happen if and only if the subset @xmath0 of the domain defining the projection satisfies the condition spelled out in the next definition . [ yielding ]
suppose that @xmath11 is a partial knowledge structure , with @xmath41 .
a subset @xmath64 is _ yielding _ if for any state @xmath7 of @xmath2 that is minimal for inclusion in some equivalence class @xmath45 $ ] , we have @xmath153 |\leq 1 $ ] .
we recall that @xmath45 $ ] is the equivalence class containing @xmath6 in the partition of @xmath2 induced by @xmath0 ( cf .
definition [ def sim ] ) . for any non trivial child @xmath54 of @xmath2 ,
we call @xmath154 a _ plus _ child of @xmath2 .
[ projectiontheorem2 ] suppose that @xmath11 is a learning space with @xmath41 , and let @xmath0 be a proper non empty subset of @xmath1 .
the two following conditions are then equivalent .
\(i ) the set @xmath0 is yielding .
\(ii ) all the plus children of @xmath2 are learning space , in which case @xmath155 ( cf .
example [ examp projection ] ) . ] .
( it is easily shown that any learning space has always at least on non trivial child . )
\(i ) @xmath156 ( ii ) .
by lemma [ proj def k space = > ] ( iii ) , we know that any non trivial child @xmath54 is@xmath21-closed .
this implies that the associated plus child @xmath157 is a knowledge space .
we use lemma [ toprovewg ] to prove that @xmath157 is also well graded .
suppose that @xmath7 and @xmath152 are states of @xmath157 , with @xmath158 and , say @xmath159 .
suppose that @xmath160 . by the definition of @xmath157
, we have @xmath161 $ ] and @xmath162 $ ] for some @xmath163 $ ] .
since @xmath164 , we must have @xmath165 .
because @xmath140 and @xmath141 are sets of @xmath45 $ ] , there exists in @xmath45 $ ] ( by the projection theorem [ projectiontheorem1 ] ( ii ) ) a tight path @xmath166 we show that this tight path defines a tight path @xmath167 lying entirely in @xmath157 ( actually , in @xmath54 ) . by definition of a tight path
, we have @xmath168 for some @xmath169 . defining @xmath170 $ ] ,
we get @xmath171 and thus @xmath172 . note that @xmath173 is in @xmath157 because @xmath174 is is @xmath45 $ ] .
the existence of the tight path ( [ wg path 2 in ( ii ) ] ) follows by induction .
suppose now that @xmath175 .
in view of what we just proved , we only have to show that , for any non empty @xmath176 , there is a singleton set @xmath177 with @xmath178 . by definition of @xmath157
, we have @xmath179 $ ] for some @xmath180 $ ] .
take a minimal state @xmath181 in @xmath45 $ ] such that @xmath182 and so @xmath183 \sb m$ ] .
since @xmath0 is yielding , we get @xmath184| \leq 1 $ ] . if @xmath184|=1 $ ] , then @xmath183=\{q\}\sb m$ ] for some @xmath185 with @xmath186 .
suppose that @xmath184| = 0 $ ] . thus @xmath183
= \es$ ] and @xmath181 must be the only minimal set in @xmath45 $ ] , which implies that @xmath187= n$ ] . by the wellgradedness of @xmath45 $ ] established in the projection theorem [ projectiontheorem1](ii ) , there exists some @xmath188 such that @xmath189 .
we get thus @xmath190 = ( n + \{p\})\setminus n = \{p\}\sb m\quad\text{with}\quad \{p\}\in\kkp.\ ] ] the tight path ( [ wg path 2 in ( ii ) ] ) from @xmath7 to @xmath152 exists thus in both cases . applying lemma [ toprovewg ]
, we can assert that @xmath157 is well - graded .
we have shown earlier that @xmath157 is a knowledge space .
accordingly , the plus child @xmath157 is a learning space .
\(ii ) @xmath156 ( i ) . if some equivalence class @xmath45 $ ] is a chain or is a single set that this is possible . ]
@xmath191 $ ] , then @xmath192| = 0 $ ] .
otherwise @xmath45 $ ] contains more than one minimal state .
let @xmath7 be one of these minimal states .
thus @xmath193 $ ] is a minimal non empty set of @xmath157 , which by hypothesis is a learning space . by the wellgradedness of @xmath157
, there is a tight path from @xmath12 to @xmath193 $ ] . because @xmath193 $ ] is non empty and minimal in @xmath54 , it must be a singleton .
we get thus @xmath194=1 $ ] .
performing an assessment in a large learning space @xmath11 may be impractical in view of memory limitation or for other reasons .
in such a case , a two - step or an n - step procedure may be applied . on step 1 ,
a representative subset @xmath0 of items from the domain @xmath1 is selected , and an assessment is performed on the projection learning space @xmath116 induced by @xmath0 ( cf .
projection theorem [ projectiontheorem1](i ) ) .
the outcome of this assessment is some knowledge state @xmath195 of @xmath116 which corresponds ( 1 - 1 ) to equivalence class @xmath45 $ ] of the partition of @xmath2 induced by @xmath0 ( cf .
lemma [ proj def k struct = > ] ( ii ) ) . on step 2 , the child @xmath54 is formed by removing all the common items in the states of @xmath45 $ ] .
the assessment can then be pursued on @xmath54 of @xmath2 which is a partial learning space ( cf .
projection theorem [ projectiontheorem1](ii ) ) .
the outcome of step 2 is a set @xmath193 $ ] , where @xmath7 is a state in the learning space @xmath2 .
this 2-step procedure can be expanded into a n - step recursive algorithm if necessary .
if the set @xmath0 is yielding , then any equivalence class @xmath45$]containing more than one set can be made into a learning space by a trivial transformation .
this property is not critical for the 2-step procedure outlined above .
d. albert and j. lukas , editors .
_ knowledge spaces : theories , empirical research , applications_. lawrence erlbaum associates , mahwah , nj , 1999 .
falmagne , e. cosyn , j .-
doignon , and n. thiry . .
in b. ganter and l. kwuida , editors , _ formal concept analysis , 4th international conference , icfca 2006 , dresden , germany , february 1317 , 2006 _ , lecture notes in artificial intelligence , pages 6179 .
springer - verlag , berlin , heidelberg , and new york , 2006 . | any proper subset @xmath0 of the domain @xmath1 of a learning space @xmath2 defines a projection of @xmath2 on @xmath0 which is itself a learning space consistent with @xmath2 .
such a construction defines a partition of @xmath1 having each of its classes either equal to @xmath3 , or preserving some key properties of the learning space @xmath2 , namely closure under union and wellgradedness .
if the set @xmath0 satisfies certain conditions , then each of the equivalence classes is essentially , via a trivial transformation , a learning space .
we give a direct proof of these and related facts which are instrumental in parsing large learning spaces .
# 1 _|q this paper is dedicated to george sperling whose curious , incisive mind rarely fails to produce the unexpected creative idea .
george and i have been colleagues for the longest time in both of our careers . the benefit has been mine .
learning spaces , which are special cases of knowledge spaces ( cf . * ? ? ?
* ) , are mathematical structures designed to model the cognitive organization of a scholarly topic , such as beginning algebra or chemistry 101 .
the definition of ` learning space ' is recalled in our definition [ ls def ] .
essentially , a learning space is a family of sets , called knowledge states , satisfying a couple of conditions .
the elements of the sets are ` atoms ' of knowledge , such as facts or problems to be solved .
a knowledge state is a set gathering some of these atoms .
each of the knowledge states in a learning space is intended as a possible representation of some individual s competence in the topic .
embedded in a suitable stochastic framework , the concept of a learning space provides a mechanism for the assessment of knowledge in the sense that efficient questioning of a subject on a well chosen subset of atoms leads to gauge his or her knowledge state .
many aspects of these structures have been investigated and the results were reported in various publications ; for a sample , see @xcite .
the monograph by @xcite contains most of the results up to that date . in practice , in an educational context for example
, a learning space can be quite large , sometimes numbering millions of states .
the concept of a ` projection ' at the core of this paper provides a way of parsing such a large structure into meaningful components .
moreover , when the learning space concerns a scholarly curriculum such as high school algebra , a projection may provide a convenient instrument for the programming of a placement test . for the complete algebra curriculum comprising several hundred types of problems , a placement test of a few dozens problems can be manufactured automatically via a well chosen projection .
the key idea is that if @xmath2 is a learning space on a domain @xmath1 , then any subset @xmath0 of @xmath1 defines a learning space @xmath4 on @xmath0 which is consistent with @xmath2 .
we call @xmath4 a ` projection ' of @xmath2 on @xmath0 , a terminology consistent with that used by @xcite and @xcite for media . moreover , this construction defines a partition of @xmath2 such that each equivalence class is a subfamily of @xmath2 satisfying some of the key properties of a learning space .
in fact , @xmath0 can be chosen so that each of these equivalence classes is essentially ( via a trivial transformation ) either a learning space consistent with @xmath2 or the singleton @xmath3 .
these results , entitled ` projection theorems ' ( [ projectiontheorem1 ] and [ projectiontheorem2 ] ) , are formulated in this paper .
they could be derived from corresponding results for the projections of media @xcite .
direct proof are given here .
this paper extends previous results from ( * ? ? ?
* theorem 1.16 and definition 1.17 ) and @xcite . |
superconductivity@xcite and quantum hall effects@xcite are known to present two distinct and interesting examples of zero - resistance states in material systems .
another interesting example occurs in the quantum hall two - dimensional electron system ( 2des ) when the 2des is irradiated with microwave and terahertz band radiation in the presence of a transverse magnetic field . here
, it becomes possible to photo - excite into zero - resistance states in this low dimensional system.@xcite = 3.0 in = 7.0 in = 3.5 in = 3.5 in photo - excited transport in the 2des has become a topic of experimental and theoretical interest since this observation.@xcite periodic in @xmath5 radiation - induced magnetoresistance oscillations , which lead into the radiation - induced zero - resistance states , are now understood to be a consequence of radiation - frequency ( @xmath6 ) and magnetic field ( @xmath7 ) dependent , scattering at impurities @xcite and/or a change in the distribution function@xcite . and , vanishing resistance at the oscillatory minima is explained as a feature of negative resistance instability and current domain formation.@xcite in spite of the progress , there remain many aspects that could be better understood from including , for example , the growth of the oscillatory effect vs. the radiation intensity , @xmath4 . here
, a number of works have numerically evaluated the radiation - induced magnetoresistance oscillations for several @xmath4 and graphically presented the results,@xcite while dmitriev et al.,@xcite have suggested that the correction to the dark dc conductivity is linear in @xmath4.@xcite a comparison of experiment with theory , so far as the @xmath4-dependence is concerned , could help to identify the importance of the invoked - mechanisms in these theories.@xcite thus , we examine the growth of the radiation - induced magneto - resistance oscillations with @xmath4 .
experiment indicates that the amplitude @xmath2 of the radiation - induced oscillatory diagonal resistance ( @xmath0 ) , grows nonlinearly with @xmath4 and can be described by @xmath1 where the exponent @xmath8 . at the same time , according to experiment , @xmath9 also depends on the lattice temperature @xmath10 and radiation frequency @xmath6 . since
such non - linear growth of @xmath2 with @xmath4 , @xmath10 , and @xmath6 had not been predicted , we propose a theoretical explanation .
the explanation utilizes the _ radiation - driven electron orbit model_,@xcite where radiation forces the orbit center of the landau states to move back and forth in the direction of the radiation electric field at the frequency of radiation , and the @xmath0 oscillations reflect the periodic motion of the electron orbits center .
this theory establishes that the amplitude ( @xmath11 ) of the orbit center motion , sets the amplitude @xmath2 of the @xmath0 oscillations , and that @xmath11 is proportional to the radiation electric field .
since the radiation power @xmath4 is proportional to the square of the radiation electric field , it follows that @xmath0 and @xmath2 will depend on the square root of @xmath4 , at the lowest temperatures .
the experiments indicate that @xmath9 is close to @xmath12 at lower temperatures .
the above mentioned theory@xcite was proposed to explain the @xmath0 of an irradiated 2des at low @xmath7 .
we obtained the exact solution of the corresponding electronic wave function@xcite : @xmath13 , where @xmath14 is the solution for the schrdinger equation of the unforced quantum harmonic oscillator , @xmath15 is the center of the orbit for the electron motion , @xmath16 is the classical solution of a forced harmonic oscillator : @xmath17 where @xmath18 is the electron charge , @xmath19 is a phenomenologically - introduced damping factor for the electronic interaction with acoustic phonons , @xmath20 the cyclotron frequency , @xmath21 the radiation electric field .
then , the obtained wave function is the same as the standard harmonic oscillator where the center is displaced by @xmath16 .
thus , the orbit centers are not fixed , but they oscillate harmonically at the radiation field frequency @xmath22 .
then , by this micro(mw)-driven periodic motion , electrons in their orbits ( quantum oscillators ) interact with the lattice ions being damped and emitting acoustic phonons . in the @xmath23 expression
, @xmath19 represents this damping .
this @xmath24 behavior affects dramatically the charged impurity scattering and eventually the conductivity .
then , first we calculate the impurity scattering rate @xmath25 between two @xmath26 landau states @xmath27 , and @xmath28@xcite .
next we find the average effective distance advanced by the electron in every scattering jump : @xmath29@xcite , where @xmath30 is the scattering time .
finally the longitudinal conductivity @xmath31 is given by : @xmath32 being @xmath33 the energy .
to obtain @xmath0 we use the relation @xmath34 , where @xmath35 and @xmath36 .
therefore , @xmath37 and the amplitude of resistance oscillations depends linearly on the radiation electric field .
@xmath4 can be related with @xmath0 through the well - known formula that gives radiation intensity @xmath38 ( power divided by surface ) in terms of the radiation electric field @xmath21 : @xmath39 , where @xmath40 is the speed of light in vacuum and @xmath41 is the permittivity in vacuum . if we want to express only the power in terms of the radiation electric field we have to take into account the sample surface .
in the particular case of gaas if @xmath42 is the sample surface , we can readily obtain : @xmath43 where @xmath44 is the speed of light in gaas and @xmath45 is the gaas dielectric constant .
accordingly , @xmath46 .
then , substituting in the expression of @xmath11 , we obtain that @xmath0 varies with @xmath4 following an square root law : @xmath47 thus , the @xmath0 response will grow as the square root of @xmath4 , i.e. , @xmath48 . from the expression of
@xmath0 we can study the influence of @xmath49 and @xmath10 on the sublinear relation . in the case of @xmath49 ,
the calculation is straightforward ( see expressions of @xmath0 or @xmath11 ) . in the case of @xmath10
, we introduce a microscopic model , which allow us to obtain the damping parameter @xmath19 and its dependence on @xmath10 @xcite . according to the model
, @xmath19 is proportional to the scattering rate of electron - acoustic phonon interaction , being eventually linear with @xmath10@xcite .
low frequency lock - in based electrical measurements were carried out at @xmath50 with the samples immersed in pumped liquid - helium and mounted near the open end of a microwave waveguide.@xcite the w ( wegscheider)-gaas / algaas single heterostructures were nominally characterized by an electron density , @xmath51 and a mobility of @xmath52 .
results are reported here for measurements on a set of three w - specimens . to examine the growth of the radiation - induced oscillations with @xmath4 , fig .
1 presents the @xmath53 of specimen w1 for several @xmath4 at @xmath54 .
also shown in the figure are data - fits to @xmath55 . here ,
a slowly varying background , approximately equaling the dark trace , was removed from the photo - excited @xmath0 data to realize @xmath53.@xcite although this fit function includes three parameters , @xmath2 , @xmath56 , and @xmath57 , the oscillation period in @xmath5 is independent of the radiation - intensity , and thus , @xmath57 is a constant .
further , the damping constant , @xmath56,@xcite turns out to be insensitive to @xmath4 , see fig .
thus , the main free parameter in the fit - function is the amplitude , @xmath2 , of the oscillations . in fig .
1(c ) , we exhibit the fit extracted @xmath2 vs. @xmath4 for w1 and w2 .
the figure shows a sub - linear growth of @xmath2 with @xmath4 .
also shown are power law fits , @xmath58 . here , @xmath59 and @xmath60 for w1 and w2 , respectively , at @xmath61 and @xmath54 .
@xmath62 varies between w1 and w2 because the effective intensity attenuation factor varies between the two experiments . for a third w - specimen labeled w3 , fig .
1(d ) and 1(e ) report the influence of the temperature on the growth of @xmath2 vs. @xmath4 , where @xmath2 is extracted , as before , from line - shape fits of the oscillatory data to @xmath63 . here , fig .
1(d ) and 1(e ) indicate that , as expected , at a constant @xmath4 , @xmath2 grows with decreasing @xmath10 both at @xmath64 [ fig .
1(d ) ] and @xmath65 [ fig 1(e ) ] .
further , the figures show that the @xmath2 vs. @xmath4 curves exhibit greater curvature at lower temperatures .
the @xmath2 vs. @xmath4 have been fit once again to @xmath66 ; the fit - extracted @xmath9 and @xmath62 have been summarized in tabular form within fig .
1(d ) and fig .
these fit - extracted @xmath9 have also been plotted vs. @xmath10 in the inset to these figures .
these insets suggest that @xmath9 decreases with decreasing temperatures , consistent with the observed increased non - linearity at lower temperatures .
note that , at @xmath67 , all three w - specimens , exhibit the same @xmath9 , within uncertainties , at the lowest pumped @xmath68he temperatures . in addition , a comparison of the @xmath9 reported in fig .
1(d ) and fig .
1(e ) also suggests that reducing the microwave frequency @xmath6 at a fixed @xmath10 tends to reduce the @xmath9 , i.e. , increase the non - linearity .
thus , the experimental results presented here suggest a nonlinear power law in a regime characterized by modest excitation . this peculiar behavior can be theoretically explained and the experimental results recovered by means of the radiation - driven electron orbits model .
next we present the calculated results that seem to reasonably agree with the present experiments .
fig . 2 presents the calculated @xmath0 as a function of @xmath7 over @xmath69 at @xmath70 and @xmath71 . as in experiments , we obtain multiple oscillations for all @xmath4 except , as expected , for @xmath72 .
it is clear the progressive collapse of @xmath0 oscillations as @xmath4 decreases .
the explanation is straightforward according to our theory and it is given by the expression of @xmath0 .
thus , for a decreasing ( increasing ) @xmath4 the @xmath0 response is progressively smaller ( larger ) .
the key issue here , and the challenge for the different available theories , is to deduce the rate of increase ( decrease ) of @xmath0 oscillations for increasing ( decreasing ) @xmath4 .
this is shown in the inset of fig .
2 , where we present the calculated amplitude versus @xmath4 . as in experiments ,
we obtain a clear sublinear growth of @xmath2 with @xmath4 : @xmath73 .
we would expect an exponent @xmath9 around @xmath12 since the model establishes an square root dependence .
thus , we have performed a fit of the calculated values , obtaining @xmath74 , very close to a square root law .
= 3.5 in = 5.2 in the dependence of the sublinear law on @xmath10 and frequency is presented in fig .
3 . in fig . 3a we represent the calculated amplitude @xmath2 of the main @xmath0 peak versus @xmath4 for a frequency of @xmath54 and @xmath10 values : @xmath75 , @xmath76 , @xmath77 and @xmath78 .
we obtain a qualitatively similar behavior as in experiments .
thus , we observe that for a constant @xmath4 , @xmath2 increases ( decreases ) with decreasing ( increasing ) @xmath10 . the physical explanation is a follows .
if one increases @xmath10 , the interaction between electrons and lattice ions also increases giving rise to a more intense damping .
this implies progressively smaller amplitudes and @xmath0 oscillations tend to vanish . on the other hand
, it is very clear that for all @xmath10 used in the calculation , the dependence between @xmath2 and @xmath4 is sub - linear and that , at least qualitatively , the curves show a bigger curvature for decreasing @xmath10 . to check this curvature ,
we have made a fit to the calculated values for each @xmath10 observing that the exponent of the fits slightly increases for increasing @xmath10 . in all cases ,
the exponent is always close to 0.5 .
this outcome is expected since , strictly speaking , our model does not predict any variation of the exponent with @xmath10 ; our theoretical framework imposes a square root relation . yet
, experimental results show a much faster increase of the exponent with increasing @xmath10 regarding the calculated ones .
the latter show a much slower variation with @xmath10 , although in the same direction . at present
, this model can not explain this quantitative discrepancy of the variation of @xmath9 with @xmath10
. this discrepancy could be the result of an statistical effect of the fit or the reflex of a real physical process .
in the latter case , this would imply an extension of the current theory that should start from the square root law obtained here .
we consider that the square root dependence is a solid result reflecting a reasonably good agreement between theory and experiments a low temperatures . a similar behavior is presented in fig .
3b where we show the calculated values of @xmath2 versus @xmath4 for different frequencies .
we observe that , as in experiments , @xmath2 decreases as the the frequency increases .
the explanation comes from the expression of @xmath0 where @xmath49 shows up in the denominator .
then , for a constant @xmath4 , a larger @xmath49 gives rise to smaller amplitudes .
we have fit these curves and , as expected , the exponent of the fit slightly increases for increasing frequency whereas the pre - factor decreases . yet ,
experiments show a much faster increase .
this discrepancy is similar to the case of the @xmath10-dependence .
thus , the explanation given in fig .
3a , applies identically here too .
all calculated results presented in figs . 2 and 3 , have been obtained directly through equation ( 6 ) .
in summary , experimental results indicate a nonlinear growth in the amplitude @xmath2 of radiation - induced magneto - resistance oscillations with @xmath4 .
these results can be explained with the radiation - driven electron orbits model . according to this model
, the amplitude of the radiation - induced oscillations should be proportional to the square root of radiation power , which implies a sub - linear relation .
we have also studied how this sub - linear law varies with lattice temperature and radiation frequency obtaining only a qualitative good agreement between theory and experiments .
the obtained quantitative discrepancies can not be explained by our theory .
the origin of them could be an statistical effect or a real physical effect . in the latter case
, this would imply an extension of the present theoretical model , always starting from the square root law here obtained .
r. g. mani , appl .
. lett . * 85 * , 4962 ( 2004 ) ; ieee trans . on nanotech .
* 4 * , 27 ( 2005 ) ; phys .
b * 72 * , 075327 ( 2005 ) ; appl . phys . lett . * 91 * , 132103 ( 2007 ) ; sol . st
. comm . * 144 * , 409 ( 2007 ) ; appl .
. lett . * 92 * , 102107 ( 2008 ) ; physica e , * 40 * , 1178 ( 2008 ) . | we find that the amplitude of the @xmath0 radiation - induced magnetoresistance oscillations in gaas / algaas system grows nonlinearly as @xmath1 where @xmath2 is the amplitude and the exponent @xmath3 .
this striking result can be explained with the radiation - driven electron orbits model , which suggests that the amplitude of resistance oscillations depends linearly on the radiation electric field , and therefore on the square root of the power , @xmath4 . we also study how this sub - linear power law varies with lattice temperature and radiation frequency . |
we assume that for a given city , we have the @xmath1 matrix @xmath2 where @xmath3 is the number of spatial units that compose the city at the spatial aggregation level considered ( for example a grid composed of square cells of size @xmath4 , see methods ) .
this od matrix @xmath2 represents the number of individuals living in the location @xmath5 and commuting to the location @xmath6 where they have their main , regular activity ( work or school for most people ) . by convention ,
when computing the numbers of inhabitants and workers in each cell we do not consider the diagonal of the od matrix .
this means that we omit the individuals who live and work in the same cell ( considered as ` immobile ' at this spatial scale ) . in order to extract a simple signature of the od matrix
, we proceed in two steps .
we first extract both the residential and the work locations with a large density the so called ` hotspots ' ( see @xcite ) .
the number of residents of cell @xmath5 is given by @xmath7 and its number of workers is given by @xmath8 .
the hotspots then correspond to local maxima of these quantities .
it is important to note that the method is general , and does not depend on how we determine these hotspots .
once we have determined the cells that are the residential and the work hotspots ( some cells can possibly be both ) , we proceed to the second and main step of the method .
we reorder the rows and columns of the od matrix in order to separate hotspots from non - hotspots .
we put the @xmath9 residential hotspots on the top lines , and do the same for columns by putting the @xmath10 work hotspots on the left columns .
the od matrix then becomes a 4-quadrants matrix where the flows @xmath2 are spatially positioned in the matrix with respect to their nature : on the top left the individuals that live in hotspots and work in hotspots ; at the top right the individuals that live in hotspots and do not work in hotspots ; at the bottom left individuals that do not live in hotspots but work in hotspots ; and finally in the bottom right corner the individuals that neither live or work in hotspots . for each quadrant
we sum the number of commuters and normalize it by the total number of commuters in the od matrix , which gives the proportion of individuals in each of the four categories of flows . in other words , for a given city , we reduce the od matrix to a @xmath0 matrix @xmath11 where @xmath12 is the proportion of * * i**ntegrated flows that go from residential hotspots to work hotspots ; by construction , we have @xmath16 $ ] and @xmath17 .
this matrix @xmath18 is thus a very simple footprint of the od matrix that gives an expressive picture of the structure of commuting in the city , as illustrated by fig .
[ fig : icdr ] . ) from hotspot to hotspot , the convergent flows ( @xmath19 ) to hotspots , the divergent flows ( @xmath20 ) originating at hotspots and finally the random flows ( @xmath21 ) , which are neither starting nor ending at hotspots . for each city with its origin - destination matrix
, we can compute the importance of each commuting flow category and get a simple picture of the mobility structure in the city . ]
large scale individual mobility networks are nowadays extracted from pervasive geolocated data , such as mobile phone , gps , public transport cards or social apps data @xcite . in particular ,
if an individual s mobile phone geolocated activity is available during a sufficiently long period of time , it is possible under certain regularity conditions to infer the most likely locations of her home and her workplace , and by aggregation to construct od matrices @xcite .
several parameters however impact the construction of od matrices such as the nature of the data source ( survey or user - generated geolocated data ) , or the spatial scale at which the od matrix is built which can be dictated by administrative units ( divisions in wards , counties , municipalities , etc . ) or by technical reasons such as the density of antennas in the case of mobile phone data .
given this variety of data collection protocols , it is thus particularly remarkable that when considering the commuting flows at a city scale , various sources of pervasive data provide a very similar mobility information when compared to the od matrices built from surveys @xcite .
this result needs to be confirmed for other cities and countries , but it already opens the door to a systematic use of pervasive , geolocated data as a relevant substitute to traditional transport surveys . in the following we apply our icdr method to od matrices that have been extracted from mobile phone records in thirty - one spanish urban areas during a five weeks period
( see the methods section for details on the dataset and the calculation of the od matrices ) . as described above
, the first step consists in determining hotspots .
several possible methods have been proposed in the literature @xcite , and we use here a parameter - free method based on the lorenz curve of the densities that we have proposed in a recent study ( see @xcite and the methods section ) .
once we have determined both the origin ( residential ) hotspots and the destination ( work ) hotspots in each city , we first observe how their number scale with the population size of the city .
both these numbers for residential and employment hotspots scale sublinearly with the population size ( see supplementary figures 4 and 5 ) .
the number of work hotspots grows significantly slower than the number of residential hotspots , showing that residential areas are ( i ) more dispersed in the city , and ( ii ) are more numerous than activity centers , as intuitively expected ( see supplementary figure 6 that displays the locations of home and work hotspots in four cities that exhibit different spatial organizations ) .
we also note here that the sublinear scaling of the work hotspots confirms previous results obtained with a totally different dataset ( the number of employment centers in us cities ) @xcite .
we now apply the second part of the method in order to calculate the @xmath22,@xmath19,@xmath20 and @xmath21 values for each od matrix . for the 31 spanish urban areas under study ( see supplementary figure 1 ) , we obtain the values shown in fig . [
fig : icdr - values ] . in fig .
[ fig : icdr - values](a ) , we plot these values versus the population size of these cities . for this sample of cities we see that globally the proportion @xmath22 of individuals that commute from hotspot to hotspot decreases as the population size increases , while the proportion @xmath21 of ` random ' flows increases and the proportions @xmath19 and @xmath20 of convergent and divergent flows seem surprisingly constant whatever the city size . in fig . [
fig : icdr - values](b ) we plot the same values but sorted by decreasing values of @xmath22 which shows clearly that the @xmath22 and @xmath21 values are the relevant parameters for distinguishing cities from each other .
we also notice that the values obtained for another spatial scale of data aggregation confirm this trend ( see supplementary figure 7 ) .
the decay of @xmath22 flows ( integrated ) flows in favor of @xmath21 flows ( random ) when @xmath23 increases shows that the population growth among spanish cities goes with a decentralization of both activity places and residences .
as cities get bigger , their numbers of residential and employment hotspots grow ( sublinearly ) , but these hotspots catch a smaller part of the commuting flows . in order to evaluate to
what extent the @xmath24 signatures of cities are characteristic of their commuting structure , we compare these values to the ones returned by a null model of commuting flows . for each city
we generate random od matrices of the same size than the reference od matrix but with random flows of individuals that preserve the in- and out- degree of each node ( see supplementary note 5 ) .
fig [ fig : icdr - values](c ) shows the average values and standard deviation obtained for 100 replications . on fig
[ fig : icdr - values](d ) we plot the z - scores of the @xmath22 , @xmath19 , @xmath20 , and @xmath21 values of each city when compared to the values @xmath25 , @xmath26 , @xmath27 and @xmath28 returned by our null model ( e.g. for the quantity @xmath21 of the city @xmath5 , the z - score is given by @xmath29 ) . essentially , we observe that the z - scores of @xmath22 and @xmath21 are positive and large , while those of @xmath19 and @xmath20 are negative ( and large in absolute value ) . also as cities grow , the z - scores of @xmath22 and @xmath21 increase while those of @xmath19 and @xmath20 decrease .
these results demonstrate that the larger a city , and the less random it appears .
this is in contrast with the naive expectation that the larger a city and the more disordered is the structure of individuals mobility . for large cities
, there seems to be a commuting backbone which can not result from purely random movements of individuals .
this backbone is the footprint of the city s structure and history , and probably results from strong constraints and efficiency considerations .
since our method first requires to determine origin and destination hotspots , one could argue that the interpretation of the @xmath22,@xmath19,@xmath20 and @xmath21 values will crucially depend on the particular method chosen to define these hotspots .
the identification of hotspots is a problem that has been broadly discussed in urban economics ( see supplementary note 3 ) . roughly speaking , starting from a spatial distribution of densities ,
the goal is to identify the local maxima and amounts to choose a threshold @xmath30 for the density @xmath31 of individuals : a cell @xmath5 is a hotspot if the local density of people is such that @xmath32 . in order to test the impact of the choice of a particular threshold @xmath30 on the resulting @xmath24 values
, we measure the sensitivity of these values as a function of the density threshold @xmath30 ( see supplementary figure 8) .
as expected , the lower the density threshold , the larger the number of origin and destination hotspots , and consequently the larger @xmath22 and the smaller @xmath21 .
in contrast , changing the density threshold has little impact on the @xmath19 and @xmath20 terms .
more importantly , the conclusions drawn from the comparison of the @xmath24 values across cities remain the same : whatever the density threshold @xmath30 chosen to define residential / employment hotspots , we still observe the same qualitative behavior : a decay of integrated flows ( @xmath22 ) in favor of ` random ' flows ( @xmath21 ) , when the population size increases .
we now want to characterize spatially these different flows and the relation between city size and the commuting distances traveled by individuals . in each city
we compute the average distance traveled by individuals per type of flows @xmath22,@xmath19,@xmath20 and @xmath21 .
the resulting average distances measured in data are plotted in red on fig .
[ fig : avg - norm - commuting - dist](a ) .
we observe that the average distance for all categories of flows increases with population size , an expected effect as the city s area also grows with population size .
we also observe that the average distance of convergent flows @xmath19 increases faster than for other types of flows ( we note that the average distance associated to convergent flows increases more than the distance associated to divergent flows @xmath20 , showing that these flows are not symmetric as one could have naively expected ) .
this result means that for this set of spanish cities , commuters from small residential areas to important activity centers travel on average a longer distance than all other individuals .
this observation could be an indication that for our set of cities , residential areas have expanded while activity centers remained at their location , leading to longer commuting distances ( see supplementary figures 10 and 11 when considering various spatial scales of aggregation ) .
another interesting information is provided by the comparison of distances measured in the data with average distances measured from the random od matrices generated by the null model .
the average distances associated to the null model are plotted in blue in fig .
[ fig : avg - norm - commuting - dist](a ) .
we see that for all types of flows the distances measured in the empirical data are shorter than those generated by the null model .
this is another clear indication of the spatial organization of individual flows in cities .
it also highlights the importance of the travel time budget in the residential locations choice .
remarkably enough , the distance of convergent flows ( c ) is both the largest and the one that increases the fastest with population , indicating a low degree of efficiency . the comparison of this behavior with the null model leads to interesting results .
[ fig : avg - norm - commuting - dist](b ) we plot the ratio @xmath33 for the four types of flow .
values lower than @xmath34 indicate that the average commuting distance generated by the null model is shorter than the distance observed in the city .
surprisingly , we observe that small cities display a value less than one indicating the lesser importance of space at this short scale .
we also see that this ratio increases faster for random flows ( @xmath21 ) than for the others ( @xmath35 ) , suggesting a remarkable spatial structure of these @xmath21 flows .
we also consider the fraction of total commuting distance by type of flow ( fig .
[ fig : total - commuting - dist ] ) .
we see that for each type of flows , their respective fraction is constant and independent of city size . with the loubar hotspots detection method @xcite ( see supplementary figure 3 ) and with a grid of @xmath36 square cells
, we measure that roughly @xmath37 of the total commuting distance is made on random flows while the other types represent each about @xmath38 .
this result shows that the method is able to identify where most of the commuting distance is traveled .
in particular , we see that the natural , obvious flows ( i ) from residential centers to activity centers are not the most important ones , and that the decentralization of commuting flows seems to be the rule for the spanish cities in our sample .
finally , the @xmath24 signature of their od matrix allows to cluster cities with respect to the structure of their commuting patterns .
we measure the euclidian distance between the cities @xmath24 signatures and we then perform a hierarchical cluster analysis . fig .
[ fig : dendrogram ] shows the dendrogram resulting from the classification .
four well - separated clusters are identified on this dendrogram , and table [ table : classe - profiles ] gives the average value of each term along with the average population of the cities composing the cluster . remarkably these summary statistics show that largest cities are clustered together and are characterized by a larger proportion of ` random ' flows ( @xmath21 ) of individuals both living and working in parts of the city that are not the dominant residential and activity centers .
this can be interpreted as an increased facility in bigger urban areas to commute from any part of the city to any other part .
further studies on other cities and countries are needed at this stage in order to discuss the relevance of the proposed classification .
it is also important to test the robustness of this classification and we show that introducing a reasonable amount of noise in the od matrices does not change the classification ( see methods and supplementary figure 12 ) .
this sensitivity test confirms that the clustering is robust against possible errors in the data source and in the extraction of the mobility networks .
the classification of cities based on their @xmath24 values is also reasonnably robust to a change of the method used to define residential and employment hotspots ( see supplementary figure 13 ) .
comparing properties of cities of very different population sizes and areas requires to rely on a harmonized , spatial delimitation of cities that goes beyond the administrative boundaries defined by national authorities .
these administrative boundaries may be somewhat arbitrary and do not follow neither the morphological nor functionnal spatial footprint of the city . to this end
we have chosen to rely on the _ urban areas _ defined by the audes initiative ( areas urbanas de espaa : audes project .
documentation and open data available at http://alarcos.esi.uclm.es/per/fruiz/audes/ ( accessed january 27 , 2014 ) ) .
this project is an attempt to capture some coherent spatial delimitations of spanish cities regarding the ( journey to work ) commuting patterns of individuals living in the core city of each urban area and in their surrounding municipalities .
these delimitations are built upon statistical criteria based on the proportion of residents of surrounding municipalities that commute to the main city to work .
the locations and spatial boundaries of the 31 spanish urban areas analysed in this study are represented on supplementary figure s1 .
we remove the btss with zero mobile phone users and we compute the voronoi cells associated with each btss of the urban area ( hereafter called ua ) .
we remark in supplementary figure s2a that there are four types of voronoi cells : 1 .
the voronoi cells contained in the ua 2 .
the voronoi cells partitionned between the ua and the territory outside the ua .
the voronoi cells between the ua and the sea ( noted s ) .
4 . the voronoi cells between the ua , the territory outside the ua and the sea . to compute the number of users associated with the intersections between the voronoi cells and the ua we have to take into account these different types of voronoi cells .
let @xmath9 be the number of voronoi cells , @xmath39 the number of mobile phone users in the voronoi cell @xmath40 and @xmath41 the area of the voronoi cell @xmath40 , latexmath:[$v
\in we note in equation [ eq : vua ] that we remove the intersection of the voronoi area with the sea , indeed , we assume that the number of users calling from the sea are negligeable .
now we consider the number of mobile phone users @xmath44 and the associated area @xmath45 of the voronoi cells intersecting the ua ( see supplementary figure s2b ) .
we can not directly extract an od matrix between the grid cells with the mobile phone data because each users home and work locations are identified by the voronoi cells .
thus , we need a transition matrix @xmath23 to transform the bts od matrix @xmath46 into a grid od matrix @xmath47 .
let @xmath9 be the number of voronoi cells covering the urban area and @xmath3 be the number of grid cells .
let @xmath46 be the od matrix between btss where @xmath48 is the number of commuters between the bts @xmath5 and the bts @xmath6 . to transform the matrix @xmath46 into an od matrix between grid cells @xmath47 we define the transition matrix @xmath23 where @xmath49 is the area of the intersection between the grid cell @xmath5 and the bts @xmath6
. then we normalize @xmath23 by column in order to consider a proportion of the btss areas instead of an absolut value , thus we obtain a new matrix @xmath50 ( equation [ eq : pchap ] ) .
a city is divided in @xmath3 cells and the data give us access to its @xmath1 od matrix extracted from the mobile phone data . after straightforward calculation
we obtain the distributions @xmath53 and @xmath54 of the numbers of residents / workers in the @xmath3 cells composing the city .
the determination of centres and subcentres is a problem which has been broadly tackled in urban economics ( see for example the references given in the paper and references in @xcite ) . starting from a spatial distribution of densities , the goal is to identify the local maxima .
this is in principle a simple problem solved by the choice of a threshold @xmath55 for the density @xmath31 : a cell @xmath5 is a hotspot if the density of users @xmath56 .
it is however clear that such method based on a fixed threshold introduce some arbitrariness due to the choice of @xmath55 , and also requires prior knowledge of the city to which it is applied to choose a relevant value of @xmath55 . in @xcite
we proposed a generic method to determine hotspots from the lorenz curve of the densities .
in the following we quickly introduce the principle of the method and its application to the determination of the residential and work hotspots of each city .
we invite the interested readers to refer to @xcite for further discussion on this method .
we first sort @xmath53 and @xmath54 in increasing order , and denote the ranked values by @xmath57 where @xmath3 is the number of cells .
the two lorenz curves of the distribution of residents / workers are constructed by plotting on the x - axis the proportion of cells @xmath58 and on the y - axis the corresponding proportion of commuters @xmath59 with : @xmath60 if all the cells had the same number of residents / workers the lorenz curves would be the diagonal from @xmath61 to @xmath62 .
in general we observe a concave curve with a more or less strong curvature . in the lorenz curve ,
the stronger the curvature the stronger the inequality and , intuitively , the smaller the number of hotspots .
this remark allows us to construct a criterion by relating the number of dominant places ( i.e. those that have a very high number of residents / workers compared to the other cells ) to the slope of the lorenz curve at point @xmath63 : the larger the slope , the smaller the number of dominant values in the statistical distribution .
the natural way to identify the typical scale of the number of hotspots is to take the intersection point @xmath64 between the tangent of @xmath65 at point @xmath63 and the horizontal axis @xmath66 ( see suplementary figure s3 ) .
this method is inspired from the classical scale determination for an exponential decay : if the decay from @xmath63 were an exponential of the form @xmath67 where @xmath4 is the typical scale we want to extract , this method would give @xmath68 ( see @xcite for further discussion ) .
on supplementary figure s8 we plot the @xmath22,@xmath19,@xmath20,@xmath21 values of the 31 spanish urban areas as a function of the density threshold @xmath30 chosen to define hotspots ( here defined relatively to the density value @xmath69 returned by the loubar method see section [ sec : hotspots - def ] ) ) . in the extreme case represented on supplementary
figure s9 all cells whose number of residents / workers are greater than the mean value of the distribution of residents / workers are tagged as hotspots ( see @xcite for a discussion of this criteria and his comparison to the " loubar criteria used in this study ) . with a much broader acceptation of what is an hotspot , it is obvious that the term @xmath22 will increase drastically since we increase both the number of residential hotspots and the number of work hotspots . still what is important to notice is that we can still observe the qualitative trend observed with the loubar method . as the population size increases , the decrease of the proportion @xmath22 of integrated flows
is accompanied by an increase of the proportion @xmath21 of random flows . in order to evaluate to
what extent the @xmath24 values of a given city are characteristic of its commuting structure , we compare these values to the values returned by a reasonable null model of commuting flows . the guiding idea is to generate od matrices that ( i ) have the same size than the city s od matrix ( i.e that contain the same total number of individuals ) ; ( ii ) that respect the city s static spatial organisation ( i.e. the in- and out - degrees of all nodes should stay constant ) ; and ( iii ) that randomize the flows between the nodes , i.e. with different weights of the edges .
such a null model of flows that respects the static organization of the city is indeed more reasonnable and realistic than a null model that would respect the total number of individuals in the matrix but that would modify the in- and out - degrees of the nodes . to generate a random graph that conserves the in- and out- degree of each node of the reference graph
, we use the molloy - reed algorithm @xcite which complexity is in @xmath70 , where @xmath3 is the sum of the weights of the edges ( i.e. the number of individuals in the od case ) . in order to evaluate the sensitivity of the classification of cities to the number of hotspots selected in cities , for each city @xmath5 we make vary the number of work hotspots between @xmath71 the reference value returned by the loubar method ( see supplementary note 3 ) , and two times the reference value @xmath72 : @xmath73 $ ] . as for the sensitivity to noise in the flows @xmath74
, we evaluate the stability of the classification of cities in @xmath75 groups with the jaccard index ( see supplementary note [ sec : icdr - sens ] ) .
the values of @xmath76 as a function of @xmath55 are represented on supplementary figure s13 . | the extraction of a clear and simple footprint of the structure of large , weighted and directed networks is a general problem that has many applications .
an important example is given by origin - destination matrices which contain the complete information on commuting flows , but are difficult to analyze and compare .
we propose here a versatile method which extracts a coarse - grained signature of mobility networks , under the form of a @xmath0 matrix that separates the flows into four categories .
we apply this method to origin - destination matrices extracted from mobile phone data recorded in thirty - one spanish cities .
we show that these cities essentially differ by their proportion of two types of flows : integrated ( between residential and employment hotspots ) and random flows , whose importance increases with city size .
finally the method allows to determine categories of networks , and in the mobility case to classify cities according to their commuting structure .
[ [ sec : intro ] ] the increasing availability of pervasive data in various fields has opened exciting possibilities of renewed quantitative approaches to many phenomena .
this is particularly true for cities and urban systems for which different devices at different scales produce a very large amount of data potentially useful to construct a ` new science of cities ' @xcite .
a new problem we have to solve is then to extract useful information from these huge datasets . in particular , we are interested in extracting coarse - grained information and stylized facts that encode the essence of a phenomenon , and that any reasonable model should reproduce .
such meso - scale information helps us to understand the system , to compare different systems , and also to propose models .
this issue is particularly striking in the study of commuting in urban systems . in transportation research and urban planning
, individuals daily mobility is usually captured in origin - destination ( od ) matrices which contain the flows of individuals going from a point to another ( see @xcite ) .
an od matrix thus encapsulates the complete information about individuals flows in a city , at a given spatial scale and for a specific purpose .
it is a large network , and as such does not provide a clear , synthetic and useful information about the structure of the mobility in the city .
more generally , it is very difficult to extract high - level , synthetic information from large networks and methods such as community detection @xcite and stochastic block modeling ( see for example @xcite and @xcite ) were recently proposed .
both these methods group nodes in clusters according to certain criteria and nodes in a given cluster have similar properties ( for example , in the stochastic block modeling , nodes in a given group have similar neighborhood ) .
these methods are very interesting when one wants to extract meso - scale information from a network , but are unable to construct expressive categories of links and to propose a classification of weighted ( directed ) networks .
this is particularly true in the case of commuting networks in cities , where edges represent flows of individuals that travel daily from their residential neighborhood to their main activity area .
several types of links can be distinguished in these mobility networks , some constitute the backbone of the city by connecting major residential neighborhoods to employment centers , while other flows converge from smaller residential areas to important employment centers , or diverge from major residential neighborhoods to smaller activity areas .
in addition , the spatial properties of these commuting flows are fundamental in cities and a relevant method should be able to take this aspect into account .
there is an important literature in quantitative geography and transportation research that focuses on the morphological comparison of cities @xcite and notably on multiple aspects of polycentrism , ranging from schematic pictures proposed by urban planners and architects @xcite to quantitative case studies and contextualized comparisons of cities @xcite .
so far most comparisons of large sets of cities have been based on morphological indicators @xcite ) built - up areas , residential density , number of sub - centers , etc . and aggregated mobility indicators @xcite motorization rate , average number of trips per day , energy consumption _ per capita _ per transport mode , etc .
, and have focused on the spatial organization of residences and employment centers .
but these previous studies did not propose generic methods to take into account the spatial structure of commuting trips , which consist of both an origin and a destination .
such comparisons based on aggregated indicators thus fail to give an idea of the morphology of the city in terms of daily commuting flows .
we still need some generic methods that are expressive in a urban context , and that could constitute the quantitative equivalent of the schematic pictures of city forms that have been pictured for long by urban planners @xcite . in this paper
we propose a simple and versatile method designed to compare the structure of large , weighted and directed networks . in the next section
we describe this method in detail .
the guiding idea is that a simple and clear picture can be provided by considering the distribution of flows between different types of nodes .
we then apply the method to commuting ( journey to work ) od matrices of thirty - one cities extracted from a large mobile phone dataset .
we discuss the urban spatial patterns that our method reveals , and we compare these patterns observed in empirical data to those obtained with a reasonable null model that generates random commuting networks . finally the method allows determining categories of networks with respect to their structure , and here to classify cities according to their commuting structure .
this classification highlights a clear relation between commuting structure and city size . |
this paper studies existence and uniqueness of a flow for the equation @xmath1 where @xmath2 and @xmath3 are in the whole @xmath4 and @xmath5 is a given function from @xmath4 to @xmath4 .
those are of course newton s equations for a particle moving in a force field @xmath5 . for many applications the force field
is in fact a potential @xmath6 even though we will not use the additional hamiltonian structure that this is providing .
this is a particular case of a system of differential equations @xmath7 with @xmath8 , @xmath9 , @xmath10 .
cauchy - lipschitz theorem applies to and gives maximal solutions if @xmath5 is lipschitz .
those solutions are in particular global in time if for instance @xmath11 .
moreover because of the particular structure of eq . , this solution has the additional for any @xmath12 the application @xmath13 is globally invertible and has jacobian @xmath14 at any @xmath15 .
it also defines a semi - group @xmath16 [ prop : hamilt ] in many cases this lipschitz regularity is too demanding and one would like to have a well posedness theory with a less stringent assumption on @xmath5 .
that is the aim of this paper .
more precisely , we prove [ thm : main ] assume that @xmath17
. then , there exists a solution to ( [ eq : newton ] ) , satisfying property [ prop : hamilt ] .
moreover this solution is unique among all limits of solutions to any regularization of .
many works have already studied the well posedness of eq . under weak conditions for @xmath18 .
the first one was essentially due to diperna and lions @xcite , using the connection between and the transport equation @xmath19 the notion of renormalized solutions for eq .
provided a well posedness theory for under the conditions @xmath20 and @xmath21 .
this theory was generalized in @xcite , @xcite and @xcite . using a slightly different notion of renormalization ,
ambrosio @xcite obtained well posedness with only @xmath22 and @xmath21 ( see also the papers by colombini and lerner @xcite , @xcite for the @xmath23 case ) .
the bounded divergence condition was then slightly relaxed by ambrosio , de lellis and maly in @xcite with then only @xmath24 ( see also @xcite ) .
of course there is certainly a limit to how weakly @xmath18 may be and still provide uniqueness , as shown by the counterexamples of aizenman @xcite and bressan @xcite .
the example by de pauw @xcite even suggests that for the general setting , @xmath23 is probably close to optimal . but as is a very special case of , it should be easier to deal with . and for instance bouchut @xcite got existence and uniqueness to with @xmath25 in a simpler way than @xcite .
hauray @xcite handled a slightly less than @xmath23 case ( @xmath26 ) . in dimension @xmath27 of physical space
( dimension @xmath28 in phase space ) , bouchut and desvillettes proved well posedness for hamiltonian systems ( thus including as @xmath5 is always a derivative in dimension @xmath14 ) without any additional derivative for @xmath5 ( only continuity ) .
this was extended to hamiltonian systems in dimension @xmath28 in phase space with only @xmath29 coefficients in @xcite and even to any system ( non necessarily hamiltonian ) with bounded divergence and continuous coefficient by colombini , crippa and rauch @xcite ( see also @xcite for low dimensional settings and @xcite with a very different goal in mind ) . unfortunately in large dimensions ( more than @xmath14 of physical space or @xmath28 in the phase space ) , the hamiltonian or bounded divergence structure does not help so much . to our knowledge , th .
[ thm : main ] is the first result to require less than @xmath14 derivative on the force field @xmath5 in any dimension . note that the comparison between @xmath0 and @xmath23 is not clear as obviously @xmath30 and @xmath31 . even if one considers the stronger assumption that the force field be in @xmath32 , that space contains by interpolation @xmath33 for @xmath34 and not @xmath0 .
as the proof of th .
[ thm : main ] uses orthogonality arguments , we do not know how to work in spaces non based on @xmath35 norms ( @xmath36 for example ) .
therefore strictly speaking th .
[ thm : main ] is neither stronger nor weaker than previous results .
we have no idea whether this @xmath0 is optimal or in which sense .
it is striking because it already appears in a question concerning the related vlasov equation @xmath37 note that this is the transport equation corresponding to eq .
, just as eq . corresponds to . as a kinetic equation , it has some regularization property namely that the average @xmath38 is more regular than @xmath39 . and precisely if @xmath40 and @xmath41 then @xmath42 ; we refer to golse , lions , perthame and sentis @xcite for this result , diperna , lions , meyer for a more general one @xcite or @xcite for a survey of averaging lemmas
of course we do not know how to use this kind of result for the uniqueness of or even what is the connection between the @xmath0 of averaging lemmas and the one found here . it _
could _ just be a scaling property of those equations .
note in addition that the method chosen for the proof may in fact be itself a limitation .
indeed it relies on an explicit control on the trajectories : for instance , we show that @xmath43 and @xmath44 remain approximately of order @xmath45 if @xmath46 however the example given in section [ example ] demonstrates that such a control in not always possible : even in @xmath47 it requires at least @xmath48 derivative on the force term ( @xmath49 ) whereas well posedness is known with essentially @xmath50 ( see the references above ) .
this kind of control is obviously connected with regularity properties of the flow ( differentiability for instance ) , which were studied in @xcite ( see also @xcite ) . the idea
to prove them directly and then use them for well posedness is quite recent , first by crippa and de lellis in @xcite with the introduction and subsequent bound on the functional @xmath51 this gave existence / uniqueness for eq . with @xmath52 for any @xmath53 and a weaker version of the bounded divergence condition .
this was extended in @xcite and @xcite .
we use here a modified version of which takes the different roles of @xmath2 and @xmath3 into account .
the way of bounding it is also quite different as we essentially try to integrate the oscillations of @xmath5 along a trajectory .
the paper is organized as follows : the next section introduces the functional that is studied , states the bounds that are to be proved and briefly explains the relation with the well posedness result th .
[ thm : main ] .
the section after that presents the example in @xmath47 and the last and longer section the proof of the bound .
[ [ notation ] ] notation + + + + + + + + * @xmath54 denotes the usual scalar product of @xmath55 and @xmath56 .
* @xmath57 denotes the @xmath58-dimensional unit sphere in @xmath4 .
* @xmath59 is the closed ball of @xmath60 for the standard euclidean norm with center @xmath61 and radius @xmath62 .
* @xmath63 denotes a positive constant that may change from line to line .
in the sequel , we give estimates on the flow to eq .
( [ eq : newton ] ) for initial values @xmath64 in a compact subset @xmath65 of @xmath66 and for time @xmath67 $ ] .
fix some @xmath68 and consider any @xmath11 with @xmath69 .
then for any solution to eq .
@xmath70 therefore , for any @xmath67 $ ] and for any @xmath64 at a distance smaller than 1 from @xmath71 , @xmath72 for some compact subset @xmath73 of @xmath66 .
moreover @xmath73 depends only on @xmath71 and @xmath74 .
similarly , we introduce @xmath75 a compact subset of @xmath66 such that the couple @xmath76 belongs to @xmath75 for any @xmath67 $ ] and any @xmath64 at a distance smaller than 1 from @xmath73 . for @xmath77 ,
define the quantity @xmath78 where @xmath79 and @xmath80 are two solutions to , satisfying @xmath81 we prove the following result [ prop : control - q ] fix @xmath77 , any @xmath68 and @xmath82 compact .
define @xmath73 and @xmath75 as in section [ sec : reduc ] .
there exists a constant @xmath83 depending only of @xmath84 , @xmath85 , @xmath86 and @xmath74 , such that , for any @xmath87 , @xmath88 with @xmath89 and any solutions @xmath90 and @xmath91 to satisfying property [ prop : hamilt ] and , one has for any @xmath92 , @xmath93 as will appear in the proof , this result can be actually extended without difficulty to any @xmath94 such that @xmath95 for some function @xmath96 such that @xmath97 when @xmath98 , where @xmath99 is the fourier transform of @xmath5 . we restrict ourselves to prop .
[ prop : control - q ] to simplify the presentation in the proof but this remark means that the following modified proposition holds [ prop : control - qbis ] fix @xmath77 , @xmath68 , @xmath82 compact and any @xmath96 such that @xmath97 when @xmath98 .
define @xmath73 and @xmath75 as in section [ sec : reduc ] .
there exists a continuous , decreasing function @xmath100 with @xmath101 s.t . for any @xmath102 with @xmath89 , for any solutions
@xmath90 and @xmath91 to satisfying property [ prop : hamilt ] and , one has for any @xmath92 , @xmath103 with @xmath104 the fourier transform of @xmath5 .
it is now well known how to pass from an estimate like the one provided by prop .
[ prop : control - q ] to a well posedness theory ( see @xcite for example ) and therefore we only briefly recall the main steps .
take any @xmath105 .
we start by the existence of a solution . for that define @xmath106 a regularizing sequence of @xmath5 .
denote @xmath107 , @xmath108 the solution to with @xmath106 instead of @xmath5 and @xmath109 . for any @xmath110 in @xmath66 , put @xmath111 the function @xmath106 and the solutions @xmath90 , @xmath91 satisfy to all the assumptions of prop .
[ prop : control - q ] , as @xmath112 , using property [ prop : hamilt ] . since @xmath113
, one may choose @xmath106 uniformly bounded in this space .
the proposition then shows that @xmath114 is uniformly bounded in @xmath115 and @xmath116 by @xmath117 by rellich criterion , this proves that the sequence @xmath118 is compact in @xmath119 .
denote by @xmath90 an extracted limit , one directly checks that @xmath90 is a solution to by compactness and satisfies , .
thus existence is proved . for uniqueness ,
consider another solution @xmath91 to , which is the limit of solutions to a regularized equation ( such as the one given by @xmath106 or by another regularizing sequence of @xmath5 ) . then with the same argument
, @xmath91 also satisfies and .
moreover @xmath120 so that @xmath90 and @xmath121 also verify for any @xmath122 . applying again prop .
[ prop : control - q ] and letting @xmath116 go to @xmath123 , one concludes that @xmath124 and @xmath125 .
note from this sketch that one has uniqueness among all solutions to satisfying and and not only those which are limit of a regularized problem .
however not all solutions to ( pointwise ) necessarily satisfy those two conditions so that the uniqueness among all solutions to is unknown .
indeed in many cases , it is not true , as there is a hidden selection principle in ( see the discussion in @xcite , @xcite or @xcite ) .
finally if @xmath126 only , then one first applies the de la valle poussin s lemma to find a function @xmath39 s.t.@xmath127 when @xmath98 and @xmath128 one proceeds as before with a regularizing sequence @xmath106 which now has to satisfy uniformly the previous estimate . using prop .
[ prop : control - qbis ] instead of prop .
[ prop : control - q ] , the rest of the proof is identical .
it is hard to know whether the condition @xmath126 is optimal and in which sense ( see the short discussion in the introduction ) .
instead the purpose of this section is to give a simple example showing that @xmath129 is a necessary condition in order to use the method followed in this paper ; namely a quantitative estimate on @xmath130 and @xmath131 .
more precisely , for any @xmath132 , we are going to construct a sequence of force fields @xmath133 uniformly bounded in @xmath134 and a sequence @xmath135 converging to 0 such that functionals like @xmath136 can not be uniformly bounded in @xmath137 .
this example is one dimensional ( @xmath28 in phase space ) where it is known that much less is required to have uniqueness of the flow ( almost @xmath5 a measure ) .
so this indicates in a sense that the method itself is surely not optimal .
moreover what this should imply in higher dimensions is not clear ... through all this section we use the notation @xmath138 if there exists a constant @xmath63 s.t .
@xmath139 in dimension @xmath14 all @xmath5 derive from a potential so take @xmath140 with @xmath141 a periodic and regular function ( @xmath142 at least ) with @xmath143 .
as @xmath144 is regular , we know that the solution @xmath90 with initial condition @xmath64 and the shifted one @xmath91 corresponding to the initial condition @xmath145 satisfy the conservation of energy or @xmath146 as @xmath144 is defined up to a constant , we do not need to look at all the trajectories and may instead restrict ourselves to the one starting at @xmath2 s.t .
@xmath147 . by symmetry
, we may assume @xmath148 and excluding the negligible set of initial data with @xmath149 , we may even take @xmath150 .
let @xmath151 and @xmath152 be the first times when the trajectories stop increasing : @xmath153 and @xmath154 .
as both velocities are initially positive , they stay so until @xmath151 or @xmath152 .
so for instance @xmath155 hence @xmath151 is obtained by @xmath156 if @xmath157 .
of course by energy conservation @xmath158 and again as we are in dimension @xmath14 this means that we may simply take @xmath159 .
[ fig : ex ] and the construction of @xmath160 and @xmath161,width=377 ] we have the equivalent formula for @xmath152 with @xmath162 ( which we may not assume equal to @xmath123 ) . put @xmath163 and note that @xmath164 , so that @xmath165 since @xmath166 for @xmath137 large enough
. then @xmath167 as the integral between @xmath2 and @xmath168 is bounded by @xmath169 ( the integrand is bounded here ) and @xmath170 note that as @xmath141 is lipschitz regular @xmath171 so subtracting the two formula and making an asymptotic expansion @xmath172 making the change of variable @xmath173 in the dominant term in the integral , one finds @xmath174 consequently as long as @xmath175 is of order @xmath14 then @xmath176 is of order @xmath177 .
note that @xmath178 is small when @xmath179 is , but it is always possible to find functions @xmath141 s.t .
@xmath178 is of order @xmath14 at least for some @xmath179 .
one way to see this is by observing that @xmath180 can not vanish for all @xmath179 and functions @xmath141 .
taking @xmath141 such that @xmath181 for @xmath179 in some non - trivial interval , we can assume that @xmath74 is of order @xmath14 for @xmath182 $ ] for some @xmath183 .
coming back to the definition of @xmath179 and @xmath162 , @xmath182 $ ] is equivalent to @xmath184).\ ] ] using the formula for @xmath144 and the fact that @xmath185 and @xmath186 are independent of @xmath137 or @xmath116 , we find @xmath187.\ ] ] so let us finally choose @xmath188 and denote by @xmath189 the space of initial velocities @xmath3 s.t
. is satisfied for @xmath137 large enough . in view of the previous computation
, there exists @xmath190 and @xmath191 such that for all @xmath192 and all @xmath193 , @xmath194 we consider now the rest of the trajectories after times @xmath151 and @xmath152 . to this aim ,
we denote by @xmath195 and @xmath196 the solution of with initial data @xmath197 . by uniqueness @xmath198
obviously one can not have @xmath199 small for all times , as initially @xmath200 was not small , and , as the force field @xmath201 is bounded , @xmath202 is lipschitz in time .
so in conclusion for any @xmath203 , there exists a time interval @xmath204 of length of order @xmath3 where @xmath202 is larger than @xmath205 .
moreover @xmath206=[\underline{\eta}/n,\bar{\eta}/n]$ ] .
now either there exists a time interval @xmath207 of order @xmath3 s.t .
@xmath208 or if it is not the case then on a subset @xmath209 of @xmath210 of size @xmath3 , one has @xmath211 note that @xmath151 may be replaced by @xmath152 in the previous inequality by reducing the interval @xmath212 ( while keeping its length of order 1 ) since @xmath213 .
therefore for @xmath214 @xmath215 as @xmath202 is larger than @xmath205 .
consequently in both situations , we have two solutions , @xmath216 and @xmath217 or @xmath90 and @xmath91 , distant of @xmath218 initially but distant of order @xmath177 on a time interval of order @xmath3 .
since @xmath141 is periodic this provides many initial conditions with such a property .
the difficulty that the distance between @xmath160 and @xmath162 is not fixed can be overcome since we are in two - dimensional setting ( another trajectory starting further than @xmath162 from @xmath160 can not approach more @xmath216 than @xmath217 does ) .
therefore we may control a functionals like @xmath219 with @xmath220 uniformly in @xmath137 only if @xmath221 since @xmath188 , this requires @xmath222 or @xmath223 in at least @xmath224 as claimed .
recall the notation @xmath104 for the fourier transform of @xmath5 .
the assumption of proposition [ prop : control - q ] corresponds to the following bound : @xmath225 we introduce a @xmath232 function @xmath233 $ ] such that @xmath234 if @xmath235 and @xmath236 if @xmath237 . writing @xmath238 ( resp .
@xmath239 ) for @xmath240 ( resp . @xmath241 ) and @xmath242 ( resp .
@xmath243 ) for @xmath244 ( resp.@xmath245 ) , and introducing @xmath246 we may write @xmath247 where @xmath248 @xmath249 @xmath250 and @xmath251 the proof is based on a control each of these terms . as proved in subsection [ sec : q^4 ] , the fourth term can be bounded with elementary computations . in subsection
[ sec : q^2-q^3 ] , the second and third terms are bounded using standard results on maximal functions . finally , the control of @xmath252 requires a more precise version of the maximal inequality , detailed in subsection [ sec : q^1 ] .
let us define the function @xmath256 since @xmath257 , we have @xmath258 where the second line follows from lemma [ lem:1 ] and from property [ prop : hamilt ] applied to the change of variables @xmath259 and @xmath260 .
then , it follows from plancherel s identity that @xmath261 we recall that the maximal function @xmath264 of @xmath265 , @xmath266 , is defined by @xmath267 we are going to use the following classical results ( see @xcite ) .
first , there exists a constant @xmath63 such that , for all @xmath268 and @xmath265 , @xmath269 second , for all @xmath270 , the operator @xmath271 is a linear continuous application from @xmath272 to itself .
we begin with the control of @xmath263 .
let @xmath273 it follows from the previous inequality that @xmath274 therefore , since @xmath275 if @xmath276 , following the same steps as for the contro , of @xmath253 , @xmath277 then @xmath278 for any @xmath268 , we introduce the hyperplane orthogonal to @xmath293 @xmath294 if @xmath295 , we define for example @xmath296 , where @xmath297 .
fix a @xmath298 function @xmath299 such that @xmath300 for @xmath301 $ ] and @xmath302 . by invariance of @xmath303 with respect to rotations
, we also have @xmath304 for all @xmath268 . observing that @xmath324 this set
may also be defined as @xmath325 \right .
\left.\mbox{\ and\ } \left|\frac{y}{|x| } -\left(\frac{y}{|x|}\cdot\frac{x}{|x|}\right ) \frac{x}{|x|}\right|\leq 4\frac{y}{|x|}\cdot\frac{x}{|x| } \left(1-\frac{y}{|x|}\cdot\frac{x}{|x|}\right)\right\}.\end{gathered}\ ] ] note that , for any @xmath326 , taking @xmath291 and @xmath141 as in ( [ eq : def - cone ] ) , we have @xmath327 . therefore , denoting by @xmath328 the angle between the vectors @xmath2 and @xmath329 , @xmath330 for fixed @xmath268
, we now introduce the application @xmath331\times\{h\in h(x , y ) : |h|\leq|y - x|\ } \rightarrow k(x - y ) \\ & ( \theta , h)\mapsto \theta ( x - y+4(1-\theta)h).\end{aligned}\ ] ] it is elementary to check that @xmath332 is a bijection when @xmath333 , with inverse @xmath334 for @xmath335 .
moreover , @xmath332 is differentiable and its differential , written in a basis of @xmath4 with first vector @xmath336 , is @xmath337 therefore , the jacobian of @xmath332 at @xmath338 is @xmath339 . making the change of variable @xmath340 in ( [ eq : def - b ] )
, we can now compute @xmath341 with @xmath342 and @xmath343 we defined , for @xmath344 , @xmath345 and @xmath346 where @xmath347 if @xmath348 $ ] , and @xmath349 otherwise .
it follows from ( [ eq : angle ] ) and from the definition of @xmath350 that these two functions have support in @xmath351.\ ] ] moreover , they belong to @xmath352 . indeed , since @xmath349 for @xmath353 , the terms @xmath354 in the denominators do not cause any regularity problem
moreover , since @xmath300 for @xmath301 $ ] and @xmath355 for all @xmath356 , the terms @xmath357 in the denominators do not cause any worry either . finally , since @xmath358 , the discontinuity of @xmath359 at @xmath123 can only cause a problem in the neighborhood of points such that @xmath360 ( @xmath361 being nonzero ) .
therefore , the previous observation also solves this difficulty .
writing @xmath363 for @xmath364 we decompose @xmath365 @xmath366 now , let us write @xmath367 for @xmath368 and let us define similarly as in ( [ eq : def - psi1_s ] ) and ( [ eq : def - chi_s ] ) the notation @xmath369 , @xmath370 and @xmath371 . note that the term @xmath372 is exactly the time derivative of @xmath373 .
so integrating by parts in time , we obtain @xmath374 with @xmath375 @xmath376 @xmath377 correspondingly @xmath378 \:dk\:dz\:ds , \end{split}\ ] ] and @xmath379 let us define @xmath380 and @xmath381 , @xmath382 , @xmath383 and @xmath384 similarly .
first , we make the change of variables @xmath388 , followed by the change of variable @xmath389 in the integral defining @xmath384 .
when @xmath390 , the variable @xmath391 belongs to the set @xmath392 .
note also that @xmath393 and @xmath394 .
writing for convenience @xmath395 instead of @xmath396 , it follows from these changes of variables and from property [ prop : hamilt ] that @xmath397 where @xmath398 and @xmath399 writing the tensor ( remember that @xmath400 ) @xmath401 and reminding that @xmath402 for all @xmath403 $ ] , we have @xmath404 where @xmath405 for any tensor with three entries @xmath406 with @xmath407 .
so @xmath408 now , on the one hand , following the same computation as in lemma [ lem:1 ] , the integral with respect to @xmath409 can be upper bounded by @xmath410 for @xmath45 small enough .
on the other hand , @xmath411 and this last term is bounded by @xmath412 where we have used that , for any @xmath413 , @xmath414 , and where the last inequality can be obtained by a spherical coordinate change of variable centered at @xmath2 in the variable @xmath415 in the first term , and centered at @xmath415 in the variable @xmath2 in the second term .
now , @xmath416 and integrating first in @xmath415 and @xmath417 , this is equal to @xmath418 therefore @xmath419 where we write the vector @xmath3 as @xmath420 in an orthonormal basis of @xmath4 with first vector @xmath421 . in conclusion
@xmath422 combining this inequality with ( [ eq : bourrin ] ) , we finally get @xmath423 applying to @xmath425 the same change of variable as we did for @xmath426 , we have @xmath427 where @xmath428 for any matrix @xmath429 , @xmath430 and @xmath431 note that , because of the properties of @xmath432 obtained in section [ sec : cv - z ] , @xmath433 for some constant @xmath63 .
now , we observe from ( [ eq : def - b1-b2 ] ) that @xmath465 has exactly the same structure as @xmath466 : a singularity of order @xmath58 in @xmath415 , a function @xmath467 that has all the required regularity , and a term @xmath468 .
it is then easy to see that this term can be treated by exactly the same method as @xmath466 .
we leave the details to the reader .
finally , the term @xmath471 of ( [ eq : def - c ] ) can be bounded exactly as @xmath318 by simply exchanging the roles of @xmath472 and @xmath473 .
therefore , the proof of proposition [ prop : control - q ] is completed .
l. ambrosio , g. crippa , existence , uniqueness , stability and differentiability properties of the flow associated to weakly differentiable vector fields .
lecture notes of the unione matematica italiana , springer verlag , to appear .
l. ambrosio , c. de lellis , j. mal , on the chain rule for the divergence of vector fields : applications , partial results , open problems , perspectives in nonlinear partial differential equations , 3167 , _ contemp .
_ , * 446 * , amer .
soc . , providence , ri , 2007 .
f. colombini , n. lerner , uniqueness of @xmath474 solutions for a class of conormal @xmath23 vector fields . geometric analysis of pde and several complex variables , 133156 , _ contemp .
_ * 368 * , amer .
soc . , providence , ri , 2005 .
stein , harmonic analysis : real - variable methods , orthogonality , and oscillatory integrals .
princeton mathematical series , 43 .
monographs in harmonic analysis , iii .
_ princeton university press , princeton , nj _ , 1993 . | we study existence and uniqueness for the classical dynamics of a particle in a force field in the phase space . through an explicit control on the regularity of the trajectories ,
we show that this is well posed if the force belongs to the sobolev space @xmath0 .
_ msc 2000 subject classifications : 34c11 , 34a12 , 34a36 , 35l45 , 37c10 _ _ key words and phrases : flows for ordinary differential equations , kinetic equations , stability estimates _ |
in 1966 , marc kac offered a poetic formulation of what came to be a much studied problem involving the geometry of planar domains : does dirichlet spectrum determine a planar domain up to isometry ?
kac s problem inspired a great deal of work much of which centered on developments involving heat invariants .
the appearance of sunada s work involving the construction of isospectral metrics for riemannian manifolds provided powerful new methods for investigating diverse spectral phenomena ; in particular , sunada s method led to the discovery of a number of negative results for inverse spectral problems . in 1992
gordon , web and wolpert used an extension of sunada s method to construct a pair of isospectral , non - isometric planar domains @xcite .
soon after , buser , conway doyle and semmler produced a collection of families of pairs of isospectral non - isometric planar polygonal domains ( we refer to these examples as _ bcds pairs _ ) . their construction ,
again going back to sunada , involves reflecting a `` seed '' triangle across edges ( see @xcite ) .
an example of their construction can be found in section 3 below and plays an important role in our work .
given these negative results in reference to kac s question , one might ask : what are good invariants for planar domains and/or quantum graphs ?
one approach to finding new invariants is to use data which , while not spectral , arise from the normalized dirichlet eigenfunctions in a natural fashion .
an example is afforded by the sequence generated by counting nodal domains .
this approach has been investigated in a recent series of papers in which quantum graph analogs of bcds pairs are produced and their corresponding nodal domain counts are compared ( see @xcite , @xcite ) . another approach to producing non - spectral invariants
is given by integration against the heat kernel .
more precisely , suppose @xmath0 is the dirichlet heat kernel for the domain @xmath1 and @xmath2 is the solution of the inital value problem @xmath3 then one can define the _ heat content of @xmath1 _ @xmath4 as well as a sequence of invariants @xmath5 the invariant @xmath6 sometimes referred to as the _ torsional rigidity of the domain @xmath7 _ can be computed via the solution of a poisson problem .
more precisely , if @xmath8 solves @xmath9 then @xmath10 torsional rigidity arises in the study of elasticity and has a long history ( a source for early work is @xcite ) .
in addition to its role in the mechanics of solid bodies , torsional rigidity is closely related to the expected exit time of brownian motion from the domain @xmath1 and higher values of @xmath11 in ( [ ak1 ] ) correspond to higher moments of the exit time ( cf @xcite , @xcite and references within ) .
for this reason the sequence @xmath12 defined by ( [ ak1 ] ) has been labeled the _
@xmath13-moment spectrum _ associated to @xmath14 it is easy to see that both heat content and the moment spectrum are invariants of the metric .
recent work suggest they play a valuable role in understanding fundamental geometric properties of a given ambient space ( for example , isoperimetry @xcite @xcite , @xcite ) , and one might reasonably ask how such objects compare to dirichlet spectrum as tools to classify domain behavior up to isometry . such a study has been initiated in @xcite .
for the purpose of this note , the main results of @xcite can be summarized as follows : 1 . for smoothly bounded domains ,
the moment spectrum determines the heat content .
2 . for generic smoothly bounded domains ,
the moment spectrum determines the dirichlet spectrum . it is a result of gilkey that heat content can not distinguish isospectral planar domains constructed via the sunada method @xcite . on the other hand , in @xcite
, the authors establish that heat content distinguishes the isosspectral domains constructed by chapman and in @xcite the authors construct combinatorial analogs of bcds pairs and explicit check that their heat contents differ ( at the fifth coefficient ) .
the main result of this paper is that with respect to heat content , quantum graph analogs of bcds pairs behave more like combinatorial graphs than planar domains , and that torsional rigidity is sufficient to distinguish well - known examples of isospectral pairs .
more precisely , we have [ theorem1]let @xmath15 and @xmath16 be the @xmath17-pair of isopectral non - isometric planar polygonal domains as constructed by buser , conway , doyle and semmler and let @xmath18 and @xmath19 be their quantum graph analogs ( cf section 3 below )
. then @xmath18 and @xmath19 are isospectral , non - isometric and distinguished by their torsional rigidity .
in addition to establishing theorem 1 , we prove [ theorem2 ] the moment spectrum of a quantum graph with dirichlet standard boundary conditions ( cf section 2 below ) determines the heat content of the quantum graph . as an immediate corollary
, we conclude that our isospectral non - isometric pairs are distinguished by their heat content .
our technique applies to the other families of isospectral pairs constructed in @xcite .
the remainder of this note is organized as follows . in the second section
we establish the required background information concerning quantum graphs , establish our notational conventions and prove theorem [ theorem2 ] . in the third section we review the sunada construction in the context of planar domains following buser , conway , doyle and semmler and include the construction of isospectral non - isometric quantum graph analogs of bcds pairs . in the fourth section of the paper
we explicitly compute the torsional rigidity for a pair of isospectral non - isometric quantum graphs and check they differ .
we end the paper with a previously studied example ( @xcite ) of a pair of isospectral non - isomorphic weighted combinatorial graphs which arise as analogs of bcds pairs and check that they too are distinguished by their torsional rigidity .
let @xmath20 be a graph with finite vertex set @xmath21 and edge set @xmath22 for @xmath23 denote by @xmath24 the cardinality of the set @xmath25 we will often find it convenient to assign an orientation to edges ; that is , for each edge @xmath26 choose @xmath27 to be either @xmath28 or @xmath29 by a _ path _ in @xmath20 we will mean a sequence of vertices , @xmath30 with @xmath31 we say @xmath20 is connected if every pair of vertices can be connected by a path in @xmath32 we restrict our attention to graphs which are connected and contain no edges of the form @xmath33 we endow @xmath20 with a metric structure as follows .
for each edge @xmath34 choose a positive real number @xmath35 identify each edge @xmath36 with the interval @xmath37,$ ] where @xmath38 is identified with @xmath27 and @xmath39 is identified with the remaining vertex of @xmath40 this identification allows us to introduce a natural coordinate on each edge ; the coordinate @xmath41 along the interval @xmath37.$ ] these coordinates give rise to a natural metric structure for @xmath32 the pair @xmath42 is called a _
metric graph_. because @xmath20 and @xmath39 are finite , the resulting metric graph is compact .
we identify functions on @xmath42 with functions along the open edges together with values at each vertex : for @xmath43 we write @xmath44 the metric structure on each edge gives rise to a natural hilbert space associated to @xmath45 write @xmath46)\ ] ] where the inner product associated to edge @xmath47 is defined by @xmath48 let @xmath49 and denote the inner product on @xmath50 by @xmath51 in particular , @xmath52 where @xmath53 is the total length of the graph @xmath32 there is a natural differential operator acting on functions on the interior of each edge : @xmath54 we wish to extend this operator to a self - adjoint operator on @xmath55-functions on the metric graph .
we sketch the approach developed in @xcite ( see @xcite for details ) . for each @xmath34
let @xmath56 denote @xmath57 let @xmath58 and set @xmath59 then the operator @xmath60 which acts on @xmath61 according to @xmath62 is closed , symmetric and densely defined .
we seek to impose boundary conditions which give self - adjoint extensions of @xmath60 to @xmath63 we begin by noting that boundary values of functions along edges lie in the space @xmath64 where the order of the components is fixed by the orientation .
more precisely , given @xmath65 let @xmath66 be given by @xmath67 boundary conditions can then be described by a pair of linear operators @xmath68 satisfying @xmath69 those pairs of linear operators @xmath70 leading to self - adjoint extensions of @xmath60 have been classified by kostrykin , potthoff and schrader @xcite using a symplectic formalism going back to at least to novikov .
we are interested in a single special case , _ dirichlet standard boundary conditions , _ which we now describe .
[ sbc ] denote by dsbc the collection of continuous functions defined by @xmath71 where the derivative occurring in ( [ standard ] ) is always directed into the vertex @xmath72 and , as before , @xmath73 is the collection of edges on which @xmath72 is incident .
we note that the literature often refers to `` standard boundary conditions '' as those involving kirchoff conditions at internal vertices ( ie vertices of degree at least 2 ) and neumann conditions at boundary vertices ( ie vertices of degree 1 ) .
we can describe standard boundary conditions using pairs of matrices @xmath70 as follows : for each vertex @xmath72 with @xmath74 introduce @xmath75 matrices @xmath76 and @xmath77 ( cf also @xcite , @xcite ) @xmath78 for vertices of degree 1 , define @xmath79 matrices @xmath80 and @xmath81 the matrices @xmath76 and @xmath77 referenced above then take a block form : @xmath82 we note that dirichlet standard boundary conditions are local in the sense of @xcite and do not mix derivatives with function values .
we note that dirichlet standard boundary conditions force continuity across all internal vertices for all elements in the corresponding domain of the self - adjoint extension of @xmath83 thus , it is the choice of boundary conditions which ultimately ties the geometry and topology of the quantum graph to the analysis of the corresponding laplace operator .
we will denote the self - adjoint extension of @xmath60 with dirichlet standard boundary conditions as @xmath84 the pair consisting of the metric graph @xmath20 and the operator @xmath85 is the quantum graph we study . the spectrum of @xmath85 is discrete and of finite multiplicity .
we will denote the spectrum of @xmath85 by @xmath86 we will assume there is at least one boundary vertex and we will list elements of @xmath87 in increasing order with multiplicity : @xmath88 there is a great deal known about the spectrum for arbitrary self - adjoint boundary conditions . for our purposes it is important that there is a heat kernel which can be written as @xmath89 the heat content of @xmath20 can then be defined via integration as for domains : @xmath90 we let @xmath91 denote the set of values defined by the dirichlet standard boundary conditions ( ie we disregard multiplicity ) for which the corresponding eigenspace is not orthogonal to constant functions .
then we can write the heat content in a fashion which makes explicit its non - spectral nature : @xmath92 where @xmath93 is the square of the @xmath55-norm of the projection of the constant function 1 on the eigenspace defined by @xmath94 the moments of @xmath95 define a sequence associated to the quantum graph @xmath32 more precisely , as in the introduction , for @xmath11 a natural number , set @xmath96 the moment spectrum of the graph @xmath20 is the sequence @xmath97 from the definition it is clear that the heat content determines the moment spectrum .
the converse ( theorem [ theorem2 ] above ) is also true : the moment spectrum of the quantum graph @xmath20 determines the heat content of the quantum graph @xmath32 proof the proof follows the case of domains , where the same result holds ( see @xcite , @xcite ) . for completeness
we provide the argument given in @xcite .
it suffices to show that the moment spectrum determines both @xmath91 and the collection of coefficients @xmath98 the important observation is that the moment spectrum describes special values of the mellin transform of the heat content .
more precisely , using ( [ heatc2 ] ) , the mellin transform of the heat content is given by the dirichlet series @xmath99 which converges for real part of @xmath100 nonnegative and admits a meromorphic extension to the plane , with poles at the negative half - integers ( see @xcite ) .
the moment spectrum then satisfies @xmath101 using ( [ zeta1 ] ) and ( [ zeta2 ] ) , there is a recursion : write @xmath102 with @xmath103 strictly increasing .
then @xmath104 and @xmath105 having determined @xmath106 and @xmath107 for @xmath108 we have @xmath109 and @xmath110 this shows that both @xmath91 and the length partition of @xmath20 are determined by the @xmath13-moment spectrum . using the spectral representation of the heat content ( [ heatc1 ] ) ,
it is clear that @xmath91 and the coefficients @xmath93 determine the heat content .
this proves that the @xmath13-moment spectrum determines heat content and completes the proof of theorem [ theorem2 ] .
our primary interest is in @xmath6 which , as in the case of smooth domains , can be described via a solution of a poisson problem on the graph @xmath32 more precisely , let @xmath8 solve @xmath111 we can apply fubini s theorem to interchange the order of integration in the definition of the heat content and @xmath112 to obtain @xmath113 given any quantum graph with dirichlet standard boundary conditions , the computation of the right - hand - side of ( [ tt2 ] ) is an exercise in linear algebra . we compute torsional rigidity for certain pairs of quantum graphs constructed in the next section .
as mentioned in the introduction , buser , conway , doyle and semmler have constructed families of pairs of isospectral planar domains that are not isometric .
their construction and investigation of the resulting pairs is spectacularly simple . to construct families of pairs of isospectral planar domains , *
fix a seed triangle with edges labeled ; * reflect the seed about its edges to produce a second generation of labeled triangles ; * for each of the progeny , reflect and label across edges determined by two rules derived from subgroups which are `` nearly conjugate '' in sunada s sense * iterate * interpret the edges as distinct lengths of an acute triangle .
an example of the construction is given in figure 1 : the so - called @xmath17 pairs ( for complete details see @xcite ) .
[ fig1 ] having constructed families of pairs as above , buser et al prove that the associated pairs are isospectral using the technique of eigenfunction transplantation .
given an eigenfunction @xmath114 on one domain , say @xmath115 they use the process by which the domain is constructed to naturally partition the domain into a collection of isometric subdomains , say @xmath116 where @xmath117 indexes the number of triangles occurring in the construction , say @xmath118 the same procedure carried out on the domain pair , say @xmath119 results in a collection @xmath120 of isometric subdomains of @xmath121 having partitioned both domains into collections of subdomains , they restrict the given eigenfunction to each subdomain to produce an associated collection of functions , @xmath122 on isometric copies of the same subdomain .
they then use the geometry of the underlying domains @xmath15 and @xmath16 ( ie , how the domains are constructed from the collections of subdomains ) to construct a linear map which prescribes how to build an associated eigenfunction on the domain @xmath16 by combining the function elements @xmath123 on each subdomain of @xmath121 the process by which an explicit transplantation map is constructed is described neatly in the references @xcite and @xcite . for our purposes
it is important to note that the process is essentially combinatorial in the sense that it depends only on the following : * the eigenfunction @xmath114 vanishes at the boundary of @xmath124 * to smoothly continue the eigenfunction @xmath114 across the boundary of the domain @xmath115 the reflection principle requires that the value of @xmath125 be prescribed to be @xmath126 where @xmath127 is the reflection image of @xmath128 * if @xmath129 and @xmath130 are subdomains which share a boundary internal to @xmath115 the associated eigenfunction elements @xmath131 and @xmath132 and their respective normal derivatives must match along the common boundary .
there are also techniques to produce _ quantum graph _
analogs @xcite , @xcite , @xcite . to proceed , we recall the required definitions .
following earlier work ( cf @xcite , @xcite , @xcite and references therein ) band , parzanchevski and ben - shach have described an extension of eigenfunction transplantation in the context of quantum graphs @xcite . they provide a method which associates to every family of pairs constructed in @xcite , a family of pairs of quantum graphs which are isospectral but not isometric .
the construction technique is straightforward , as is the check that the objects satisfy the required conditions .
we sketch the process below and produce the pairs associated to the @xmath17 pairs in figure 2 .
for each pair of isospectral nonisometric domains found in @xcite , * for each triangle appearing in the bcds construction , introduce a 3-star graph consisting of three edges , with edges labeled with a label from the edges of the corresponding triangle , joined at a central vertex . *
join edges in 3-stars at a common degree two vertex if corresponding edges in the triangles to which they are associated overlap .
we carry out the process on quantum graphs corresponding to @xmath17 pairs in figure [ fig3 ] .
note that for the graphs to be isometric , it would be necessary to map the 3-stars corresponding to generating triangles onto each other .
[ fig2 ] for neumann standard boundary conditions , the quantum graph analogs are shown to be isospectral and non - isometric in @xcite ( see also @xcite ) .
this is accomplished by showing that the graphs share the same secular equation .
the same argument can be made to work in the current context , but instead we provide an explicit transplantation map .
in fact , it s easy to check that the transplantation map used to establish that the bcds pairs are isometric works to show that the associated quantum graphs pairs are isometric .
we do this as follows : * label the edges of the quantum graph @xmath18 and the corresponding partner @xmath133 * given an eigenfunction on @xmath134 restrict to 3-star subgraphs to obtain a collection of functions defined on isometric 3-stars .
* use the linear map constructed following @xcite to construct the required transplantation map taking eigenfunctions of @xmath18 to eigenfunctions of @xmath133 we can formalize these observations with [ 71pairs]the quantum graph ( with dirichlet standard boundary conditions ) analogs of the pairs constructed in @xcite are isospectral and non - isometric .
_ proof _ by construction , for each domain constructed in @xcite the corresponding quantum graph is obtained by replacing subdomains with 3-stars , gluing 3-stars along boundaries which correspond to boundaries which are glued in subdomains .
the transplantation map for domains may be represented by a matrix acting on function elements defined on subdomains , all of which can be identified .
this defines a linear map from eigenfunctions on @xmath18 to functions on @xmath135 which we will write as @xmath136 give an eigenfunction @xmath137 on @xmath18 corresponding to eigenvalue @xmath138 to see that the image @xmath139 is an eigenfunction on @xmath19 first note that on the interior of every edge , @xmath140 by linearity .
thus , it suffices to check that @xmath139 has value zero at the boundary of @xmath19 and that @xmath139 behaves as it should across glued edges .
but for this to be the case requires only the three properties listed in our description of the transplantation process given above .
since the gluing relations for quantum graphs are precisely those that are prescribed by the domains and the boundary conditions are dirichlet in both cases , these three properties must hold for the image of @xmath137 under @xmath136 in particular , @xmath139 must be an eigenfunction corresponding to eigenvalue @xmath94 we give the required transplantation map in figure 3 .
we are now in a position to prove theorem [ theorem1 ] for the @xmath17 pairs of figure 2 .
we begin by noting that we can eliminate vertices of degree 2 and replace the corresponding two edges by a single edge of length equal to the sum of the corresponding edges ( following friedlander @xcite , we call such quantum graphs _ clean _ ) .
this results in the pair of quantum graphs given in figure 4 .
note that there is a symmetry : interchanging @xmath141 and @xmath142 in the two graphs maps one graph onto the other .
we use this symmetry in the computation of the torsional rigidity of each graph .
suppose we denote by @xmath143 for the coordinate on the @xmath144th edge of the @xmath117th graph .
then we must solve @xmath145 with boundary conditions determined by our choice of dirichlet standard boundary conditions .
the solution on each edge is of the form @xmath146 where the constants @xmath147 and @xmath148 are to be determined from the dirichlet standard boundary conditions .
thus , for each graph there are 30 constants to be determined using nine dirichlet boundary conditions at vertices of degree 1 , seven kirchoff boundary conditions at vertices of degree 3 , and fourteen continuity conditions at vertices of degree 3 .
this determines a linear system @xmath149 where the coefficient matrix @xmath150 is a sparse @xmath151 matrix and the constant vector @xmath152 is defined in terms of continuity constraints and conservation at internal nodes .
the structure of the matrix will be determined by our choice of edge labels and orientation .
we choose the first nine edges to be those incident to boundary vertices and we orient the edges to be inward facing relative to the boundary . with this choice the coefficients @xmath153 for @xmath154 and @xmath155 with the remaining edges oriented to point toward the `` central node ''
, the representation of @xmath150 for @xmath18 as pictured in figure 4 , is given by the coefficient matrix @xmath156 @xmath157 the continuity and conservation constraints define a vector , @xmath158 given by @xmath159 where `` @xmath160 '' denotes transpose . to obtain the data for the second graph ,
simply interchange @xmath141 and @xmath161 we can use these systems to carry out the required calculation @xmath162 \nonumber \\ & = \frac{1}{2 } \sum_{j=1}^{15 } \left(\alpha_{1,j } l_{1,j}^2 - \alpha_{2,j } l_{2,j}^2 \right ) + \sum_{j=1}^{15}\left ( \beta_{1,j}l_{1,j } - \beta_{2,j}l_{2,j}\right ) \nonumber \\ & = \langle l_1^{-1}{\bf v}_1,{\bf l}_1\rangle - \langle l_2^{-1}{\bf v}_2,{\bf l}_2 \rangle \label{tor33}\end{aligned}\ ] ] where the vectors @xmath163 are constructed from the lengths of the edges of the graphs @xmath164 @xmath165 and @xmath166 is obtained from @xmath167 by interchanging @xmath141 and @xmath161 the right hand side of ( [ tor33 ] ) is a rational function in the variables @xmath168 @xmath169 and @xmath170 @xmath171 where the rational function @xmath172 has a sign : @xmath173 from ( [ sign1 ] ) and the explicit expression of @xmath174 it is easy to see that the difference in torsional rigidities is nonzero whenever @xmath168 @xmath141 and @xmath142 are distinct .
this proves theorem [ theorem1 ] . as a corollary of theorem [ theorem1 ] and theorem [ theorem2 ]
, we obtain : the quantum graphs @xmath18 and @xmath19 are distinguished by their heat content . [ fig4 ]
in @xcite the authors construct combinatorial weighted graph analogs of bcds pairs .
the result of these constructions for the @xmath17-pairs given above are the two isospectral , non - isomorphic weighted combinatorial graphs given in figure 5 : there is a natural weighted laplacian associated to such weighted graphs . for functions on the vertices
we write @xmath175 where the sum is over vertices @xmath176 adjacent to @xmath127 and @xmath177 denotes the weight along the edge defined by the vertices @xmath127 and @xmath178 for the above two graphs this results in laplace operators given by @xmath179 it is easy to check that these matrices have the same characteristic equation . to compute the difference in torsional rigidity for the corresponding combinatorial graphs we let @xmath180 and we compute : @xmath181 where @xmath182 is the rational function defined by @xmath183 in particular , for distinct values of @xmath168 @xmath169 and @xmath142 the isospectral non - isomorphic combinatorial graphs constructed in @xcite are distingusihed by their torsional rigidity .
v. kostrykin , j. potthoff and r. schrader ( 2007 ) _ heat kernels on metric graphs and a trace formula .
_ adventures in mathematical physics , 175198 , contemp . math .
, 447 , amer .
soc . , providence , ri .
v. kostrykin , j. potthoff and r. schrader ( 2008 ) _ contraction semigroups on metric graphs .
_ analysis on graphs and its applications , 423458 , proc .
pure math .
, 77 , amer .
soc . , providence , ri v. kostrykin , j. potthoff and r. schrader ( 2005 ) _
laplacians on metric graphs : eigenvalues resolvents and semigroups _ proceeding of the conference on quantum graphs and their applications , amer .
soc . , providence , ri m. solomyak ( 2002 ) _ on the eigenvalue estimates for the weighted laplacian on metric graphs .
_ nonlinear problems in mathematical physics and related topics , i , int . math .
1 , kluwer / plenum , new york , 327347 . | we study torsional rigidity for graph and quantum graph analogs of well - known pairs of isospectral non - isometric planar domains . we prove that such isospectral pairs are distinguished by torsional rigidity . |
several large - scale gravitational wave ( gw ) interferometers have achieved long term operation at design sensitivity .
different detectors can have very different noise levels and different frequency bandwidth .
the directional sensitivity of different detectors can also be very different .
for instance , the most sensitive gw detectors in the us , namely the ligo detector at livingston , louisiana , and the two co - located detectors at hanford , washington ( abbreviated as l1 and h1/h2 ) , are designed to be nearly aligned .
detectors in europe ( geo600 in germany , and virgo in italy ) and in asia ( tama in japan ) are incidentally nearly orthogonal in directional sensitivities to the ligo detectors .
the questions arise on how to best combine data from these detectors in gw data analysis .
we present in this paper an application of the singular value decomposition ( svd ) method@xcite to data analysis for a network of gw detectors .
we show that the svd method provides simple solutions to detection , waveform extraction , source localization , and signal - based vetoing . by means of svd
, the response matrix of the detector network can be decomposed into a product of two unitary matrices and a pseudo - diagonal matrix containing singular values .
the unitary matrices can be used to form linear combinations of data from all detectors that have one to one correspondence to linear combinations of the gravitational wave signal polarization components .
each newly formed data stream has a corresponding singular value representing the network s response to the new signal polarizations .
data streams with non - zero singular values represent the signal components while data with zero singular values ( or zero multiplication factors ) represent the null streams .
the null streams have null response to gws and can be used for localization of gw sources@xcite and for identifying detector glitches from that of real gravitational wave events as proposed by wen and schutz ( 2005)@xcite for ground - based gw detectors .
the statistical uncertainty in estimating the gw waveforms from the data can be shown to be related to the inverse of the singular values .
this can be used to reduce errors in signal extraction by enabling `` bad '' data with relatively small singular values to be discarded .
the observed strain of an impinging gw by an interferometric detector is a linear combination of two wave polarizations @xmath0 where @xmath1 is time , @xmath2 and @xmath3 are the detector s response ( antenna beam pattern functions ) to the plus and cross polarizations ( @xmath4 , @xmath5 ) of a gw wave .
these antenna beam patterns depend on source sky directions , wave polarization angle , and detector orientation .
the @xmath6 subscript is a label for the @xmath6th detector , indicating the dependence of the observed quantities on detectors .
data from a gw detector can be written as the sum of the detector s response and noise , @xmath7 , where , @xmath8 $ ] , @xmath9 is the duration of the data used , @xmath10 is the wavefront arrival time at the @xmath6-th detector relative to the reference detector ( labeled as detector 1 ) . the arrival time delay can be calculated as @xmath11 , where @xmath12 is location vector of the @xmath6-th detector relative to the reference detector , @xmath13 is the speed of light , @xmath14 is the unit vector along the wave propagation direction . for a given source direction ,
the time - delay corrected data from a network of @xmath15 gw detectors can be written in the frequency domain as , @xmath16 where , @xmath17 is the label for frequency bins , @xmath18 is the response matrix of the detector network at each frequency , @xmath19 and @xmath20 [ a_0 ] where @xmath21 is the noise variance of the @xmath22-th detector at the @xmath17-th frequency bin .
the response of the detector network to gws in all frequencies can be further written in terms of vectors and matrices as , @xmath23 where , @xmath24 @xmath25 of dimension @xmath26 is therefore the response matrix of the detector network for all frequencies .
the singular value decomposition of the network response matrix @xmath25 can be written as , @xmath27 where singular values @xmath28 . @xmath29 and @xmath30 are unitary matrices of dimensions @xmath31 and @xmath32 respectively with @xmath33 .
note that @xmath34 are eigenvalues of @xmath35 .
note also that the singular values of @xmath25 are only a rearrangement of singular values of @xmath18 .
we propose to construct new data streams by linearly recombining data from different detectors and construct new signal polarizations using the unitary matrices resulting from the svd of the network response matrix @xmath25 , @xmath36 these newly constructed data streams and new signal polarizations are one - to - one related to each other via the pseudo - diagonal matrix containing singular values , @xmath37 at the presence of a gw signal , there are at most two data streams at each frequency corresponding to non - zero singular values . if the data contain pure stationary gaussian noise only , the new data stream also follows gaussian distribution and are statistically uncorrelated with each other . these data streams
can therefore be treated effectively as one - detector data streams containing signals . all detection algorithms for a single detector
can be applied to these `` signal '' data streams .
the singular values of the response matrix @xmath25 of the detector network are directly related to the statistical uncertainties in estimating waveforms from the data .
the fisher information matrix for estimating wave polarizations @xmath38 is , @xmath39 where we define the variance of a complex number to be the sum of the variance of the real and complex parts .
the lower - bound@xcite for the variance of the estimated gw signals @xmath40 due to statistical errors are therefore a linear combination of @xmath41 .
singular values are therefore indicators for the accuracy of waveform estimation and can be used for regularization of the solutions to the waveforms . at the end ,
data streams corresponding to zero singular values or zero multiplication factors have null response to signals ( eq . [ new_data ] )
. there are at least @xmath42 `` null streams '' due to the fact that there are only two polarizations for a plane gw based on einstein s theory of general relativity .
null streams can therefore be used to test the consistency of the detected gw events and to veto against signal - like noise.@xcite in this section , we discuss the construction of detection statistic for gws of unknown waveforms using data from a network of detectors .
we discuss two possible approaches , ( 1 ) a direct application of maximum likelihood ratio@xcite ( mlr ) principle by constructing detection statistic from adding powers of all data with non - zero singular values , ( 2 ) optimized detection statistic based on mlr together with assumptions on wave spectrum and principle of maximum signal - to - noise ratio ( snr ) .
snr is a good indicator for detection efficiency in our case as , when adding sufficient large number of noise powers together , central limit theorem states that these snrs , regardless of underlying noise distributions , follow roughly gaussian distribution .
the `` standard '' solution is based on a direct application of the mlr principle for unknown waveforms . when the signal polarizations @xmath43 and @xmath44 are assumed to be independent variables for all frequencies ,
the components of the new signal @xmath45 are also independent from one another .
the standard mlr detection statistic can then be constructed by adding equally all powers of `` signal '' data - streams of non - zero singular values , @xmath46 where @xmath47 is the number of non - zero singular values .
note that singular values have already been ranked ( eq . [ svd ] ) . in the absence of noise , @xmath48 . in the presence of pure stationary gaussian noise
, @xmath49 follows @xmath50 distribution with variance of @xmath51 . by discarding data of null response to gw signals
, this detection statistic is optimal in snrs than simply adding powers of all ( noise - weighted ) data together .
however , in the presence of extremely small singular values ( e.g , when two detectors are nearly aligned ) or extremely weak signals , this detection statistic does not necessarily maximize the snr .
singular values can be used to further optimize the detection statistic by making reasonable assumptions about the signal spectrum .
for instance , we can omit signal powers of small singular values with hopes that the chance is very small for the new signal component @xmath52 to be very strong at frequencies and polarizations where the network is not sensitive . under this assumption
, the detection statistic can be , @xmath53 that is , only @xmath54 number of powers of reasonable large singular values are included .
if we assume that the detection statistic is a linear combination of individual powers and that the signal power @xmath55 is the same in both frequency and new polarizations , maximization of the snr leads to , @xmath56 where , @xmath57 if we assume flat power in signal polarizations only , summation and normalization of @xmath34 terms in eq . [ p1_alpha ] apply only to the two polarizations within each frequency .
we have also investigated the approach of adding powers within a window of the ranked singular values as a general approach to unknown waveforms , @xmath58 where the bandwidth of @xmath59 is determined by maximizing the snr , @xmath60 this is an ad hoc approach that works better for signal powers that are narrow - banded in frequencies .
while this approach can be time - consuming , it performs pretty well in our test with the bh - bh merger signals . early work on optimization of the detection statistic based on directional sensitivity of gw detectors can be found in rajesh nayak et al .
( 2003 ) for future space detector lisa@xcite and in klimenko et al .
( 2005 , 2006 ) for ground - based detectors.@xcite it has been demonstrated in these papers that further optimization from the standard mlr detection statistic can be obtained by constraining or eliminating contributions of data corresponding to weaker network sensitivity . other work can be found in mohanty et al .
( 2006)@xcite which is based on the snr variability regulator and in summerscales et al .
( 2006)@xcite which is based on the maximum entropy principle .
we show that the svd method provides a simple mean to the construction of the detection statistic and its further optimization .
the standard detection statistic based on the maximum likelihood ratio principle can be easily reproduced from the newly constructed data .
various optimization strategies can then be obtained based on general assumptions of the signals together with the singular values which represent the sensitivity of the detector network .
we have also proposed a strategy by maximization of the snrs from the summation of powers corresponding to a window of ranked singular values .
a comparison of the performance of different detection statistic can be found in a follow - up paper.@xcite the `` standard '' estimator for @xmath61^t$ ] according to the maximum likelihood ratio principle of `` burst '' gws ( defined as any gws of unknown waveforms ) is @xmath62 where @xmath47 is the number of non - zero singular values .
note the components of @xmath63 include two polarization components for each frequencies .
the `` standard '' solutions to wave polarizations can be extracted using the svd components of the network response ( eq
. [ svd ] , eq .
[ new_data ] ) @xmath64 note that the solutions of two signal polarization components at each frequency depends on quantities within that frequency only .
faster calculations can therefore be carried out independently at each frequencies . in the limit when @xmath65 , the solution given in eq .
[ h_k0 ] is equivalent to that written with the moore - penrose inverse , @xmath66 .
note that this is also the same as the effective one - detector strain for data from a network of gw detectors introduced by flanagan and hughes ( 1998).@xcite the `` standard '' solution @xmath67 from eq .
[ h_k0 ] however can possibly lead to unstable solutions in the sense that a small error in the data can lead to large amplified error in the solution .
fisher information matrix ( eq . [ fisher ] ) indicates that the best possible statistical variance for the estimated @xmath68 is a linear combination of @xmath69 .
the extracted wave polarizations @xmath68 can contain large errors if we include data corresponding to very small singular values .
this situation occurs when the response matrix @xmath25 is `` ill - conditioned '' with @xmath70 .
the small singular values can result from machine truncation errors instead of zero values or from nearly degenerated solutions to the equations , e.g. , in our case , when antenna beam patterns of two detectors are nearly aligned .
regularization is needed in order to have stable solutions to @xmath40 .
one simple solution is to apply the `` truncated singular value decomposition '' ( tsvd ) method by omitting data with very small singular values in eq .
@xmath71 where @xmath72 is the number of data included .
the main problem is then the decision on where to start to truncate the data streams with small singular values .
it depends on the accuracy requirement in waveform extraction , type of waveforms and type of constraints that can be put on the solutions .
the fractional error due to truncation , defined as the ratio of the sum of error - square and the sum of detector - response square from all frequencies , is , @xmath73 truncation of terms with very small singular values can therefore retain the least square fit of the detector response to the data while greatly reduce the statistical errors when extracting individual signal polarizations .
a recent work on introducing the tikhonov regularization to waveform extraction of gw signals can be found in rakhmanov ( 2006).@xcite note that the tikhonov regularization is equivalent to introducing a new parameter to filtering out data associated with small singular values .
there are at least @xmath42 data streams with zero singular values or zero multiplication factors .
these are data streams that have null response to signals .
the `` standard '' null streams can be written in terms of the svd of @xmath25 ( eq . [ svd ] ) as , @xmath74 for stationary gaussian noise , @xmath75 follows gaussian distribution of zero mean and unity variance and are statistically independent with each other .
consistency check for a detected gw event therefore requires that all the null - streams are consistent with ` noise - only' at the source direction . using null - streams as a tool for consistency check of gw events against signal - like glitches for ground - based gw detectors was first proposed by wen and schutz (
2005).@xcite further investigations on consistency check using the null - streams and its relation to the null space of the network - response matrix @xmath25 can be found in chatterji et al .
( 2006).@xcite there are also `` semi- '' null streams corresponding to data streams with very small but non - zero singular values , @xmath76 where @xmath77 is the starting indexes for @xmath78 .
these semi - null streams exist when the equations are nearly degenerated .
this happens , for example , for response of the two ligo detectors of h1 ( h2 ) and l1 that are designed to be nearly aligned .
the semi - null stream can be also simply caused by combination of weak directional sensitivity and/or high noise level instrument of all detectors in the network .
consistency check of gw events , veto against noise , and localization of the source can be further improved by including both the `` standard '' null - streams of analytically zero singular values and also these semi - null streams . @xmath79
the usage of semi - null streams depends on waveforms and therefore efficiency studies should be carried out before - hand .
a proposal of using the semi - null stream for veto against detector glitches and for source localization can be found in wen and schutz ( 2005)@xcite for the two - detector network of h1-l1 .
a new data analysis approach based on the singular value decomposition method has been proposed for the data analysis of gws using a network of detectors .
we show that singular values of the response matrix of the gw network directly encode the sensitivity of the gw network to signals and the uncertainties in waveform estimation .
the svd method is particularly useful for constructing null data streams that have no or very little response to gw signals .
we argue that the svd method provide a simple recipe for data analysis of gws for a network of detectors .
note that the svd method is widely used in engineering signal processing for image processing , noise reduction and geophysical inversion problems .
an application of the svd method to detecting gws from the extreme - mass - ratio - inspiral sources using the space gw detector lisa can be found in wen et al .
( 2006).@xcite the svd software package is also available in matlab
. strategies on construction of the detection statistic , stable solutions to waveforms , and null streams based on the svd method are discussed .
we show that the detection statistic should be constructed from data streams of non - zero singular values .
we discuss how detection efficiency can be improved from a direct application of mlr by incorporating our knowledge of the network s directional response to gws and our assumptions on distribution of the signal power .
we also give expressions of null - streams for arbitrary number of gw detectors using components from the svd of the network - response matrix .
the concept of the semi - null streams that are characterized by small singular values is also introduced .
we argue that the exploration of the null - space by including semi - null streams can help improving the source localization and consistency check .
analytical study for the angular resolution of a network of gw detectors and their relations with the null space will be found in wen@xcite ( 2007 ) .
we also show how a stable solution to the waveform can be constructed based on information from the singular values .
we conclude that a gw event should be identified only when both of the following two conditions are satisfied at the same sky direction , ( 1 ) high probability of detection for the optimal statistic constructed from `` signal '' streams with non - zero singular values ( eq . [ new_data ] , sec .
[ detection ] ) and ( 2 ) high probability that the null streams corresponding to zero singular values are consistent with noise ( eq . [ new_data ] , eq .
[ null_data ] , sec .
[ null ] ) .
results of an extended investigation will be published elsewhere.@xcite
we are very grateful to yanbei chen for his critical discussions .
we also thank wei - tou ni , sergey klimenko , soumya mohanty , and david blair for careful review and helpful comments of this manuscript .
this work is under the support by the alexander von humboldt foundation s sofja kovalevskaja programme funded by the german federal ministry of education and research .
this paper has been assigned ligo document number ligo - p070012 - 00-z .
christian hansen , `` the truncated svd as a method for regularization '' , _ bit numerical mathematics _ , * 27 * , 534 ( 1987 ) .
y. grsel and m. tinto , _ phys .
d _ , * 40 * , 3884 ( 1989 ) .
l. wen and b. f. schutz , _ classical and quantum gravity _
, * 22 * , s1321 ( 2005 ) .
h. cramer , _ mathematical methods of statistics _ ( princeton university press , princeton , n. j. 1946 ) .
p. g. hoel , s. c. port , and c. j. stone ,
testing hypotheses " in _ introduction to statistical theory _
3 ( new york : houghton mifflin , 1971 ) , p. 56 .
k. rajesh nayak , s. v. dhurandhar , a. pai and j - y vinet , _ phys .
d _ , * 68 * , 122001 ( 2003 )
. s. klimenko , s. mohanty , m. rakhmanov and g. mitselmakher , _ phys .
d _ , * 72 * , 122002 ( 2005 ) .
sergei klimenko , soumya d. mohanty , malik rakhmanov , guenakh , mitselmakher , _ j. phys .
_ , * 32 * 12 - 17 ( 2006 ) .
soumya d. mohanty , malik rakhmanov , sergei klimenko , guenakh , mitselmakher , _ classical quant .
* 23 * , 4799 ( 2006 ) . t. z. summerscales , a. burrows , c. d. ott , and l. s. finn , maximum entropy for gravitational wave data analysis : inferring the physical parameters of core - collapse supernovae , submitted to _ astrophys .
j. _ ( 2007 ) .
l. wen , data analysis of gravitational waves using a network of detectors , in preparation ( 2007 ) . .
flanagan and s. a. hughes , _ phys .
d _ , * 57 * 4566 ( 1998 ) .
m. rakhmanov , _ classical and quantum gravity _ * 23 * , s673 ( 2006 ) .
s. chatterji , a. lazzarini , l. stein , p. j. sutton , a. searle , & m. tinto , _ phys .
d _ , * 74 * , 082005 ( 2006 ) .
l. wen , y. chen and j. gair , _ aip conf .
proc . 873 : laser interferometer space antenna : 6th international lisa symposium _ , * 873 * , 595 ( 2006 ) .
l. wen , angular resolution of a network of gravitational - wave detectors , 2007 , apjl , submitted . | several large - scale gravitational wave ( gw ) interferometers have achieved long term operation at design sensitivity .
questions arise on how to best combine all available data from detectors of different sensitivities for detection , consistency check or veto , localization and waveform extraction .
we show that these problems can be formulated using the singular value decomposition ( svd)@xcite method .
we present techniques based on the svd method for ( 1 ) detection statistic , ( 2 ) stable solutions to waveforms , ( 3 ) null - stream construction for an arbitrary number of detectors , and ( 4 ) source localization for gws of unknown waveforms . |
for any finite group @xmath0 and prime number @xmath1 , it is an interesting question to ask which ordinary irreducible representations of @xmath0 remain irreducible in characteristic @xmath1 .
this question was answered some time ago for the symmetric groups @xcite , and more recently @xcite for the alternating groups . in this paper
we address double covers of the symmetric groups ; since ( in all but finitely many cases ) these are schur covers , we implicitly also address the question of which projective representations of the symmetric groups remain irreducible in characteristic @xmath1 .
we denote a chosen double cover of @xmath3 by @xmath4 ( though in fact our results apply equally well to the other double cover ) ; thus @xmath4 has a central involution @xmath5 such that @xmath6 . in any characteristic
one can separate the irreducible representations of @xmath4 into `` linear '' representations ( on which @xmath5 acts trivially ) and `` spin '' representations ( on which @xmath5 acts as @xmath7 ) .
the linear representations are also irreducible representations of @xmath3 , and so the answer to our main question for linear representations follows at once from the answer for symmetric groups .
so we need only consider spin representations .
in fact in this paper we prefer to work in the language of characters : the ordinary irreducible spin characters of @xmath4 are labelled by @xmath8-regular partitions of @xmath9 : there is a single character @xmath10 for every @xmath8-regular partition @xmath11 of @xmath9 with an even number of even parts , and a pair of associate characters @xmath12 for every @xmath11 with an odd number of even parts . in the latter case ,
we write @xmath10 to mean `` either @xmath13 or @xmath14 '' , so our main question becomes `` for which @xmath11 is the @xmath1-modular reduction of @xmath10 irreducible ? '' .
there has been a recent resurgence of interest in the representation theory of @xmath4 in odd characteristic .
we hope to address this case in a future paper , but in the present paper we work instead in characteristic @xmath8 . apart from the fundamental paper @xcite of bessenrodt and olsson in 1997 from which we take some important results ) , this situation has been relatively neglected , but we hope that the present paper will spark a revival .
note that in characteristic @xmath8 the central involution @xmath5 must act as @xmath15 on an irreducible representation , so there are no irreducible spin representations .
so the irreducible brauer characters in characteristic @xmath8 are precisely those coming from @xmath3 , namely the characters @xmath16 of the `` james modules '' @xmath17 . thus the @xmath8-modular reduction of any ordinary irreducible spin character decomposes as a sum of characters @xmath16 .
this means that the decomposition matrix of @xmath4 in characteristic @xmath8 consists of the decomposition matrix of @xmath3 extended downwards by a `` spin matrix '' with rows corresponding to ordinary irreducible spin characters .
although this spin matrix can be computed fairly easily given the decomposition matrix of @xmath3 , it remains fairly mysterious in general .
a key contribution in the present paper is to describe part of this matrix ( specifically , the part with rows labelled by @xmath8-regular partitions having no even parts greater than @xmath8 ) for rouquier blocks of @xmath4 .
we now indicate the structure of the paper and of the proof of our main theorem .
in [ backsec ] we set out the background information we shall need ; we try to provide enough information to make the paper reasonably self - contained , though we do make some effort at concision . in [ mainthmsec ]
we recall the notion of a _ carter partition _ which enables us to state our main theorem . in [ degreesec ]
we examine character degrees , which are an important ingredient in our proof .
the key observation here is that if @xmath10 and @xmath18 are ordinary characters having a @xmath8-modular constituent in common , and if the degree of @xmath10 is greater than the degree of @xmath18 , then the @xmath8-modular reduction of @xmath10 can not be irreducible .
it turns out that ( in contrast to the more famous hook - length formula for the degrees of linear characters ) the bar - length formula for spin characters allows us to compare degrees quite easily . in [ rouqandsep ]
we consider spin characters lying in a `` rouquier '' block of @xmath4 .
while rouquier blocks of symmetric groups have been studied extensively in the last few years , this appears to be the first time they have been considered for the double covers in characteristic @xmath8 . by constructing various projective characters and exploiting connections with symmetric functions ,
we describe part of the decomposition matrix for a rouquier block .
we then extend this beyond rouquier blocks to spin characters labelled by what we call _ separated _ partitions , and give a proof of our main theorem for separatedpartitions . in [ mainproofsec ] we complete the proof of our main theorem , by considering non - separatedpartitions .
here we use an inductive proof ( similar in spirit to that used for the alternating groups in @xcite ) , based on kleshchev s modular branching rules . in [ futuresec ]
we consider possibilities for future work , and in [ indexnotnsec ] we provide an index of notation .
the research presented in this paper would not have been possible without extensive computations using gap @xcite .
_ throughout this section @xmath19 is a fixed field of characteristic @xmath8 . _
we let @xmath3 denote the symmetric group of degree @xmath9 , and @xmath20 its group algebra over @xmath19 .
the representation theory of @xmath20 is well - studied , and in the last twenty years or so has become increasingly intertwined with the representation theory of a certain ( iwahori)hecke algebra .
specifically , we let @xmath21 denote the iwahori hecke algebra of @xmath3 over @xmath22 with quantum parameter @xmath7 ; this has generators @xmath23 , with defining relations @xmath24 for each @xmath25 , together with the braid relations @xmath26 for @xmath27 and @xmath28 for @xmath29
. then the representation theory of @xmath21 is closely related to the representation theory of @xmath20 : the canonical labelling sets for the specht modules and simple modules for these algebras are the same , and ( as we shall see ) there is a close relationship between their decomposition matrices .
the main advantage of introducing @xmath21 into the subject is that its decomposition numbers are better understood than those of @xmath20 ( indeed , they can be computed using the _ llt algorithm _
we shall exploit this later , although the relationship between the symmetric group and the hecke algebra does not extend well to double covers of @xmath3 .
another important family of algebras in this area is the family of _ schur algebras _ introduced by green @xcite , and more generally the _
@xmath30-schur algebras _ defined by dipper and james ( * ? ? ?
* definition 2.9 ) .
we just concentrate on two of these algebras , for each @xmath9 : let @xmath31 denote the classical schur algebra over @xmath19 ( which is @xmath32 in the notation of dipper james ) , and let @xmath33 denote the @xmath30-schur algebra over @xmath22 with quantum parameter @xmath34 , i.e. @xmath35 in the dipper james notation .
@xmath31 is related to @xmath20 via the _ schur functor _ , which shows in particular that the decomposition matrix of @xmath20 can be obtained from that of @xmath31 by deleting some of the columns .
@xmath33 is related to @xmath21 is a similar way .
as we shall see , the decomposition matrix of @xmath31 also appears in the adjustment matrices of rouquier blocks .
now we consider double covers . in this paper
we work with the group @xmath4 with generators @xmath36 , subject to relations @xmath37 then @xmath4 is a double cover of the symmetric group @xmath3 : the subgroup @xmath38 is central , and setting @xmath5 to @xmath15 in the above presentation yields the coxeter presentation of @xmath3 .
in fact , for @xmath39 @xmath4 is a _ schur cover _ of @xmath3 .
this means that linear representations of @xmath4 are equivalent to projective representations of @xmath3 , i.e. homomorphisms from @xmath3 to @xmath40 for a vector space @xmath41 .
thus our main theorem can be phrased in terms of projective representations . for @xmath42
, @xmath3 is its own schur cover ( note that the final relation in the presentation above never occurs when @xmath42 , so that @xmath4 is just the direct product of @xmath3 and @xmath38 ) .
nevertheless , it makes sense to work with the group @xmath4 for all @xmath43 , and we do so in this paper to help with our induction proofs .
we remark that for @xmath39 there is an alternative schur cover obtained by replacing the relation @xmath44 with @xmath45 ; this is isomorphic to @xmath4 only when @xmath46 , but in general is isoclinic to @xmath4 , and so from our point of view its representation theory is the same .
now we consider the different types of irreducible representations of @xmath4 . in an irreducible representation of @xmath4 over any field ,
the central element @xmath5 must act as @xmath15 or @xmath7 .
irreducible representations in which @xmath5 acts as @xmath15 correspond exactly to irreducible representations of @xmath3 , and we refer to these as _ linear _ representations . studying linear representations of @xmath4
is equivalent to studying representations of @xmath3 , so in this paper we shall be chiefly concerned with the representations where @xmath5 acts as @xmath7 , which we call _ spin _ representations . for an introduction to the representation theory of @xmath4
, we refer the reader to the book by hoffman and humphreys @xcite .
the representations of the symmetric groups and their double covers are conventionally labelled by partitions . in this paper ,
partition _ is an infinite weakly decreasing sequence @xmath47 of non - negative integers with finite sum , which we denote @xmath48 .
if @xmath49 , then we say that @xmath11 is a partition of @xmath9 .
the integers @xmath50 are called the _ parts _ of @xmath11 . when writing partitions , we conventionally omit the trailing zeroes and group together equal parts with a superscript ( so that we write the partition @xmath51 as @xmath52 ) , and we write the unique partition of @xmath53 as @xmath54 .
a partition is _
@xmath8-regular _ if it does not have two equal non - zero parts , and _
@xmath8-singular _ otherwise .
we write @xmath55 for the set of partitions of @xmath9 , and @xmath56 for the set of @xmath8-regular partitionsof @xmath9 , and set @xmath57 and @xmath58 . if @xmath59 , the _ conjugate partition _
@xmath60 is defined by setting @xmath61 for all @xmath62 .
the _ dominance order _ on @xmath55 is defined by saying that @xmath11 _ dominates _
@xmath63 ( and writing @xmath64 ) if @xmath65 for all @xmath66 .
the _ young diagram _ of a partition @xmath11 is the set @xmath67 we often identify a partition with its young diagram ; for example , we may write @xmath68 to mean that @xmath69 for all @xmath25 . the elements of the young diagram of @xmath11 are called the _ nodes _ of @xmath11 .
more generally , we use the term _ node _ for any element of @xmath70 .
we draw young diagrams as arrays of boxes using the english convention , in which @xmath62 increases down the page and @xmath71 increases from left to right .
a node of @xmath11 is _ removable _ if it can be removed from @xmath11 to leave the young diagram of a smaller partition , while a node not in @xmath11 is an _ addable node _ of @xmath11 if it can be added to @xmath11 to leave the young diagram of a larger partition .
the _ residue _ of a node @xmath72 is the residue of @xmath73 modulo @xmath8 .
we refer to a node of residue @xmath25 as an @xmath25-node .
now we introduce some notions relating to spin representations .
the _ spin residue _ of a node @xmath72 is the residue of @xmath74 modulo @xmath8 ; that is , @xmath53 if @xmath75 or @xmath76 , and @xmath15 otherwise .
if @xmath11 is a @xmath8-regular partitionand @xmath77 , we write @xmath78 for the smallest @xmath8-regular partitionwhich can be obtained from @xmath11 by removing nodes of spin residue@xmath25 ; the nodes removed are called _ @xmath25-spin - removable nodes _ of @xmath11 .
for example , let @xmath79 . then @xmath80 and @xmath81 .
the following diagram shows the spin residuesof the nodes of @xmath11 , with the spin - removable nodeshighlighted .
@xmath82 * some notation for partitions * in this paper we shall use some notation for partitions which , although quite natural , is not particularly standard .
suppose @xmath11 and @xmath63 are partitions and @xmath83 .
then we write * @xmath84 for the partition @xmath85 ; * @xmath86 for the partition @xmath87 ; * @xmath88 for the partition obtained by combining all the parts of @xmath11 and @xmath63 and arranging them in decreasing order . for example ,
if @xmath89 , @xmath90 and @xmath91 , then we can define the partition @xmath92 we now set out some of the combinatorics of partitions relevant to the @xmath8-modular representation theory of the symmetric group .
everything we shall say has analogues for all primes @xmath1 , but we stick to @xmath2 in the interests of brevity .
suppose @xmath59 .
a _ rim @xmath8-hook _ of @xmath11 is a pair of ( horizontally or vertically ) adjacent nodes of @xmath11 which may be removed to leave a smaller young diagram ; we call the rim hook _ horizontal _ or _ vertical _ accordingly
. the _ @xmath8-core _ of @xmath11 is the partition obtained by repeatedly removing rim @xmath8-hooks until none remain .
the @xmath8-core is well defined , and has the form @xmath93 for some @xmath94 .
the _ @xmath8-weight _ of @xmath11 is the number of rim @xmath8-hooks removed to reach the @xmath8-core .
this may be visualised using the abacus .
take an abacus with two infinite vertical runners , and label these @xmath53 and @xmath15 ( with @xmath53 being the runner on the left ) . on runner @xmath25 mark positions @xmath95
, so that position @xmath96 is directly to the right of position @xmath97 for all @xmath25 . now given a partition @xmath11 ,
place a bead on the abacus at position @xmath98 for each @xmath25 .
the resulting configuration is called the _ abacus display _ for @xmath11 . the key observation motivating
the abacus is that removing a rim @xmath8-hook corresponds to sliding a bead into an empty position immediately above .
hence an abacus display for the @xmath8-core of @xmath11 may be obtained by repeatedly sliding beads up until every bead has a bead immediately above it .
the abacus also allows us to define the @xmath8-quotient of a partition .
given the abacus display for @xmath11 and given @xmath99 , examine runner @xmath100 in isolation , and let @xmath101 be the number of empty positions above the @xmath25th lowest position , for each @xmath25 .
then @xmath102 is a partition , and the pair @xmath103 is the _
@xmath8-quotient _ of @xmath11 .
the partition @xmath104 has @xmath8-core @xmath105 , as we see from the young diagram ( in which we indicate removed rim @xmath8-hooks ) and the abacus displays of these two partitions .
@xmath106{\tyng(0cm,0cm,6,4 ^ 2,3)\yfillopacity{0}{\ylinethick{1.3pt}}\tgyoung(0cm,0cm,::_2_2,:|2|2|2:,|2:,:_2)}\qquad\qquad \abacus(vv , bb , nn , nb , nb , bn , nb , nn , vv)\qquad\qquad\abacus(vv , bb , bb , nb , nb , nn , nn , nn , vv)\ ] ] examining the runners in the display for @xmath11 , we see that the @xmath8-quotient of @xmath11 is @xmath107 .
later we shall need the following simple lemma ( a version of this for arbitrary @xmath1 appears in @xcite , among other places ) .
[ quoconj ] suppose @xmath108 , and let @xmath109 be the @xmath8-quotient of @xmath63 .
then the @xmath8-quotient of @xmath110 is @xmath111 .
the final thing we need relating to cores and quotients is the _
@xmath8-sign _ of a partition .
this was originally introduced by littlewood @xcite , but we use the alternative definition from @xcite .
suppose @xmath63 is a partition , and construct the @xmath8-core of @xmath63 by repeatedly removing rim @xmath8-hooks .
let @xmath100 be the number of vertical hooks removed .
@xmath100 is not well - defined , but its parity is , and so we may safely define @xmath112 .
@xmath113 is then called the @xmath8-sign of @xmath63 .
we will need the following very simple observation .
[ parityconj ] suppose @xmath63 is a partition of @xmath8-weight @xmath114
. then @xmath115 .
having established the necessary background relating to the combinatorics of partitions , we can discuss modules and decomposition numbers for the groups and algebras studied in this paper . for each @xmath116 , we write @xmath117 for the _ specht module
_ for @xmath3 , as defined ( over an arbitrary field ) by james @xcite ; in particular , @xmath118 is the trivial module . working over our field @xmath19 of characteristic
@xmath8 , suppose @xmath119 ; then @xmath117 has a unique irreducible quotient @xmath120 , which we call the _
james module_. the modules @xmath120 give all the irreducible @xmath20-modules as @xmath11 ranges over @xmath56 .
a similar situation applies for the hecke algebra @xmath21 : for each @xmath116 there is a specht module for @xmath21 ( which we shall also denote @xmath117 ) , which has a unique irreducible quotient @xmath120 when @xmath11 is @xmath8-regular , and the modules @xmath120 are the only irreducible @xmath21-modules up to isomorphism .
now we consider the schur algebras . for every @xmath116
there is a _ weyl module _
@xmath121 for @xmath31 , which has a unique irreducible quotient @xmath122 .
note that the labelling we use for these modules is that used in james s paper @xcite ( and is the opposite of that in most other places ) , so that the specht module @xmath117 is the image of @xmath121 ( and not @xmath123 ) under the schur functor . similarly for the @xmath124-schur algebra @xmath33 we have a weyl module @xmath121 and an irreducible module @xmath122 for every @xmath116 .
with these conventions established , we can define decomposition numbers . for any @xmath125 , we define @xmath126 to be the composition multiplicity @xmath127}$ ] for the schur algebra @xmath31 .
the fact that @xmath31 is a quasi - hereditary algebra gives the following .
[ unitriang ] suppose @xmath125 .
then @xmath128 , and @xmath129 unless @xmath130 .
we emphasise that we use the convention from @xcite for labelling weyl modules ; with the more usual convention the symbol @xmath131 would become @xmath132 in the above .
this convention also means that when @xmath133 , @xmath126 is also the composition multiplicity @xmath134}$ ] for @xmath3 in characteristic @xmath8 . the matrix with entries @xmath135 is the _ decomposition matrix _ for @xmath31 , and the matrix with entries @xmath136 is the decomposition matrix for @xmath3 in characteristic @xmath8 .
we may write either of these matrices simply as @xmath137 , if the context is clear .
a similar situation applies for hecke algebras and @xmath124-schur algebras .
for @xmath125 we write @xmath138 for the composition multiplicity @xmath127}$ ] for the @xmath124-schur algebra @xmath33 ; if @xmath133 , then this also equals the multiplicity @xmath134}$ ] for the hecke algebra @xmath21 .
in addition , [ unitriang ] holds with @xmath137 replaced by @xmath139 .
furthermore , the decomposition numbers @xmath126 and @xmath138 are related by the theory of adjustment matrices : if we fix @xmath9 and let @xmath137 denote the decomposition matrix of @xmath31 and @xmath139 the decomposition matrix of @xmath33 , then the square matrix @xmath140 defined by @xmath141 is called the _ adjustment matrix _ for @xmath31 .
@xmath140 is automatically lower unitriangular ( i.e. [ unitriang ] holds with @xmath137 replaced by @xmath140 ) , and remarkably @xmath140 has non - negative integer entries ; this statement appears in ( * ? ? ?
* theorem 6.35 ) , and derives from the theory of decomposition maps @xcite . by taking only the rows and columns of @xmath140 labelled by @xmath8-regular partitionswe
obtain the adjustment matrix for @xmath3 in characteristic @xmath8 , which we also call @xmath140 ; then if we let @xmath142 denote the decomposition matrices for @xmath3 and @xmath21 , we again have @xmath141 .
now we come to double covers . here
( and for the rest of the paper ) we work with characters , rather than modules .
this is because we do not have particularly nice constructions of the modules affording the irreducible characters of double covers , and because statements about the decomposition of projective modules are more readily expressed in terms of projective characters . since we are only concerned with composition factors and not with module structures , there is no loss in working with characters . with this in mind , we introduce notation for the characters of the @xmath3-modules described above : let @xmath143 denote the character of the specht module @xmath117 over @xmath22 , so that @xmath144 is the set of ordinary irreducible characters of @xmath3 . for @xmath119 ,
let @xmath145 denote the @xmath8-modular brauer character of the james module @xmath120 .
since @xmath3 is a quotient of @xmath4 , the characters @xmath143 , @xmath145 naturally become characters of @xmath4 , and we shall always view them in this way . note that in characteristic @xmath8 , the central element @xmath5 lies in the kernel of any irreducible representation of @xmath4 , so the irreducible representations of @xmath4 are precisely those arising from the irreducible representations of @xmath3 .
thus the brauer characters @xmath145 for @xmath119 are all of the irreducible @xmath8-modular brauer characters of @xmath4 .
now we look at spin characters . for each @xmath119 , let @xmath146 denote the number of positive even parts of @xmath11 , and write @xmath147 then for each @xmath148 , there is an irreducible spin character @xmath10 of @xmath4 , while for each @xmath149 , there are two irreducible spin characters @xmath13 and @xmath14 of @xmath4 .
the definition of these characters goes back to schur , who showed that they are all the irreducible spin characters of @xmath4 .
a uniform construction of irreducible representations affording these characters was given much later , by nazarov @xcite .
we do not give the representations or their characters explicitly in this paper ; the properties we need will be summarised in later sections . in fact , for @xmath149 the characters @xmath13 and @xmath14 behave very similarly : their values differ only up to sign , and in particular their degrees are the same and their @xmath8-modular reductions are the same .
so to make things simpler in this paper we adopt the convention that we write @xmath10 to mean `` either @xmath13 or @xmath14 '' when @xmath149 .
we also follow the usual convention of omitting brackets from a partition when using the notation @xmath150 or @xmath151 ; thus we may write @xmath152 instead of @xmath153 .
we remark that historically the spin characters have been denoted @xmath154 or @xmath155 ; we use slightly different notation here because the brackets @xmath156 are already used for two other purposes in this paper .
we trust that there will be no confusion with the use of @xmath18 in @xcite to denote a projective character of @xmath4 in characteristic @xmath157 .
we write @xmath158 for the usual inner product on ordinary characters , with respect to which the irreducible characters are orthonormal . for any character @xmath159 of @xmath4
, we write @xmath160 for the @xmath8-modular reduction of @xmath159 , i.e. the @xmath8-modular brauer character obtained by restricting @xmath159 to the @xmath8-regular conjugacy classes of @xmath4 . our main question is then : for which ordinary irreducible characters @xmath159 is @xmath160 irreducible ? we write @xmath161}$ ] for the inner product on @xmath8-modular brauer characters with respect to which the irreducible characters are orthonormal .
hence the decomposition number @xmath126 defined above equals @xmath162}$ ] , for @xmath116 and @xmath133 .
given @xmath163 , we write @xmath164}$ ] .
the integers @xmath126 and @xmath165 together constitute the decomposition matrix of @xmath4 in characteristic @xmath8 .
the decomposition matrix of @xmath166 in characteristic @xmath8 is as follows .
the top part of the matrix is the decomposition matrix of @xmath167 , with the @xmath168-entry being @xmath126 .
the lower part gives the decomposition numbers for spin characters , with entries @xmath165 ; note that for @xmath169 there are identical rows corresponding to @xmath13 and @xmath14 .
( in all matrices given explicitly in this paper , we use a dot to mean @xmath53 . )
@xmath170 this example shows an interesting feature of the answer to our main question : we can re - cast our main question and ask `` which irreducible brauer characters arise as the @xmath8-modular reductions of ordinary irreducible characters ? '' . in this case
we see that the brauer character @xmath171 does not arise as @xmath172 for any @xmath11 , but does arise as @xmath173 .
so introducing spin characters can provide constructions of irreducible brauer characters which might otherwise be hard to obtain . a central result in modular representation theory ( which we shall use without comment )
is brauer reciprocity , which gives a connection between decomposition numbers and indecomposable projective characters . given @xmath133 , the projective cover of the james module @xmath17
may be lifted to an ordinary representation of @xmath4 , and we write @xmath174 for the character of this representation ; this is called an _ indecomposable _ projective character , and the characters @xmath174 for @xmath133 give a basis for the space spanned by all projective characters ( i.e. characters which vanish on @xmath1-singular elements ) .
brauer reciprocity says that @xmath174 is given in terms of irreducible characters by the entries in the column of the decomposition matrix corresponding to @xmath63 , i.e. @xmath175 we will use this extensively in [ rouqandsep ] to derive information on decomposition numbers .
now we state some results which we shall need later concerning the inverses of the matrices @xmath137 , @xmath139 and @xmath140 .
these are taken from james s seminal paper @xcite , and derive ultimately from steinberg s tensor product theorem .
given partitions @xmath176 with @xmath177 , let @xmath109 be the @xmath8-quotient of @xmath63 , and define @xmath178 here ( and henceforth ) @xmath179 denotes the littlewood
richardson coefficient corresponding to partitions @xmath180 with @xmath181 .
now we can state two results on the inverse of the decomposition matrix .
we begin with a result for even @xmath9 ; this is a special case of ( * ? ? ?
* corollary 6.9 ) . recall that @xmath113 denotes the @xmath8-sign of a partition @xmath63 , defined in [ 2coresec ] .
[ steinberg ] suppose @xmath125 and all the columns of @xmath11 are of even length , and write @xmath182
. then @xmath183 as a consequence we can derive some information on the adjustment matrix @xmath140 . [ adj ] suppose @xmath125 and all the columns of @xmath11 are of even length , and write @xmath182
. then @xmath184 and hence @xmath185 the two parts of [ steinberg ] combine to give @xmath186 for all @xmath63 . since the rows of @xmath187 are linearly independent , this enables us to deduce the expression for @xmath188 . the result for @xmath189
then follows by inverting .
now we give a corresponding result for odd @xmath9 ; this is also a special case of ( * ? ? ?
* corollary 6.9 ) .
[ steinbergod ] suppose @xmath125 and all the columns of @xmath11 are of even length except the first column .
write @xmath190 , and let @xmath191 be the set of partitions obtained which can be obtained by removing one node from @xmath63
. then @xmath192 again , we deduce information about the adjustment matrix @xmath140 ; this is proved in exactly the same way as [ adj ] .
[ adjod ] suppose @xmath125 and all the columns of @xmath11 are of even length except the first , and write @xmath190
. then @xmath193 and hence @xmath194 let @xmath195 denote the degree of an irreducible character , i.e. the value @xmath196 .
in this section we recall the `` bar - length formula '' which gives the degrees of the irreducible spin characters .
this goes back to schur @xcite .
recall that for @xmath197 we write @xmath10 to mean either @xmath13 or @xmath14 .
[ barlength ] suppose @xmath119 has length @xmath198 .
then @xmath199 we shall use this formula extensively in [ degreesec ] .
one of the most useful results in the modular representation theory of @xmath3 is james s regularisation theorem , which gives an explicit composition factor ( occurring with multiplicity @xmath15 ) in the @xmath1-modular reduction of @xmath143 , for each @xmath11 . to state this result for @xmath2 , we need to introduce the @xmath8-regularisation of a partition . for @xmath200 , define the @xmath201th _ ladder _ in @xmath70 to be the set of nodes @xmath72 for which @xmath202 .
the intersection of this ladder with the young diagram of a partition @xmath11 is referred to as the @xmath201th ladder of @xmath11 .
now define the _
@xmath203 of @xmath11 to be the @xmath8-regular partitionwhose young diagram is obtained by moving the nodes of @xmath11 as far up their ladders as possible .
for example , if @xmath204 , then @xmath205 , as we see from the following diagrams , in which we label nodes according to the ladders in which they lie .
@xmath206 now we can state a simple form of james s theorem .
j1theorem a[jreg ] suppose @xmath59
. then @xmath207 .
of course , this theorem is extremely useful in determining whether @xmath208 is irreducible , since it tells us that if @xmath208 is irreducible , then @xmath209 . for the main result in the present paper it will be useful to have an analogue of [ jreg ] for spin characters .
fortunately there is such a result ; this is due to bessenrodt and olsson , though a special case was proved earlier by benson ( * ? ? ?
* theorem 1.2 ) . to state
this , we need another definition .
given @xmath197 , define the _
double _ of @xmath11 to be the partition @xmath210 in other words , @xmath211 is obtained from @xmath11 by replacing each part @xmath212 with two parts which are equal ( if @xmath212 is even ) or differ by @xmath15 ( if @xmath212 is odd ) . the fact that @xmath11 is @xmath8-regular guarantees that @xmath211 is a partition .
we write @xmath213 to mean @xmath214 .
now we can give the `` spin - regularisation theorem '' ; recall that @xmath146 denotes the number of positive even parts of a partition @xmath11 .
botheorem 5.2[spinreg ] suppose @xmath197
. then @xmath215 .
this result is extremely useful for our main problem : it tells us that @xmath216 can only be irreducible if @xmath11 has at most one non - zero even part , and that if @xmath216 is irreducible , then @xmath217 . in order to exploit [ spinreg ] , we will often want to show that @xmath218 for certain @xmath163 without actually calculating @xmath213 or @xmath219 .
we do this using the combinatorics of _
slopes_. for @xmath200 , define the @xmath201th slope in @xmath70 to be the set of all nodes @xmath72 satisfying @xmath220 . define the @xmath201th slope of a partition @xmath11 to be the intersection of this slope with the young diagram of @xmath11 .
the slopes are essentially the `` ladders in the @xmath221-residue diagram '' considered in @xcite , but we prefer not to over - tax the word `` ladder '' . the reader should also note that unlike in @xcite we do not consider shifted young diagrams . )
now we have the following result , which is implicit in @xcite . [ slopelad ]
suppose @xmath11 is a @xmath8-regular partitionand @xmath200 .
then the number of nodes in the @xmath201th slope of @xmath11 equals the number of nodes in the @xmath201th ladder of @xmath211 . hence if @xmath63 is another @xmath8-regular partition , then @xmath218 @xmath11 and @xmath63 have the same number of nodes in slope @xmath201 , for each @xmath201 .
for example , in the following diagram we label the nodes of @xmath222 with the numbers of the slopes containing them , and the nodes of @xmath223 with the numbers of the ladders containing them .
@xmath224 at the level of young diagrams , replacing @xmath11 with @xmath211 involves replacing the nodes @xmath225 with @xmath226 for each @xmath25 .
the former nodes lie in slopes @xmath227 respectively , while the latter nodes lie in ladders @xmath228 respectively .
the result follows .
we now recall the @xmath8-block classification for @xmath4 , which will be very useful in this paper .
we begin by looking at the @xmath8-blocks of @xmath3 .
the following result is a special case of the brauer
robinson theorem @xcite , first conjectured by nakayama .
[ brarob ] suppose @xmath125 .
then @xmath143 and @xmath229 lie in the same @xmath8-block of @xmath3 @xmath11 and @xmath63 have the same @xmath8-core .
the same statement applies for @xmath8-blocks of @xmath4 , and a corresponding statement holds for blocks of @xmath21 .
since we only consider characteristic @xmath8 in this paper , we will henceforth say `` block '' to mean `` @xmath8-block '' .
given [ brarob ] , we may speak of the _ core _ of a @xmath8-block @xmath230 of @xmath3 , meaning the common @xmath8-core of the partitions labelling the linear irreducible characters in @xmath230 .
these partitions necessarily have the same @xmath8-weight as well , and we call this the _ weight _ of the block
. we can immediately deduce the distribution of irreducible brauer characters into blocks : if @xmath133 , then since @xmath16 occurs as a composition factor of @xmath231 , it lies in the same block as @xmath229 .
the block classification may alternatively be expressed in terms of residues .
recall that the _ residue _ of a node @xmath72 is the residue of @xmath73 modulo @xmath8 .
define the _
@xmath8-content _ of a partition @xmath11 to be the multiset of @xmath53s and @xmath15s comprising the residues of all the nodes of @xmath11 .
we write a multiset of @xmath53s and @xmath15s in the form @xmath232 .
for example , the @xmath8-content of the partition @xmath233 is @xmath234 , as we see from the following `` @xmath8-residue diagram '' . @xmath235
it is an easy combinatorial exercise to show that two partitions have the same @xmath8-content they have the same @xmath8-core and @xmath8-weight , so the block classification may alternatively be stated by saying that @xmath143 and @xmath229 lie in the same block @xmath11 and @xmath63 have the same @xmath8-content .
accordingly , we can define the _ content _ of a block @xmath230 to be the common @xmath8-content of the partitions labelling the linear irreducible characters in @xmath230 .
now we consider spin characters . since the only irreducible @xmath8-modular characters are the characters @xmath145 , the spin characters fit into the @xmath8-blocks described above , i.e. those containing linear characters .
the distribution of the spin characters among these blocks is given by the following theorem , which was originally conjectured by knrr and olsson @xcite .
botheorem 4.1[spinblockclass ] suppose @xmath119 , and let @xmath236 be the @xmath8-core of @xmath211 .
then @xmath10 lies in the @xmath8-block of @xmath4 with core @xmath236
. it will be helpful to have alternative descriptions of this block classification , for which we need different notions of core and content .
if @xmath197 , we define the _ @xmath237-bar - core _ of @xmath11 to be the @xmath8-regular partitionobtained by repeatedly applying the following operations to @xmath11 : * removing all even parts ; * removing any two parts whose sum is a multiple of @xmath237 ; * replacing any odd part @xmath238 with @xmath239 , if @xmath239 is not already a part of @xmath11 .
the @xmath237-bar - coreof @xmath11 is easily seen to be well defined , and equals either @xmath240 or @xmath241 for some @xmath200 .
note that these are precisely the partitions whose double is a @xmath8-core .
moreover , the double of the @xmath237-bar - coreof @xmath197 coincides with the @xmath8-core of @xmath211 ( * ? ? ?
* lemma 3.6 ) .
so [ spinblockclass ] may alternatively be stated by saying that if @xmath242 is the @xmath237-bar - coreof @xmath119 , then @xmath10 lies in the block of @xmath4 with core @xmath243 .
in particular , two spin characters @xmath10 and @xmath18 lie in the same block @xmath11 and @xmath63 have the same @xmath237-bar - core .
so we can define the _ bar - core _ of a block @xmath230 to be the common @xmath237-bar - coreof the @xmath8-regular partitionslabelling spin characters in @xmath230 .
the @xmath237-bar - coreof @xmath11 is a partition of @xmath244 for some @xmath114 , which we call the _ @xmath237-bar - weight _ of @xmath11 . by the comments above
, the @xmath237-bar - weightof @xmath11 equals the weight of the block containing @xmath10 .
consider the block @xmath230 of @xmath245 with core @xmath246 .
the partitions of @xmath247 with @xmath8-core @xmath246 are @xmath248 , @xmath249 , @xmath250 , @xmath251 and @xmath252 .
@xmath246 is the double of the @xmath237-bar - core@xmath253 , so @xmath253 is the bar - coreof @xmath230 .
the partitions in @xmath254 with @xmath237-bar - core@xmath253 are @xmath255 and @xmath249 .
so the ordinary irreducible characters in @xmath230 are @xmath256 @xmath257 and the irreducible @xmath8-modular characters in @xmath230 are @xmath258 and @xmath259 . if @xmath260 , then @xmath261 is naturally embedded in @xmath4 , and a lot of information can be obtained by inducing and restricting characters between these two groups . in the modular representation theory of the symmetric groups ,
more delicate information can be gleaned by using the so - called _ @xmath25-induction _ and _ @xmath25-restriction _ functors .
we summarise the key points here , specialising to characteristic @xmath8 ( where the results automatically extend to @xmath4 ) .
the definition of the @xmath25-induction and @xmath25-restriction functors goes back to robinson @xcite , though our main reference will be the survey by brundan and kleshchev @xcite .
given a character @xmath159 of @xmath4 , we write @xmath262 for its restriction to @xmath261 , and @xmath263 for the corresponding induced character for @xmath264 .
now suppose @xmath159 lies in a single block @xmath230 , with content @xmath232
. then we write @xmath265 for the component of @xmath262 lying in the block with content @xmath266 if there is such a block , and set @xmath267 otherwise .
similarly , we write @xmath268 for the component of @xmath262 lying in the block with content @xmath269 if there is such a block , and set @xmath270 otherwise .
we extend the functions @xmath271 linearly .
these functions @xmath271 can be applied to either ordinary characters or @xmath8-modular brauer characters . in any case
, it turns out that for any character @xmath159 we have @xmath272 .
the effect of these functions on ordinary irreducible characters is well understood , via the following `` branching theorems '' .
the first of these dates back to young , but the second is much more recent , due to dehuai and wybourne @xcite . to state these
, we introduce some notation .
suppose @xmath11 and @xmath63 are partitions , and @xmath77 .
we write @xmath273 to mean that @xmath11 is obtained from @xmath63 by adding @xmath62 addable @xmath25-nodes ( omitting the superscript @xmath62 when it equals @xmath15 ) .
similarly , we write @xmath274 to mean that @xmath11 is obtained from @xmath63 by adding @xmath62 @xmath25-spin - addable nodes .
[ classbrnch ] suppose @xmath116 and @xmath77 , and let @xmath275 then @xmath276 [ spinbrnch ] suppose @xmath119 and @xmath77 , and let @xmath277 1 .
if @xmath148 , then @xmath278 to state further results , we need to consider powers : @xmath279 and @xmath280 are defined for any @xmath9 , so we can define powers @xmath281 and ( if we allow @xmath282-linear combinations of characters ) divided powers @xmath283 . for any non - zero character @xmath159 and @xmath77
, we can then define @xmath284 the following result comes immediately from the classical branching rule .
[ ordmaxbranch ] suppose @xmath59 and @xmath77 , and let @xmath285 be the partition obtained by removing all the removable @xmath25-nodes from @xmath11
. then @xmath286 . for the spin branching rule , a result of the same form holds , but it is more difficult to keep track of multiplicities .
we begin with the following result , which will also be useful in later sections .
[ spinbranchpower ] suppose @xmath119 , @xmath287 , @xmath77 , and @xmath274 .
let @xmath288 * if @xmath148 and @xmath289 , then @xmath290 recall that we write @xmath291 for the number of positive even parts of @xmath292 , and that by definition @xmath293 @xmath291 is even .
when we remove a spin - removable nodefrom @xmath294 , we change the parity of @xmath291 , unless the removed node is the only node in its row .
applying this observation @xmath62 times , we find the following , when @xmath274 : * if @xmath295 is even , then none of the nodes of @xmath296 lies in the first column ( i.e. @xmath11 and @xmath63 have the same length ) ; * if @xmath295 is odd , then one of the nodes of @xmath296 does lie in column @xmath15 ( and so in particular @xmath297 ) .
now we use induction on @xmath62 , with the case @xmath298 being trivial . for the inductive step ,
we consider only the case where @xmath148 , @xmath289 and @xmath62 is odd ; the other cases are similar ( and often simpler ) .
let @xmath299 be the set of column labels of the nodes of @xmath296 . by the above remarks
, we must have @xmath297 in this case , and @xmath300 .
apply the spin branching rule , and consider the partitions @xmath301 such that @xmath302 .
these come in three types . 1 . the partition @xmath294 obtained from @xmath11 by removing the node at the bottom of column @xmath15 .
then @xmath303 , @xmath304 by the branching rule and by induction @xmath305 .
2 . there are @xmath306 different partitions @xmath294 obtained by removing a node in column @xmath307 , where neither @xmath308 nor @xmath309 lies in @xmath299 . in these cases
@xmath310 , @xmath311 by the spin branching rule , and by induction @xmath312 .
3 . for each pair
@xmath313 in @xmath299 , there is one @xmath8-regular partition@xmath294 obtained by removing a node in column @xmath314 or @xmath309 : if the nodes of @xmath296 lying in columns @xmath314 and @xmath309 are of the form @xmath315 , then @xmath294 is obtained by removing the node @xmath316 , while if they have the form @xmath317 , then @xmath294 is obtained by removing @xmath318 . in any case @xmath310 ,
@xmath311 by the spin branching rule and by induction @xmath319 .
so in total we find that @xmath320 which is what we need .
we give an example which illustrates a different case of [ spinbranchpower ] .
take @xmath321 and @xmath322 .
the @xmath53-spin - removable nodesof @xmath11 are highlighted in the following diagram .
@xmath323 the spin branching rule gives @xmath324 as predicted by [ spinbranchpower ] .
now we can deduce the following about @xmath325 for @xmath197 ; this will be central to the proof of our main theorem .
[ spinmaxbranch ] suppose @xmath197 and @xmath77 .
suppose @xmath11 has @xmath62 @xmath25-spin - removable nodes , and let @xmath285 be the partition obtained by removing all these nodes .
then : 1 . if @xmath326 , then @xmath327 for some @xmath83 ; 2 . if @xmath328 , then @xmath329 with @xmath330 .
analogous results apply to induced modules .
we have @xmath331 , where ( if @xmath159 lies in the block with content @xmath232 ) @xmath332 is the component of @xmath263 lying in the block with content @xmath333 , and @xmath334 is defined similarly . for a non - zero character @xmath159
we define @xmath335 , and we set @xmath336
. then we have analogues of [ classbrnch , spinbrnch , ordmaxbranch , spinbranchpower , spinmaxbranch ] ( which follow from these results by frobenius reciprocity ) , in which we add nodes rather than removing nodes .
we will not state these results explicitly , but instead refer to the `` induction versions '' of the restriction results above .
next we recall some of kleshchev s `` modular branching rules '' , to describe what happens to the irreducible brauer characters @xmath16 under the @xmath25-restriction operators .
this involves the combinatorics of _ normal nodes_. suppose @xmath337 and @xmath77 , and construct a sequence of @xmath338 and @xmath339 signs by reading along the edge of the young diagram of @xmath63 from top to bottom and writing a @xmath338 for each addable @xmath25-node and a @xmath339 for each removable @xmath25-node .
this sequence is called the _ @xmath25-signature _ of @xmath63 .
now successively delete all adjacent pairs @xmath340 from this sequence until none remain .
the resulting sequence is called the _ reduced @xmath25-signature _ of @xmath63 .
the removable nodes corresponding to the @xmath339 signs in the reduced @xmath25-signature are called the _
normal @xmath25-nodes _ of @xmath63 , and the addable nodes corresponding to the @xmath338 signs are the
_ conormal @xmath25-nodes _ of @xmath63 .
now we have the following ( see @xcite for this and more general modular branching results ) . [ modbranch ]
suppose @xmath337 and @xmath77 .
let @xmath341 be the @xmath8-regular partitionobtained by removing all the normal @xmath25-nodes from @xmath63 , and let @xmath342 be the partition obtained by adding all the conormal @xmath25-nodes to @xmath63 .
then @xmath343 are @xmath8-regular , and @xmath344 in particular , if @xmath159 is an irreducible brauer character of @xmath4 , then @xmath345 and @xmath346 are irreducible brauer characters .
the last statement in this theorem is extremely useful for us , and will form the basis of our induction proof of our main theorem .
let @xmath347 and @xmath297 .
the young diagram of @xmath63 with the residues of addable and removable nodes indicated , is as follows .
01 @xmath348 so the @xmath53-signature of @xmath63 is @xmath349 , and hence the reduced @xmath53-signature is @xmath350 , with the normal nodes being @xmath351 and @xmath352 , and the conormal node @xmath353 .
hence @xmath354 the combinatorics of partitions involved in representation theory is closely connected with the combinatorics involved in symmetric functions , and in this paper we shall exploit this connection . in this section we set out the notation and background results we shall need .
we just give the bare essentials , since symmetric function theory is covered in detail elsewhere .
we consider the space @xmath355 of symmetric functions over @xmath22 in infinitely many variables . for @xmath59 ,
let @xmath356 denote the corresponding _
schur function_. let @xmath357 be the inner product on @xmath355 for which the schur functions are orthonormal . for each @xmath62
, we let @xmath358 denote the complete homogeneous symmetric function of degree @xmath62 ( which coincides with the schur function @xmath359 ) . for any partition @xmath63
, we write @xmath360 for the product @xmath361 .
the _ pieri rule _ says that for any @xmath62 and for @xmath59 we have @xmath362 , summing over all partitions @xmath294 which can be obtained from @xmath11 by adding @xmath62 nodes in distinct columns .
inductively , this enables us to express any @xmath360 as a sum of schur functions ; the coefficients obtained are the _
kostka numbers_. as a consequence , @xmath363 is non - zero only if @xmath64 , and @xmath364 .
we write @xmath365 for the @xmath62th elementary symmetric function ( which coincides with @xmath366 ) , and for any @xmath108 we write @xmath367 .
the _ dual pieri rule _ says that for any @xmath62 and any @xmath11 we have @xmath368 , summing over all partitions @xmath294 which can be obtained from @xmath11 by adding @xmath62 nodes in distinct rows .
inductively , this enables us to express any @xmath369 as a sum of schur functions ( again , in terms of kostka numbers ) .
now @xmath370 is non - zero only if @xmath371 , and @xmath372 .
the combination of the pieri and co - pieri rules yields the identity @xmath373 .
the sets @xmath374 , @xmath375 and @xmath376 are all bases for @xmath355
. we shall need to consider the transition coefficients which allow us to express a complete homogeneous function @xmath377 in terms of the elementary functions @xmath369 .
so define @xmath378 for all @xmath379 by @xmath380 . in the case
where @xmath381 , these coefficients are given by a special case of the _ second jacobi
trudi formula _ , which says that @xmath382 equals the determinant of the @xmath383 matrix @xmath384 we now derive some consequences of this for partitions with only even parts .
[ ekdouble ] if @xmath385 , then @xmath386 for any partition @xmath63 .
using the second jacobi
trudi formula and expanding the determinant in terms of permutations , we find that @xmath387 is sum of the signatures of all the permutations @xmath388 with the property that @xmath389 for each @xmath25 and the integers @xmath390 equal @xmath391 in some order .
let @xmath392 be the set of such permutations .
then it suffices to show that there is a bijection from @xmath392 to @xmath393 which preserves the signature of every permutation if @xmath394 is even , or changes it if @xmath394 is odd . to construct the required map from @xmath392 to @xmath393 ,
we take @xmath395 , and define @xmath396 by @xmath397 to construct the inverse map , we take @xmath398 , and observe that because @xmath399 is odd and @xmath400 for every @xmath25 , we must have @xmath401 for all even @xmath25 .
now we can define @xmath402 by @xmath403 for each @xmath25 .
it is clear that these two maps are mutually inverse , so it remains to show that @xmath404 for every @xmath395 . regarding @xmath405 as an element of @xmath406 by the usual embedding of @xmath407 in @xmath406 , we can write @xmath408 , where @xmath409 are given by @xmath410 it is easy to see that the signatures of @xmath411 and @xmath412 are @xmath413 and @xmath414 respectively , which gives the result . by writing @xmath377 as the product @xmath415 and applying [ ekdouble ] , we obtain the following result .
[ ekdoublep ] suppose @xmath416
. then @xmath417 .
finally we note the following simple result .
[ simpleep ] suppose @xmath385 and @xmath418 with @xmath212 odd for at least one value of @xmath25
. then @xmath419 .
@xmath377 includes a factor @xmath420 ; when this is expressed in terms of elementary symmetric polynomials @xmath421 , each @xmath294 satisfies @xmath422 , and so must have at least one odd part .
hence each @xmath421 appearing when @xmath377 is expressed in terms of elementary symmetric polynomials must contain a factor @xmath423 for some odd @xmath201 .
having set out the background results we need , we are now able to state our main theorem .
this relies on another combinatorial definition : say that @xmath59 is _ @xmath8-carter _ if for every @xmath66 , @xmath424 is divisible by a power of @xmath8 greater than @xmath425 .
the carter partitions @xmath11 with @xmath426 are @xmath427 the notion of a carter partition was introduced for the classification of irreducible specht and weyl modules .
we shall need the following result later ( recall that @xmath31 is the schur algebra over a field @xmath19 of characteristic @xmath8 ) .
jm1theorem 4.5[irredweyl ] suppose @xmath116 .
then the weyl module @xmath121 for @xmath31 is irreducible @xmath11 is a @xmath8-carter partition .
now we come to the classification of ordinary irreducible characters for @xmath4 that remain irreducible in characteristic @xmath8 . for the characters @xmath143 of the specht modules ,
the answer is the same as for @xmath3 , and is given by the following theorem .
jmp2main theorem[jmp2main ] suppose @xmath59 .
then @xmath172 is irreducible either @xmath11 or @xmath60 is a @xmath8-carter partition , or @xmath428 .
so it remains to consider spin characters .
recall that when @xmath149 we write @xmath10 to mean either @xmath13 or @xmath14 , and @xmath216 is then unambiguously defined , since @xmath429 .
so we can phrase the main question for spin characters simply as `` for which @xmath197 is @xmath216 irreducible ? '' .
now we can state our main theorem .
[ main ] suppose @xmath197
. then @xmath216 is irreducible one of the following occurs . 1 .
@xmath430 , where @xmath431 for some @xmath200 and @xmath432 is a @xmath8-carter partition with @xmath433 .
@xmath430 , where @xmath434 for some @xmath435 and @xmath432 is a @xmath8-carter partition with @xmath433 .
@xmath436 , where @xmath431 for some @xmath200 and @xmath432 is a @xmath8-carter partition with @xmath433 .
@xmath436 , where @xmath434 for some @xmath435 and @xmath432 is a @xmath8-carter partition with @xmath437 .
@xmath11 equals @xmath438 or @xmath439 for some @xmath440 .
@xmath441 .
[ degreesec , rouqandsep , mainproofsec ] are devoted to the proof of this theorem .
in this section we use the bar - length formula to compare the degrees of certain characters .
the motivating observation here is the following .
[ samerdoub ] suppose @xmath163 , with @xmath218 and @xmath442 .
then @xmath216 is reducible . by [ spinreg ]
, @xmath443 occurs as a constituent of @xmath216 , so if @xmath216 were irreducible , we would have @xmath217 , and in particular @xmath444 . but
@xmath445 also occurs as a constituent of @xmath446 , so @xmath447 . to exploit this , we construct various families of pairs of partitions @xmath168 satisfying the hypotheses of [ samerdoub ] .
our first result is as follows .
[ dimen1 ] suppose @xmath448 , and let @xmath449 then @xmath450 , and @xmath451 .
@xmath452 is obtained from @xmath453 by replacing the node @xmath454 with the node @xmath455 .
since these nodes both lie in the same slope , we have @xmath450 by [ slopelad ] .
now we consider the degrees .
we use induction on @xmath198 , with the case @xmath456 an easy check . for the inductive step
, it suffices to prove that when @xmath457 @xmath458 directly from the bar - length formula , this ratio equals @xmath459 to show that this is always greater than @xmath15 , we show that the numerator exceeds the denominator for all @xmath448 .
the difference between the numerator and the denominator is @xmath460 .
we have @xmath461 and @xmath462 positive , while @xmath463 and @xmath464 are negative , so that all the roots of @xmath465 are less than @xmath8 .
now we give a similar dimension argument which will help when we consider rouquier blocks .
[ firstrdim ] given @xmath466 and @xmath467 , define @xmath468 then @xmath469 , and @xmath470 . to see that @xmath469 , let @xmath140 be the set of nodes of @xmath471 comprising
* the last four nodes in each of rows @xmath472 , and * the unique node in row @xmath473 .
now observe that @xmath474 can be obtained from @xmath471 by replacing each node @xmath475 with @xmath476 .
since @xmath72 and @xmath476 lie in the same slope , the result follows .
now we consider degrees .
direct from the bar - length formula we get @xmath477 now we proceed by induction on @xmath100 . in the case
@xmath478 , ( ) becomes @xmath479 , which is greater than @xmath15 .
for the inductive step , it suffices to show that @xmath480 from ( ) , this ratio equals @xmath481 the difference between the numerator and denominator in this fraction is @xmath482 which is obviously positive .
the next three results are proved in exactly the same way .
[ secondrdim ] given @xmath466 and @xmath467 , define @xmath483 then @xmath469 , and @xmath470 .
[ thirdrdim ] given @xmath466 and @xmath467 , define @xmath484 then @xmath469 , and @xmath470 .
[ fourthrdim ] given @xmath466 and @xmath467 , define @xmath485 then @xmath469 , and @xmath470 . to make these results more general ,
we now consider row - removal .
we start with the following lemma .
[ domdim ] suppose @xmath125 with @xmath486 , and that @xmath201 is an integer greater than @xmath487 .
then @xmath488 we may assume that @xmath63 covers @xmath11 in the dominance order , in which case @xmath11 is obtained from @xmath63 by moving a single node down to a lower row .
so suppose that for some @xmath489 we have @xmath490 , @xmath491 , @xmath492 , @xmath493 , and that @xmath494 for all @xmath495 .
then the ratio of the left - hand expression to the right - hand expression is @xmath496 the difference between the numerator and denominator here is @xmath497 , which is positive , so the ratio is greater than @xmath15 .
[ dimrowrem ] suppose @xmath163 and @xmath201 is an odd integer with @xmath498 .
define @xmath499 then @xmath500 furthermore , if @xmath218 , then @xmath501 . from the bar - length formula , the ratio of the left - hand side to the right - hand side is @xmath502 and by [ domdim ] this is greater than @xmath15 .
the second statement follows by considering the slopes containing the nodes of @xmath11 and the corresponding nodes of @xmath503 .
rouquier blocks are a certain class of particularly well - behaved blocks of symmetric groups ( and more generally of iwahori hecke algebras and @xmath30-schur algebras ) . in this section
we summarise some of the important properties of rouquier blocks of symmetric groups ( and hence of their double covers ) in characteristic @xmath8 , and then examine the decomposition numbers for rouquier blocks of @xmath4 .
we compute the rows of the decomposition matrix labelled by @xmath8-regular partitionswith only odd parts , showing that some of the corresponding spin characters are irreducible in characteristic @xmath8 .
we extend these results to what we call `` separated '' @xmath8-regular partitions , and prove our main theorem for spin characters labelled by separatedpartitions .
suppose @xmath230 is a block of @xmath4 , with core @xmath504 and weight @xmath114 .
we say that @xmath230 is _ rouquier _ if @xmath505 .
the first thing that makes rouquier blocks easy to understand is a simple description of the @xmath8-regular partitionslabelling characters in a rouquier block . [
trpsinrouq ] suppose @xmath504 and that @xmath11 is a partition with @xmath8-core @xmath236 and @xmath8-weight @xmath505 . . 1
. @xmath11 is @xmath8-regular .
the length of @xmath11 is at most @xmath506 .
3 . @xmath11 has the form @xmath507 , where @xmath508 .
( 1@xmath5092 ) : : suppose @xmath11 is @xmath8-regular , and suppose for a contradiction that the length of @xmath11 is greater than @xmath506 .
then @xmath510 and @xmath511 are nodes of @xmath11 , so ( since @xmath11 is @xmath8-regular ) @xmath512 are all nodes of @xmath11 .
but then @xmath513 , a contradiction .
( 2@xmath5093 ) : : we use induction on @xmath114 , with the case @xmath514 being trivial . assuming @xmath515 ,
let @xmath516 be maximal such that @xmath517 ( there must be such an @xmath62 , since @xmath518 ) .
then we claim that @xmath519 : note that @xmath520 is a node of @xmath11 but not of @xmath236 , so must be part of one of the rim @xmath8-hooks added to obtain @xmath11 from @xmath236 .
the other node in this rim @xmath8 must be @xmath521 , since @xmath522 and @xmath523 are not nodes of @xmath524 , and ( by the choice of @xmath62 ) neither is @xmath525 .
hence @xmath519 . by the choice of @xmath62 either @xmath526 or @xmath527 , which means that we can remove the rim @xmath8-hook @xmath528 from @xmath11 and leave a partition @xmath63 which is still @xmath8-regular . by induction @xmath529 for some @xmath530 , and hence @xmath531 , where @xmath432 is the partition obtained from @xmath532 by adding a node at the end of row @xmath62 .
( 3@xmath5091 ) : : suppose @xmath531 with @xmath508 .
then the length of @xmath432 is at most @xmath505 .
so @xmath533 for @xmath534 , and @xmath535 for @xmath536 , and hence @xmath11 is @xmath8-regular .
we can also describe the @xmath8-regular partitionslabelling spin characters in rouquier blocks .
this requires some preliminary work .
[ rouqlem3 ] suppose @xmath431 for @xmath200 , and @xmath11 is a @xmath8-regular partitionwith @xmath237-bar - core@xmath242 and @xmath237-bar - weight@xmath537 .
then @xmath11 has no parts congruent to @xmath15 modulo @xmath237 , and @xmath11 includes all the integers @xmath538 , where @xmath539 .
we use induction on @xmath114 . in the case
@xmath514 we have @xmath540 and the result holds . now take @xmath515 and suppose the result holds for smaller @xmath114 . by the definition of @xmath237-bar - core , there is a @xmath8-regular partition@xmath63 satisfying one of the following .
* @xmath63 is obtained from @xmath11 by reducing some @xmath212 by @xmath237 ( and re - ordering ) . by the inductive hypothesis @xmath63
has no parts congruent to @xmath15 modulo @xmath237 , and hence neither does @xmath11 .
@xmath63 has @xmath237-bar - weight@xmath541 , so by hypothesis @xmath63 includes all the integers @xmath542 .
hence @xmath11 includes the integers @xmath543 .
* @xmath544 . in this case
, the assumption that @xmath63 satisfies the given conditions means that @xmath11 does too .
* @xmath545 .
this means that @xmath546 ; but since @xmath63 has @xmath237-bar - weight@xmath541 the inductive hypothesis says that @xmath547 , so that @xmath548 , i.e. @xmath549 . on the other hand
, the fact that @xmath230 is a rouquier block means that @xmath550 , a contradiction .
[ rouqspinform3 ] suppose @xmath230 is a rouquier block of @xmath4 with bar - core@xmath431 and @xmath237-bar - weight@xmath114 , and that @xmath119 such that @xmath10 lies in @xmath230
. then @xmath11 can be written in the form @xmath551 , where @xmath552 , @xmath553 and @xmath554 .
furthermore , if @xmath555 is any even part of @xmath11 , then @xmath11 includes all positive integers less than @xmath555 which are congruent to @xmath157 modulo @xmath237 .
if we sort the parts of @xmath11 according to parity , we can write @xmath556 where all the positive parts of @xmath294 are odd . by [ rouqlem3 ] all the positive parts of @xmath294 are congruent to @xmath157 modulo @xmath237 , so if we let @xmath557 and let @xmath558 denote the @xmath237-bar - core@xmath559 , then we have @xmath560 for some partition @xmath432 .
but then the @xmath237-bar - coreof @xmath294 ( and hence of @xmath11 ) is clearly @xmath558 , and so we have @xmath561 .
so we can write @xmath562 , and it follows that @xmath554 .
the @xmath237-bar - weightof @xmath294 is @xmath563 , so applying the second statement of [ rouqlem3 ] with @xmath294 in place of @xmath11 , we find that @xmath294 ( and hence @xmath11 ) contains all the integers @xmath564 .
the fact that @xmath230 is a rouquier block means that @xmath537 .
now any even part of @xmath11 equals @xmath565 for some @xmath566 , and @xmath567 hence @xmath11 contains all positive integers less than @xmath565 which are congruent to @xmath157 modulo @xmath237 . in a very similar way
, we obtain the following .
[ rouqlem1 ] suppose @xmath434 for @xmath435 , and @xmath11 is a @xmath8-regular partitionwith @xmath237-bar - core@xmath242 and @xmath237-bar - weight@xmath568 .
then @xmath11 has no parts congruent to @xmath157 modulo @xmath237 , and @xmath11 includes all the integers @xmath569 , where @xmath539 .
[ rouqspinform1 ] suppose @xmath230 is a rouquier block of @xmath4 with bar - core@xmath431 and @xmath237-bar - weight@xmath114 , and that @xmath119 such that @xmath10 lies in @xmath230
. then @xmath11 can be written in the form @xmath551 , where @xmath552 , @xmath553 and @xmath554 .
furthermore , if @xmath555 is any even part of @xmath11 , then @xmath11 includes all positive integers less than @xmath555 which are congruent to @xmath15 modulo @xmath237 .
note in particular that by [ rouqspinform3,rouqspinform1 ] , if @xmath197 has no even parts and @xmath10 lies in a rouquier block , then the @xmath237-bar - weightof @xmath11 must be even .
an advantage of working with rouquier blocks is that their decomposition numbers are relatively well understood .
we begin with hecke algebras , where the decomposition numbers for rouquier blocks are known explicitly in terms of littlewood richardson coefficients .
this result is due to james and mathas ( * ? ? ?
* corollary 2.6 ) ; here we only give two special cases .
[ heckerouqdec ] suppose @xmath504 and @xmath505 . 1 . if @xmath570 , then @xmath571 now we consider rouquier blocks of symmetric groups .
the following result is due to turner .
turntheorem 132[snrouqdec ] suppose @xmath230 is a rouquier block of @xmath3 with core @xmath504 and weight @xmath505 .
then @xmath572 for all @xmath168 in @xmath573 .
hence the adjustment matrix @xmath140 for the block @xmath230 is just the decomposition matrix of @xmath574 , i.e. @xmath575 for @xmath168 in @xmath573 .
we now examine the decomposition numbers in a rouquier block by considering projective characters .
we fix some notation .
* assumptions and notation in force for [ projcharsec , somedecevensec ] : * @xmath230 is a rouquier block of @xmath4 with core @xmath236 and weight @xmath114 .
@xmath242 is the bar - coreof @xmath230 , i.e. the common @xmath237-bar - coreof the @xmath8-regular partitionslabelling spin characters in @xmath230 .
recall that a projective character is one which vanishes on @xmath8-singular elements ; a _ virtual _ projective character is a @xmath576-linear combination of projective characters .
it is well known that induction and restriction send ( virtual ) projective characters to ( virtual ) projective characters , and hence so do the functors @xmath577 and @xmath578 .
recall that @xmath158 is the usual inner product on ordinary characters , and that @xmath357 is the standard inner product on the space of symmetric functions .
[ induceproj ] suppose @xmath579 .
there is a projective character @xmath580 of @xmath4 with the following properties .
* @xmath581 for any @xmath582 .
in particular , @xmath583 , while @xmath584 if @xmath585 .
* if the column lengths of @xmath63 are all even , say @xmath586 , then @xmath587 for every @xmath588 .
in particular , @xmath589 while @xmath590 if @xmath591 .
* if @xmath114 is even but @xmath63 has at least one column of odd length , then @xmath590 for every @xmath588 .
we prove [ induceproj ] by applying induction functors .
note that all the addable nodes of @xmath236 have the same residue ; we assume for the rest of this section that they have residue @xmath53 .
( the proof in the opposite case is identical , but with @xmath53 and @xmath15 swapped throughout . )
given @xmath592 , consider the induction functor @xmath593 .
the classical branching rule gives the following .
[ induceord ] suppose @xmath594 , and @xmath595 is a partition with @xmath8-core @xmath236 and @xmath8-weight @xmath596 .
suppose @xmath597 . 1 .
if @xmath595 is @xmath8-regular , say @xmath598 , then @xmath599 .
2 . if @xmath595 is @xmath8-singular , then @xmath600 .
this result is effectively a special case of ( * ? ? ?
* lemma 3.1 ) , but exploiting this requires a lot of translation of notation and transfer of results from one context to another , so we give a full proof here . first note that @xmath601 lies in the block with core @xmath236 and weight @xmath602 , and the condition @xmath603 guarantees that this is a rouquier block .
now the branching rule says that @xmath601 is the sum of @xmath604 over all pairs of partitions @xmath605 such that @xmath606 .
this gives ( 2 ) straight away , because if @xmath595 is @xmath8-singular then by [ trpsinrouq ] the length of @xmath595 is at least @xmath607 , so @xmath608 .
so we are left with ( 1 ) .
suppose @xmath595 is @xmath8-regular and write @xmath598 . if @xmath294 is obtained from @xmath412 by adding @xmath62 nodes in distinct rows , say rows @xmath609 , then let @xmath405 be the partition obtained from @xmath595 by adding one node at the end of each of these rows .
then ( since all the addable nodes of @xmath595 have residue @xmath53 ) we have @xmath606 , so @xmath604 occurs in @xmath601 , and clearly occurs once only .
conversely , if we have @xmath606 , then @xmath610 , so @xmath611 .
moreover , the nodes added to @xmath595 to obtain @xmath405 all have the same residue , so lie in different rows , and so @xmath612 differs from @xmath595 in at least @xmath62 rows .
so @xmath294 differs from @xmath412 in at least @xmath62 rows , so @xmath294 must be obtained from @xmath412 by adding @xmath62 nodes in distinct rows .
now we prove a corresponding result for spin characters .
recall that we write @xmath613 to mean `` either @xmath614 or @xmath615 '' when @xmath616 .
[ inducespin ] suppose @xmath594 with @xmath602 even , and @xmath595 is a @xmath8-regular partitionwith @xmath237-bar - core@xmath242 and @xmath237-bar - weight@xmath596 .
suppose @xmath617 . 1 .
suppose all the parts of @xmath595 are odd , say @xmath618 for @xmath619
. then @xmath620 .
2 . if @xmath595 has at least one even part ( and in particular if @xmath62 is odd )
, then @xmath621 .
let @xmath622 . by the spin branching rule
, @xmath623 occurs in @xmath624 if and only if there is @xmath405 such that @xmath625 , i.e. @xmath595 can be obtained from @xmath612 by removing @xmath62 nodes of spin residue@xmath15 followed by @xmath62 nodes of spin residue@xmath53 .
the form of @xmath612 means that there are two nodes of spin residue@xmath15 at the end of each non - empty row of @xmath612 ; removing both of these nodes from a row then exposes two nodes of spin residue@xmath53 at the end of that row ( or one , if the row initially only has length @xmath157 ) .
so the only way @xmath405 can have @xmath62 nodes of spin residue@xmath53 that can be removed is if @xmath62 is even , and the nodes removed from @xmath612 to obtain @xmath405 occur in pairs at the end of @xmath626 different rows ; then the only possible way to remove @xmath62 nodes of spin residue@xmath53 from @xmath405 is to remove them in pairs from these same @xmath626 rows . thus if @xmath625 , then we must have @xmath627 for @xmath626 different values of @xmath25 , with @xmath628 for all other values of @xmath25
. hence @xmath618 , where @xmath432 is obtained from @xmath532 by removing @xmath626 nodes from distinct rows .
this is sufficient to prove the , except that in the case where @xmath532 is obtained from @xmath432 by adding nodes in distinct rows we must show that the coefficient of @xmath623 in @xmath629 is @xmath15 . but
this follows from the induction version of [ spinbranchpower ] .
now we can prove [ induceproj ] .
we use induction on @xmath114 . in the case
@xmath514 , we can define @xmath630 to be the unique indecomposable projective character in the block with core @xmath236 and weight @xmath53 , i.e. @xmath631 . by [ jreg , spinreg ] ,
this equals @xmath632 .
now suppose @xmath633 , and let @xmath612 be the partition obtained by removing the last non - empty column ( of length @xmath62 , say ) from @xmath63 , so that @xmath634 . by the inductive hypothesis the holds with @xmath612 in place of @xmath63 , and we define @xmath635 . we must check that the conditions given in the hold for @xmath580 , given that they hold for @xmath636 .
suppose @xmath582 . by [ induceord ]
we have @xmath637 now suppose @xmath114 is even and @xmath588 .
if @xmath62 is odd , then by [ inducespin](2 ) we have @xmath638 so suppose instead that @xmath62 is even .
if any of the column lengths of @xmath612 are odd , then by the inductive hypothesis @xmath639 for every @xmath640 , so by [ inducespin](2 ) @xmath590 .
on the other hand , if the column lengths of @xmath63 are all even , then write @xmath586 , so that @xmath641 where @xmath642 is obtained by removing the last column from @xmath432 . then @xmath643 we take @xmath644 and @xmath645 , so that @xmath646
. we will show how to construct the characters @xmath647 and @xmath648 . in both cases
we start from @xmath649 . applying @xmath650 to @xmath630
, we obtain @xmath651 where @xmath652 is a sum of characters of the form @xmath653 with @xmath294 @xmath8-singular , and @xmath654 is a sum of characters @xmath613 for @xmath595 not of the form @xmath655 .
note that the characters @xmath656 and @xmath657 do not contribute any terms @xmath653 with @xmath294 @xmath8-regular , and the characters @xmath658 do not contribute any terms @xmath659 .
now we look at @xmath660 . starting with @xmath630 and applying @xmath661
, we obtain @xmath662 where @xmath159 is a sum of characters of the form @xmath653 with @xmath294 @xmath8-singular , or @xmath613 with @xmath595 not of the form @xmath655 .
next we want to prove a kind of dual result to [ induceproj ] , using virtual projective characters .
[ induceprojdual ] suppose @xmath663 .
there is a virtual projective character @xmath664 with the following properties .
* @xmath665 for any @xmath582 . *
if @xmath212 is even for every @xmath25 , say @xmath666 , then @xmath667 for every @xmath588 .
* if @xmath114 is even but @xmath212 is odd for some @xmath25 , then @xmath668 for every @xmath588 . recall the coefficients @xmath378 defined by @xmath380 , and define @xmath669 now for @xmath582 we have @xmath670 now suppose @xmath114 is even .
then for @xmath588 we have @xmath671 by [ induceproj ] . if @xmath212 is odd for some @xmath25 , then by [ simpleep ] @xmath672 for every @xmath432 , so that @xmath668 . on the other hand ,
if @xmath666 , then @xmath673 by [ ekdoublep ] , so that @xmath674 now we come to the main aim of this section , which is to examine another family of virtual projective characters . by [ trpsinrouq ] , the indecomposable projective characters are the characters @xmath675 for @xmath579 , and by [ snrouqdec ] these satisfy @xmath676 for all @xmath570 .
the invertibility of the decomposition matrix of the schur algebra then implies that for each @xmath63 there is a unique virtual projective character @xmath677 such that @xmath678 for each @xmath11 .
in fact , we can write @xmath679 where ( as set out in [ inversesec ] ) @xmath137 is the decomposition matrix of the schur algebra in characteristic @xmath8 , i.e. @xmath680}$ ] . our aim is to study the multiplicities of the spin characters @xmath659 in the @xmath677 when @xmath114 is even , which will enable us to deduce information about the decomposition numbers .
the triangularity of the projective characters @xmath580 introduced in [ induceproj ] means that each @xmath677 can be written uniquely as a linear combination of the @xmath580 .
so define coefficients @xmath681 by @xmath682 .
the next result shows that these coefficients can also be used to write @xmath683 in terms of the virtual characters @xmath684 .
[ omegaups ] for any @xmath663 , @xmath685 for any @xmath582 we have @xmath686 now the uniqueness property defining @xmath683 gives the result .
now we can deduce the following symmetry property of the coefficients @xmath687 .
[ virtualsym ] suppose @xmath114 is even , @xmath663 and @xmath588 .
then @xmath688 on the one hand , we have @xmath689 replacing @xmath690 with @xmath691 gives the result .
we now use the results of the previous subsection to find some explicit decomposition numbers for spin characters in rouquier blocks .
we retain the assumptions set out at the start of [ projcharsec ] , and _ we assume throughout [ somedecevensec ] that @xmath114 is even_. our aim is to study the decomposition numbers @xmath692 , for @xmath693 and @xmath579 , and to express them in terms of the adjustment matrix of the schur algebra .
we start by defining several matrices .
let @xmath694 be the matrix with rows indexed by @xmath695 and columns by @xmath573 , with @xmath696 let @xmath137 be the decomposition matrix of the schur algebra @xmath574 in characteristic @xmath8 , i.e. @xmath697 $ ] . then ( as explained in [ inversesec ] ) @xmath141 , where @xmath139 is the decomposition matrix of the @xmath124-schur algebra @xmath698 and @xmath140 is its adjustment matrix .
now we can state the main result of this section .
[ allodd ] suppose @xmath230 is a rouquier block of even weight @xmath114 , and define @xmath699 as above
. then @xmath700 .
for example , if @xmath644 then we have the following matrices , and we see that [ allodd ] is true in this case .
@xmath701 d&=\begin{array}{@{}r@{\,}|@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c } & { \rotatebox{90}{$(4)$}}&{\rotatebox{90}{$(3,1)$}}&{\rotatebox{90}{$(2 ^ 2)$}}&{\rotatebox{90}{$(2,1 ^ 2)$}}&{\rotatebox{90}{$(1 ^ 4)$}}\\\hline ( 4)&1&\cdot&\cdot&\cdot&\cdot\\ ( 3,1)&1&1&\cdot&\cdot&\cdot\\ ( 2 ^ 2)&\cdot&1&1&\cdot&\cdot\\ ( 2,1 ^ 2)&1&1&1&1&\cdot\\ ( 1 ^ 4)&1&\cdot&1&1&1
\end{array}&\qquad { \mathring d}&=\begin{array}{@{}r@{\,}|@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c } & { \rotatebox{90}{$(4)$}}&{\rotatebox{90}{$(3,1)$}}&{\rotatebox{90}{$(2 ^ 2)$}}&{\rotatebox{90}{$(2,1 ^ 2)$}}&{\rotatebox{90}{$(1 ^ 4)$}}\\\hline ( 4)&1&\cdot&\cdot&\cdot&\cdot\\ ( 3,1)&1&1&\cdot&\cdot&\cdot\\ ( 2 ^ 2)&\cdot&1&1&\cdot&\cdot\\ ( 2,1 ^ 2)&1&1&1&1&\cdot\\ ( 1 ^ 4)&1&\cdot&\cdot&1&1 \end{array } \\[6pt ] a&=\begin{array}{@{}r@{\,}|@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c } & { \rotatebox{90}{$(4)$}}&{\rotatebox{90}{$(3,1)$}}&{\rotatebox{90}{$(2 ^ 2)$}}&{\rotatebox{90}{$(2,1 ^ 2)$}}&{\rotatebox{90}{$(1 ^ 4)$}}\\\hline ( 4)&1&\cdot&\cdot&\cdot&\cdot\\ ( 3,1)&\cdot&1&\cdot&\cdot&\cdot\\ ( 2 ^ 2)&\cdot&\cdot&1&\cdot&\cdot\\ ( 2,1 ^ 2)&\cdot&\cdot&\cdot&1&\cdot\\ ( 1 ^ 4)&\cdot&\cdot&1&\cdot&1 \end{array}&&\end{aligned}\ ] ] [ induceproj ] gives us the following information about the matrix @xmath694 , which shows that [ allodd ] is true up to a triangular adjustment .
[ triang ] suppose @xmath588 and @xmath579 . 1 .
if @xmath63 has at least one column of odd length , then @xmath702 .
2 . if the columns of @xmath63 all have even length , say @xmath586 , then @xmath703 while @xmath702 unless @xmath704 .
@xmath705 is the coefficient @xmath706 .
so we examine the projective characters @xmath707 , for @xmath663 . by the triangularity of the decomposition matrix of the symmetric group we have @xmath708 by [ induceproj ] the same property holds
if we replace @xmath675 with the projective character @xmath580 introduced in that . since @xmath580 is a linear combination of the characters @xmath707 with non - negative coefficients , we can obtain @xmath675 from @xmath580 by subtracting multiples of the characters @xmath709 for @xmath710 .
that is , @xmath711 with each @xmath712 a non - negative integer . in particular , for any @xmath588 @xmath713 so ( by [ induceproj ] ) @xmath714 if @xmath63 has at least one column of odd length or @xmath586 with @xmath591 .
furthermore , in the case where @xmath586 , the coefficient @xmath715 is zero for any @xmath710 by induction on @xmath63 ( using the dominance order ) , and so @xmath716 now we note that the same property holds for the matrix @xmath717 .
[ ehv ] suppose @xmath588 and @xmath579 . 1 .
if @xmath63 has at least one column of odd length , then @xmath718 .
if the columns of @xmath63 all have even length , say @xmath586 , then @xmath719 while @xmath718 unless @xmath704 .
we have @xmath720 , and by [ triang](1 ) we may restrict the range of summation to those @xmath11 with all columns of even length ; that is , @xmath721 if @xmath63 has a column of odd length , then by [ adj ] @xmath722 for every @xmath612 , which gives ( 1 ) . if @xmath586 , then by [ adj ] @xmath723 where @xmath724 is the decomposition matrix of the schur algebra @xmath725 .
the term @xmath726 is non - zero only if @xmath727 ( and equals @xmath15 when @xmath728 ) , while the term @xmath729 is non - zero only if @xmath730 ( and is @xmath15 if @xmath731 ) .
the result follows .
[ ehv ] may be alternatively phrased as follows : there is a square matrix @xmath732 with rows and columns indexed by @xmath695 , such that @xmath733 .
furthermore , @xmath732 is lower unitriangular in the sense that @xmath734 for each @xmath432 , while @xmath735 for @xmath736 .
our aim is to prove that @xmath732 is the identity matrix . to do this , we consider the matrices @xmath737 and @xmath738 , and show that they both satisfy a certain symmetry property . given a matrix @xmath230 with rows indexed by @xmath695 and columns by @xmath573 , we say that @xmath230 is _ conjugate - symmetric _ if @xmath739 for every @xmath693 and @xmath663 . [ jfunny ]
the matrix @xmath738 is conjugate - symmetric .
the definition of @xmath740 means that @xmath741 , and this is given explicitly by [ steinberg ] as @xmath742 , where @xmath743 is defined in [ inversesec ] .
now @xmath63 and @xmath110 have the same @xmath8-core , and if this @xmath8-core if not @xmath54 then @xmath744 , so that @xmath745 .
so assume that @xmath63 and @xmath110 have empty @xmath8-core , and therefore have @xmath8-weight @xmath746 .
then by [ parityconj ] we have @xmath747 , so we just need to show that @xmath748 for any @xmath432 . since @xmath749 for any @xmath180 , the result follows from [ quoconj ] .
[ efunny ] the matrix @xmath737 is conjugate - symmetric .
recall that @xmath750 is the multiplicity @xmath751 .
recall also that the virtual projective characters @xmath677 from [ projcharsec ] are given by @xmath752 this means that @xmath753 .
now [ virtualsym ] gives the required result .
for example , in the case @xmath644 the matrix @xmath754 is the following .
@xmath755 now we can complete the proof of [ allodd ] .
we have seen that there is a matrix @xmath732 such that @xmath733 , and that @xmath732 is unitriangular in the sense that @xmath734 and @xmath735 if @xmath756 .
so to prove that @xmath732 is the identity matrix , we just need to prove that @xmath735 for all @xmath757 , and we do this by induction on @xmath532 using the dominance order .
so assume that we have proved that @xmath758 whenever @xmath759 .
multiplying the equation @xmath733 by @xmath187 , we obtain @xmath760 .
consider the column of @xmath738 labelled by the partition @xmath761 .
for any @xmath690 we have @xmath762 , and the triangularity of @xmath139 means that @xmath763 if @xmath764 , and @xmath53 if @xmath765 .
the triangularity of @xmath732 then means that the same is true with @xmath738 replaced by @xmath766 . using the fact that both @xmath738 and @xmath737 are conjugate - symmetric(and the fact that the dominance order is reversed by conjugating partitions ) we get @xmath767 and the same with @xmath737 in place of @xmath738 . hence for @xmath757 we have @xmath768 the term @xmath769 is zero unless @xmath770 . if @xmath771 then by our induction hypothesis @xmath758 , so we only need to consider the term with @xmath772 , and we obtain @xmath773 .
hence @xmath732 is the identity matrix , and therefore @xmath700 .
we now prove a similar result to [ allodd ] for the case where @xmath114 is odd .
we retain the assumptions set out at the start of [ projcharsec ] , and _ we assume throughout [ somedecoddsec ] that @xmath114 is odd_. as before , we let @xmath137 be the decomposition matrix of the schur algebra @xmath574 , and factorise @xmath137 as @xmath774 , where @xmath139 is the decomposition matrix of the corresponding @xmath124-schur algebra and @xmath140 is the adjustment matrix .
we now define @xmath694 to be the matrix with rows indexed by @xmath775 and columns by @xmath573 , and @xmath776 now we have the following . [ allodd2 ]
suppose @xmath230 is a rouquier block of odd weight @xmath114 , and define @xmath699 as above
. then @xmath700 .
for example , taking @xmath777 , we have the following matrices .
@xmath778 d&=\begin{array}{@{}r@{\,}|@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c } \phantom{(3,1 ^ 2)}&{\rotatebox{90}{$(5)$}}&{\rotatebox{90}{$(4,1)$}}&{\rotatebox{90}{$(3,2)$}}&{\rotatebox{90}{$(3,1 ^ 2)$}}&{\rotatebox{90}{$(2 ^ 2,1)$}}&{\rotatebox{90}{$(2,1 ^ 3)$}}&{\rotatebox{90}{$(1 ^ 5)$}}\\\hline ( 5)&1&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot\\ ( 4,1)&\cdot&1&\cdot&\cdot&\cdot&\cdot&\cdot\\ ( 3,2)&1&\cdot&1&\cdot&\cdot&\cdot&\cdot\\ ( 3,1 ^ 2)&2&\cdot&1&1&\cdot&\cdot&\cdot\\ ( 2 ^ 2,1)&1&\cdot&1&1&1&\cdot&\cdot\\ ( 2,1 ^ 3)&\cdot&1&\cdot&\cdot&\cdot&1&\cdot\\ ( 1 ^
5)&1&\cdot&\cdot&1&1&\cdot&1 \end{array}&\qquad { \mathring d}&=\begin{array}{@{}r@{\,}|@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c } \phantom{(3,1 ^ 2)}&{\rotatebox{90}{$(5)$}}&{\rotatebox{90}{$(4,1)$}}&{\rotatebox{90}{$(3,2)$}}&{\rotatebox{90}{$(3,1 ^ 2)$}}&{\rotatebox{90}{$(2 ^ 2,1)$}}&{\rotatebox{90}{$(2,1 ^ 3)$}}&{\rotatebox{90}{$(1 ^ 5)$}}\\\hline ( 5)&1&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot\\ ( 4,1)&\cdot&1&\cdot&\cdot&\cdot&\cdot&\cdot\\ ( 3,2)&\cdot&\cdot&1&\cdot&\cdot&\cdot&\cdot\\ ( 3,1 ^ 2)&1&\cdot&1&1&\cdot&\cdot&\cdot\\ ( 2 ^ 2,1)&\cdot&\cdot&1&1&1&\cdot&\cdot\\ ( 2,1 ^ 3)&\cdot&1&\cdot&\cdot&\cdot&1&\cdot\\ ( 1 ^ 5)&1&\cdot&\cdot&1&\cdot&\cdot&1 \end{array}\\[6pt ] a&=\begin{array}{@{}r@{\,}|@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c@{\,}c } \phantom{(3,1 ^ 2)}&{\rotatebox{90}{$(5)$}}&{\rotatebox{90}{$(4,1)$}}&{\rotatebox{90}{$(3,2)$}}&{\rotatebox{90}{$(3,1 ^ 2)$}}&{\rotatebox{90}{$(2 ^ 2,1)$}}&{\rotatebox{90}{$(2,1 ^ 3)$}}&{\rotatebox{90}{$(1 ^ 5)$}}\\\hline ( 5)&1&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot\\ ( 4,1)&\cdot&1&\cdot&\cdot&\cdot&\cdot&\cdot\\ ( 3,2)&1&\cdot&1&\cdot&\cdot&\cdot&\cdot\\ ( 3,1 ^ 2)&\cdot&\cdot&\cdot&1&\cdot&\cdot&\cdot\\ ( 2 ^ 2,1)&\cdot&\cdot&\cdot&\cdot&1&\cdot&\cdot\\ ( 2,1 ^ 3)&\cdot&\cdot&\cdot&\cdot&\cdot&1&\cdot\\ ( 1 ^ 5)&\cdot&\cdot&\cdot&\cdot&1&\cdot&1 \end{array}&&\end{aligned}\ ] ] to prove [ allodd2 ] , we use restriction , comparing @xmath230 with the corresponding block of weight @xmath779 and exploiting [ allodd ] .
we define @xmath780 to be : * @xmath781 if @xmath236 is one of @xmath782 ( and hence @xmath242 is one of @xmath783 ) ; * @xmath784 if @xmath236 is one of @xmath785 ( and hence @xmath242 is one of @xmath786 ) . as in the even - weight case , for each @xmath663 we define @xmath683 to be the unique virtual projective character for which @xmath787 .
now consider applying @xmath780 to a character in @xmath230 , and examining the coefficients of @xmath788 and @xmath789 in this restriction .
[ ebullord ] suppose @xmath790 and @xmath791 with @xmath792 in @xmath230 .
then @xmath793 equals : * @xmath15 , if @xmath794 where @xmath63 is obtained by adding a node to @xmath11 ; * @xmath15 , if @xmath795 ; * @xmath53 , otherwise .
this is a simple application of the branching rule .
assume that @xmath796 ( the other case is very similar ) .
then by the branching rule @xmath793 equals @xmath15 if there is a partition @xmath405 with @xmath797 , and @xmath53 otherwise .
since @xmath798 , all the addable nodes of @xmath799 have residue @xmath53 , so if @xmath405 exists then the two nodes added to @xmath799 to obtain @xmath595 must be adjacent .
if these nodes are both in the same row , say row @xmath25 , then @xmath794 , where @xmath63 is obtained from @xmath11 by adding a node in row @xmath25 . on the other hand , if the added nodes are both in the same column , then ( since @xmath799 is @xmath8-regular ) this must be column @xmath15 , and hence @xmath795 .
[ ebullomega ] suppose @xmath579 and @xmath790
. then @xmath800 equals @xmath8 if @xmath801 , and @xmath53 otherwise .
hence @xmath802 , summing over those partitions @xmath11 obtained by removing a node from @xmath63 . to apply [ ebullord ] , we need to know the multiplicities @xmath803 .
we have @xmath804 and by [ heckerouqdec](2 ) this equals @xmath15 if @xmath801 and @xmath53 otherwise .
now the result follows from [ ebullord ] .
now we consider applying @xmath780 to spin characters .
[ ebullspin ] suppose @xmath805 and @xmath806 with @xmath807 in @xmath230
. then @xmath808 equals @xmath8 if @xmath809 , and @xmath53 otherwise .
we apply the spin branching rule .
assume first that @xmath796 . in order to have @xmath810 , we need @xmath811 for some @xmath812 .
we have @xmath813 , so every addable node of @xmath814 has spin residue@xmath53 , and every addable node also has a node of spin residue@xmath53 to its immediate right , with the exception of the addable node in the first column .
so the only possible way of adding two nodes of spin residues@xmath53 and then @xmath15 is to add the addable node at the bottom of the first column , and then the addable node at the bottom of the second column , which gives the partition @xmath809 .
the spin branching rule then gives the coefficient @xmath808 as @xmath8 .
the case where @xmath815 is very similar : in this case @xmath816 , so @xmath814 has addable nodes in columns @xmath15 and @xmath8 , and the only way to add a node of spin residue@xmath15 followed by a node of spin residue@xmath53 and end up with a @xmath8-regular partitionis to add the node in column @xmath8 followed by the node in column @xmath15 .
take @xmath579 and @xmath805 , and let @xmath817 .
let @xmath818 , @xmath819 , @xmath820 , @xmath821 be the matrices defined in [ somedecevensec ] for the block with core @xmath236 and weight @xmath779 ( so @xmath820 is the decomposition matrix of @xmath822 , and so on )
. then @xmath823 now comparing [ steinberg , steinbergod ] we see that this equals @xmath824 .
hence @xmath825 , and so @xmath700 . from the results of the last two subsections
we can finally deduce some information about irreducibility of spin characters modulo @xmath8 .
we continue to assume that @xmath230 is a rouquier block with weight @xmath114 , core @xmath236 and bar - core@xmath242 .
we no longer make any assumption on the parity of @xmath114 . in this section
we will classify the spin characters in @xmath230 that remain irreducible in characteristic @xmath8 .
[ rouqirred ] suppose @xmath826 , and let @xmath827 then @xmath216 is irreducible @xmath432 is a @xmath8-carter partition . by [ allodd ] or [ allodd2 ]
, the row of the decomposition matrix corresponding to @xmath11 equals the row of the adjustment matrix of @xmath574 corresponding to @xmath828 or @xmath829 . by [ adj ] or [ adjod ] , the sum of the entries of this row equals the sum of the entries of the row of the decomposition matrix of @xmath830 labelled by @xmath432 .
thus @xmath216 is irreducible the weyl module @xmath831 is irreducible , and by [ irredweyl ] this happens @xmath432 is a @xmath8-carter partition .
we now want to show that if @xmath11 is not of the form @xmath814 or @xmath832 then @xmath216 is reducible .
[ sameregdoub ] suppose @xmath10 lies in @xmath230 , and that @xmath11 has the form @xmath833 for some @xmath834 .
then there is @xmath133 such that @xmath218 and @xmath835 .
we assume first that @xmath431 for @xmath200 .
let @xmath394 be the length of @xmath432 , and let @xmath836 be the partition obtained from @xmath11 by removing the part @xmath837 .
then @xmath836 has @xmath237-bar - weight@xmath838 , so by [ rouqspinform3 ] includes all the integers @xmath839 .
hence @xmath11 includes all these integers too , which is the same as saying that @xmath840 .
so if we let @xmath841 , then the partition @xmath842 equals @xmath843 , i.e. the partition @xmath471 from [ thirdrdim ] .
furthermore , if @xmath844 then @xmath845 , so we can define a partition @xmath846 and we then have @xmath494 for @xmath847 , while the partition @xmath848 coincides with the partition @xmath474 from [ thirdrdim ] . now by combining [ thirdrdim , dimrowrem ] we obtain @xmath218 and @xmath442 . the case where @xmath434 is the same , using [ firstrdim , rouqspinform1 ] instead of [ thirdrdim , rouqspinform3 ] . in the same way ( using [ secondrdim , fourthrdim ] rather than [ firstrdim , thirdrdim ] ) we obtain the following .
[ sameregdoub2 ] suppose @xmath10 lies in @xmath230 , and @xmath11 has the form @xmath849 for some @xmath834 .
then there is @xmath133 such that @xmath218 and @xmath835 .
now we can give the main result for rouquier blocks .
[ mainirrrouq ] suppose @xmath197 with @xmath237-bar - core@xmath242 , and that @xmath10 lies in a rouquier block .
then @xmath216 is irreducible @xmath11 has the form @xmath814 or @xmath832 for @xmath432 a @xmath8-carter partition . by [ spinreg ] , we may assume @xmath11 has at most one non - zero even part
. hence by [ rouqspinform3 ] or [ rouqspinform1 ] we can write @xmath430 or @xmath850 for some natural number @xmath851 . in the case where @xmath430 or @xmath832 , the result follows from [ rouqirred ] .
if @xmath852 with @xmath440 , then by [ sameregdoub ] or [ sameregdoub2 ] we can find another @xmath8-regular partition@xmath63 with the same regularised double as @xmath11 and such that @xmath18 has smaller degree than @xmath10 .
hence by [ samerdoub ] @xmath216 is reducible .
our aim in the remainder of this section is to extend the results of [ irrspinrouq ] beyond rouquier blocks .
so now we drop the assumption that we are working in a rouquier block , but we make a definition motivated by [ rouqspinform3,rouqspinform1 ] : given @xmath197 and @xmath853 or @xmath157 , we say that @xmath11 is _ @xmath25-separated _ if * @xmath11 has at most one non - zero even part , * all the odd parts of @xmath11 are congruent to @xmath25 modulo @xmath237 , and * if @xmath11 includes an even part @xmath854 , then @xmath11 includes all positive integers less than @xmath854 which are congruent to @xmath25 modulo @xmath237 .
say that @xmath11 is _ separated _ if it is either @xmath15- or @xmath157-separated .
if we let @xmath242 be the @xmath237-bar - coreof @xmath11 , then @xmath11 is separatedif @xmath11 has one of the following two forms : * @xmath814 , where @xmath552 with @xmath855 ; * @xmath850 , where @xmath856 , @xmath552 and @xmath857 is at most the number of parts of @xmath242 which are greater than @xmath854 .
so let s write @xmath858 and @xmath859 when @xmath242 is a @xmath237-bar - coreand @xmath432 a partition .
the idea of the definition of separated@xmath8-regular partitionsis that the corresponding characters @xmath10 are well - behaved , inheriting properties of corresponding spin characters in rouquier blocks .
the terminology ` separated ' is intended to reference the term ` @xmath8-quotient separated ' introduced by james and mathas @xcite for a class of partitions labelling characters which behave in a similar way for symmetric groups and hecke algebras .
our aim is to prove the following , which generalises [ rouqirred ] .
[ voddirr ] suppose @xmath11 is a separated@xmath8-regular partition , and write @xmath860 with @xmath861
. then @xmath216 is irreducible @xmath862 and @xmath432 is a @xmath8-carter partition .
to prove [ voddirr ] , we use downwards induction on the @xmath237-bar - weightof @xmath11 .
the inductive step is achieved via the following s. [ voindstep ] given @xmath435 , let @xmath434 and let @xmath863 .
suppose @xmath860 is separated , and let @xmath864 . 1 . if @xmath865 , then @xmath866 2 .
if @xmath867 , then @xmath868 hence @xmath216 is irreducible @xmath446 is . 1 .
@xmath11 has two @xmath15-spin - addable nodesin each of rows @xmath869 .
so by the induction version of [ spinbranchpower ] , @xmath870 . similarly , since @xmath63 has two @xmath15-spin - removable nodesin each row , @xmath871 .
2 . we let @xmath394 be such that @xmath872 , and consider two cases according to the parity of @xmath851 .
if @xmath851 is even , then @xmath11 has two @xmath15-spin - addable nodesin each of rows @xmath873 except row @xmath394 . adding these nodes yields @xmath63 .
@xmath63 has no other @xmath15-spin - removable nodes , so removing all @xmath15-spin - removable nodesfrom @xmath63 leaves @xmath11 .
now the expressions for @xmath874 and @xmath875 follow from [ spinbranchpower ] and its induction inversion .
+ if @xmath851 is odd , then each of rows @xmath873 contains two @xmath15-spin - addable nodesof @xmath11 except rows @xmath394 and @xmath876 , which each contain one ( note that the separatedcondition means that @xmath877 , so that @xmath878 is not a spin - addable nodeof @xmath11 ) .
adding all these nodes yields @xmath63 , and @xmath63 has no other @xmath15-spin - removable nodes .
again , the expressions for @xmath874 and @xmath875 follow from [ spinbranchpower ] and its induction inversion . in either case , we obtain @xmath879 so the final statement follows from [ modbranch ] . in a similar way we obtain the following . [ voindstep2 ] given @xmath200 ,
let @xmath431 and @xmath880 .
suppose @xmath860 is separated , and let @xmath864 . 1 .
if @xmath865 , then @xmath881 2 . if @xmath867 , then @xmath882 hence @xmath216 is irreducible @xmath446 is .
we use downwards induction on the @xmath237-bar - weightof @xmath11 . note that by [ mainirrrouq ] the result holds if @xmath10 lies in a rouquier block . in general
if we assume that @xmath11 is separatedand write @xmath860 , then for any @xmath237-bar - core@xmath558 larger than @xmath242 , the partition @xmath883 is also separatedand has the same @xmath237-bar - weightas @xmath11 , i.e. @xmath884 .
so if we fix @xmath432 and @xmath851 and allow @xmath242 to vary , then for sufficiently large @xmath242 the character @xmath885 lies in a rouquier block .
so to prove [ voddirr ] for a specific @xmath860 , it suffices to show that @xmath216 is irreducible @xmath886 is irreducible , for some @xmath237-bar - core@xmath558 larger than @xmath242 . if @xmath11 is @xmath15-separatedthis follows from [ voindstep ] , while if @xmath11 is @xmath157-separatedthen we use [ voindstep2 ] instead .
now we come to the conclusion of this section , which is that [ main ] is true for spin characters labelled by separatedpartitions .
looking at the six cases listed in [ main ] , we see that @xmath11 is separatedin ( and only in ) cases 14 : in cases 1 and 2 we have @xmath887 with @xmath432 a @xmath8-carter partition , and cases 3 and 4 we have @xmath888 with @xmath432 a @xmath8-carter partition .
so when @xmath11 is separated , [ main ] is equivalent to [ voddirr ] .
now we prove [ main ] in the case where @xmath11 is not separated .
to do this , we introduce some notation for the various classes of partitions appearing in [ main ] : * let @xmath889 denote the set of @xmath8-regular partitions@xmath814 , where @xmath431 with @xmath200 and @xmath432 a @xmath8-carter partition with @xmath433 ; * let @xmath890 denote the set of @xmath8-regular partitions@xmath814 , where @xmath434 with @xmath435 and @xmath432 a @xmath8-carter partition with @xmath433 ; * let @xmath891 denote the set of @xmath8-regular partitions@xmath832 , where @xmath431 with @xmath200 and @xmath432 a @xmath8-carter partition with @xmath433 ; * let @xmath892 denote the set of @xmath8-regular partitions@xmath832 , where @xmath434 with @xmath435 and @xmath432 a @xmath8-carter partition with @xmath437 ; * let @xmath893 denote the set of @xmath8-regular partitions@xmath438 or @xmath439 for @xmath440 .
now let @xmath894 . to prove the `` if '' part of [ main ] in the case where @xmath11 is not separated , we need the following theorem of wales . waltheorem 7.7[wales ] suppose @xmath1 is a prime .
then the @xmath1-modular reduction of @xmath895 is irreducible unless @xmath9 is odd and @xmath1 divides @xmath9 . in particular , we have @xmath896 irreducible for every @xmath851 .
next consider the case @xmath897 and let @xmath898 . then
( as we have just noted ) @xmath446 is irreducible , and by examining spin residueswe see that @xmath899 , so by [ modbranch ] @xmath216 is irreducible too . the final case to consider to complete the `` if '' part of the proof is @xmath441 , for which one can just check the readily - available decomposition matrix ( for example , at the modular atlas homepage www.math.rwth-aachen.de/homes/moc/ ) .
now we proceed with the `` only if '' part of [ main ] , which is more complex .
we will prove by induction on @xmath48 that if @xmath11 is a @xmath8-regular partitionnot in @xmath900 then @xmath216 is a reducible @xmath8-modular character .
the key result which underlies the combinatorics of the proof is the following ( recall that we write @xmath78 for the partition obtained by removing all the @xmath25-spin - removable nodesfrom @xmath197 ) . [ rsprop ] suppose @xmath197 and @xmath77 .
if @xmath216 is irreducible , then so is @xmath901 .
if @xmath216 is irreducible , then by [ modbranch ] so is @xmath902 .
but by [ spinmaxbranch ] , @xmath902 equals @xmath903 for some @xmath83 , so @xmath901 must be irreducible ( and in fact @xmath100 must equal @xmath15 ) . so using our inductive hypothesis we may assume that for @xmath904 either @xmath905 or @xmath906 .
we also assume that @xmath11 is not separated , and by [ spinreg ] we can assume that @xmath11 has at most one non - zero even part .
we summarise our assumptions for ease of reference . *
assumptions and notation in force for the rest of [ mainproofsec ] : * @xmath11 is a @xmath8-regular partitionwhich has at most one non - zero even part , is not separatedand does not lie in @xmath900 . for @xmath904 , either @xmath905 or @xmath906 .
we let @xmath907 .
recall that by [ spinreg ] , if @xmath216 is irreducible , then @xmath908 .
we will often exploit this by finding an induction or restriction functor ( or a composition of such functors ) which kills @xmath16 but not @xmath216 , so that @xmath16 and @xmath216 can not be equal .
the assumptions above have several immediate consequences .
* @xmath909 can not lie in @xmath889 or @xmath891 , since partitions in @xmath910 all have @xmath15-spin - removable nodes(with the exception of @xmath54 , but in this case @xmath11 would also equal @xmath54 , contrary to assumption ) . * similarly , @xmath911 can not lie in @xmath890 or @xmath892 . *
neither @xmath911 nor @xmath909 can lie in @xmath893 , since then @xmath11 would lie in @xmath893 as well . * neither @xmath911 nor @xmath909 can equal @xmath246 .
so our assumptions give @xmath912 and @xmath913 .
we now consider several cases .
0 1 in this case , write @xmath916 with @xmath294 a @xmath8-carter partition of length @xmath917 .
we can obtain @xmath11 from @xmath909 by adding @xmath15-spin - addable nodes , and there are two of these nodes which can be added in each row from @xmath15 to @xmath201 .
we must add at least one node in each of these rows , since otherwise @xmath11 would have a @xmath53-spin - removable node , so that @xmath918 . on the other hand , @xmath11 has at most one even part , so there can be at most one row where we add exactly one node .
if we add two nodes in each row , then @xmath919 contrary to assumption , so instead there must be some @xmath394 such that we add one node in row @xmath394 and two nodes in each other row .
that is , @xmath920
hence @xmath921 we now consider several subcases ; recall that @xmath198 is the length of @xmath294 . 1 .
if @xmath922 , then @xmath11 is separated , contrary to assumption .
2 . if @xmath923 , then @xmath924 , again contrary to assumption .
[ subef ] if @xmath925 , then @xmath926 we apply the modular branching rules to compute @xmath927 . note that @xmath63 has an addable @xmath53-node at the end of each row from @xmath15 to @xmath928 , and ( since @xmath929 ) a removable @xmath53-node at the end of row @xmath930 and an addable @xmath53-node at the start of row @xmath931 . hence all the addable @xmath53-nodes except the one in row @xmath928 are conormal , so @xmath932 , where @xmath933 applying the modular branching rules again we obtain @xmath934 . on the other hand ,
the spin branching rules give @xmath935 , so we can not possibly have @xmath908 .
4 . if @xmath936 , then @xmath937 ( since @xmath294 is a @xmath8-carter partition and hence @xmath8-regular ) , which means that @xmath938 this partition has an addable @xmath15-node in row @xmath939 and removable @xmath15-nodes in all other rows from @xmath15 to @xmath930 .
but the fact that @xmath940 means that one of these removable nodes ( namely , the one in row @xmath941 ) is not normal , so @xmath942 .
but the spin branching rules give @xmath943 , so @xmath944 .
take @xmath945 .
then @xmath914 and @xmath946 so that @xmath947 .
we have @xmath948 giving @xmath949 now we examine the residues of the nodes of @xmath63 and the spin - residues of the nodes of @xmath11 : @xmath950 the branching rules give @xmath951 and @xmath952 . and
since @xmath953 we deduce that @xmath944 , so that @xmath216 is reducible . 1 in this case , write @xmath955 with @xmath294 a @xmath8-carter partition of length @xmath917 .
@xmath11 is obtained from @xmath909 by adding @xmath15-spin - addable nodes , and sufficiently many such nodes must be added that @xmath11 has no @xmath53-spin - removable nodes .
in particular , at least one node must be added in each of rows @xmath869 , and exactly one node must be added in each of rows @xmath956 and @xmath957 . but
this means in particular that @xmath958 . since by assumption @xmath11
has at most one even part , all of @xmath959 must be odd , and therefore we must have @xmath960 contrary to assumption .
1 now write @xmath963 with @xmath294 a @xmath8-carter partition of length @xmath917 .
@xmath11 is obtained from @xmath911 by adding nodes of spin residue@xmath53 , and at least one such node must be added in each of rows @xmath869 to ensure that @xmath962 .
@xmath11 can have at most one even part , so there can be at most one of these rows where we add a single @xmath53 .
we then also have the possibility of adding a @xmath53 in row @xmath956 . if we add two nodes to each of rows @xmath869 , then ( regardless of whether we also add a node in row @xmath956 ) @xmath11 is separated , contrary to assumption .
so there is @xmath964 such that we add a single node in row @xmath394 and two nodes in all other rows from @xmath15 to @xmath201 .
again we consider several subcases .
first suppose @xmath11 does have a node in row @xmath956 ; that is , @xmath965 then @xmath966 1 . [ subc ]
if @xmath967 , then @xmath968 and the modular branching rules give @xmath969 .
but @xmath970 , so @xmath944 .
2 . if @xmath922 , then @xmath11 is separated , contrary to assumption .
now suppose @xmath11 does not have a node in row @xmath956 , i.e. @xmath971 1 . if @xmath972 , then ( similarly to subcase [ subc ] ) @xmath973 , while @xmath974 , so @xmath944 .
2 . if @xmath975 , then @xmath924 , contrary to assumption .
3 . if @xmath976 , then ( similarly to subcase [ subef ] ) @xmath977 kills @xmath16 but not @xmath216 so we can not have @xmath908 .
4 . finally suppose @xmath978 , and define @xmath979 then by applying [ dimrowrem ] @xmath980 times followed by [ dimen1 ] , we obtain @xmath981 and @xmath982 , so that by [ samerdoub ] @xmath216 is reducible .
1 now write @xmath984 with @xmath294 a @xmath8-carter partition .
we obtain @xmath11 from @xmath911 by adding sufficiently many @xmath53-spin - addable nodesto eliminate any @xmath15-spin - removable nodes .
we certainly have to add the node @xmath985 ( to eliminate the spin - removable node@xmath986 ) .
we also have to add at least one node in each of rows @xmath987 ; since @xmath11 has at most one even part , we must add two nodes in each of these rows . if in addition we add two nodes in row @xmath201 , then @xmath988 , contrary to assumption .
so we must add no nodes in row @xmath201 , and this means that @xmath989 , since otherwise @xmath11 has a @xmath15-spin - removable node .
so we have @xmath990 in particular , this implies @xmath991 , since by assumption @xmath992 .
so @xmath993 now the modular branching rules ( together with the fact that @xmath991 ) give @xmath994 , while @xmath995 . 1 in this case , let @xmath851 denote the total number of spin - removable nodesof @xmath11 .
we claim first that the length of @xmath63 is at most @xmath851 .
this follows by considering the various possibilities for @xmath911 and @xmath909 .
note that for each @xmath25 we have @xmath998 ; since @xmath11 can have at most two spin - removable nodesin row @xmath25 , this means that @xmath999 and @xmath1000 differ by at most @xmath8 .
moreover , @xmath1001 is either @xmath53 or @xmath15 .
this leaves six distinct possibilities .
we tabulate these below , giving the number @xmath851 for each , together with the length of @xmath211 .
@xmath1002 now note that in the third and fourth cases the last two parts of @xmath211 are both equal to @xmath15 , so that the length of @xmath63 is less than the length of @xmath211 , and hence at most @xmath851 . in the fifth and sixth cases , the last three parts of @xmath211 are all equal to @xmath15 ,
so the length of @xmath63 is at most @xmath1003 , and hence less than or equal to @xmath851 .
so our claim is proved .
now if @xmath63 has fewer than @xmath851 normal nodes , then for some @xmath77 and some @xmath62 we have @xmath1004 while @xmath1005 , so that @xmath944 .
so we can assume that @xmath63 has @xmath851 normal nodes , and since @xmath63 has length at most @xmath851 this means that @xmath63 must have a normal node at the end of every non - empty row .
but the only way this can happen is if the nodes at the ends of the non - empty rows of @xmath63 all have the same residue ( since if the nodes at the ends of rows @xmath394 and @xmath876 have different residues , then the one at the end of row @xmath876 is not normal ) .
but if the removable nodes of @xmath63 all have residue @xmath25 , then @xmath1006 , while @xmath1007 ( since by assumption @xmath1008 ) .
take @xmath1009 .
then @xmath1010 so we are in the final case in the table above , with @xmath1011 .
@xmath11 has seven spin - removable nodes .
we have @xmath1012 so that @xmath63 can have at most seven normal nodes .
but in fact @xmath63 has only three normal nodes ( all of residue @xmath53 ) .
hence @xmath1013 , so that @xmath216 is reducible .
this completes the proof of [ main ] .
let @xmath1014 denote the alternating group , and define define @xmath1015 to be the inverse image of @xmath1014 under the natural homomorphism @xmath1016 . then @xmath1015 is a double cover of @xmath1014 , and ( except when @xmath42 or @xmath1017 ) is the unique schur cover of @xmath1014 .
we can ask our main question for characters of @xmath1015 as well , and the answer is closely related to that for @xmath4 . in this section
we make a few observations about this problem , which we hope to solve in a future paper . throughout this section
we work over fields large enough to be splitting fields for @xmath1014 . as with @xmath4 ,
the case of linear characters follows from the corresponding results for @xmath1014 , so here we need only be concerned with spin characters of @xmath1015 .
we say that two characters of @xmath1015 are _ conjugate _ if we can transform one into the other by conjugating all elements of @xmath1015 by an odd element of @xmath4 .
the ordinary irreducible spin characters of @xmath1015 are easily constructed from those for @xmath4 : for every @xmath148 the restriction @xmath1018 is a sum of two conjugate irreducible characters @xmath12 , while for @xmath149 the restriction @xmath1019 is a self - conjugate irreducible character which we denote @xmath10 . the set @xmath1020 is a complete set ( without repeats ) of ordinary irreducible spin characters of @xmath1015 . to help us to determine which of these characters remain irreducible in characteristic @xmath8 , we consider restricting from @xmath4 , using the fact that restriction to subgroups commutes with modular reduction .
_ we now no longer allow ourselves to write @xmath10 to mean `` either @xmath13 or @xmath14 '' when @xmath149 .
_ so @xmath10 for @xmath149 will unambiguously mean the spin character for @xmath1015 labelled by @xmath11 . as with @xmath4 , there are no irreducible spin characters of @xmath1015 in characteristic @xmath8 , so the irreducible @xmath8-modular characters of @xmath1015 are the same as those for @xmath1014 .
these are also constructed by restriction from @xmath3 , but to explain the situation here we need a definition : say that @xmath197 is an _ s - partition _ if for every odd @xmath201 the @xmath201th ladder of @xmath11 contains an even number of nodes .
then @xmath1021 is a sum of two conjugate irreducible characters @xmath1022 if @xmath11 is an s - partition , and otherwise @xmath1021 is a self - conjugate irreducible character .
the irreducible characters arising in this way are ( without repeats ) all the irreducible @xmath8-modular characters of @xmath1014 .
this result is due to benson ( * ? ? ?
* theorem 1.1 ) , though the as far as we are aware the characterisation of s - partitions we have given is new .
now we consider what happens when we take an ordinary irreducible spin character of @xmath4 , restrict to @xmath1015 and reduce modulo @xmath8 .
take @xmath149 .
then the fact that restriction commutes with modular reduction gives @xmath1023 so in order for @xmath216 to be irreducible , @xmath1024 must be irreducible , and must remain irreducible on restriction to @xmath1015 . by [ spinreg ] , this means that @xmath1025 , and that @xmath213 is not an s - partition .
in fact , the second condition is automatic : it is an easy exercise that an s - partition must have even @xmath8-weight , whereas all the partitions in [ main ] that lie in @xmath1026 have odd @xmath237-bar - weight .
so for @xmath1027 we conclude that @xmath216 is irreducible @xmath1024 is , i.e. @xmath11 is one of the partitions appearing in [ main ] .
now consider @xmath148 . in this case
restricting and reducing modulo @xmath8 gives @xmath1028 since @xmath13 and @xmath14 are conjugate characters , @xmath1024 is irreducible @xmath1029 is , and this happens @xmath1030 has exactly two irreducible constituents . hence either * @xmath216 is irreducible , or * @xmath216 has exactly two irreducible constituents , which restrict to conjugate irreducible @xmath8-modular characters of @xmath1014 .
the main theorem of the present paper tells us exactly when the first situation occurs . for the second situation to occur
, the two irreducible constituents of @xmath216 would have to be equal in order for their restrictions to @xmath1015 to be conjugate and irreducible , so by [ spinreg ] we would have to have @xmath1031 . in particular , this implies that @xmath146 would have to be @xmath8 .
so to solve the main problem for spin characters of @xmath1015 , it remains to classify partitions @xmath197 with exactly two non - zero even parts , such that @xmath1031 .
such partitions do occur ; the first example is @xmath1032 , and we will see a family of examples below .
one of the main results in this paper ( [ allodd , allodd2 ] ) is the calculation of part of the decomposition matrix for a rouquier block of @xmath4 ; specifically , the rows corresponding to @xmath8-regular partitionswith no even parts greater than @xmath8
. it would be interesting to extend this to determine the whole of the decomposition matrix for a rouquier block .
we expect this would be amenable to the same techniques , once suitable combinatorial expressions for the entries are found . by way of example
, we consider the spin part of the decomposition matrix for a rouquier block of weight @xmath1033 .
let @xmath236 be a @xmath8-core of length at least @xmath1034 and @xmath242 the corresponding @xmath237-bar - core , and let @xmath230 be the block with core @xmath236 and weight @xmath1033 .
then we write @xmath694 for the spin part of the decomposition matrix of @xmath230 ; we label the columns of @xmath694 by partitions of @xmath1033 , and the rows by pairs @xmath1035 of partitions with @xmath532 @xmath8-regular and @xmath1036 , and set @xmath1037 we also define @xmath140 to be the adjustment matrix for @xmath1038 .
the matrices @xmath694 and @xmath717 are given in [ eav ] .
@xmath1039 @xmath1040 note that the submatrix consisting of the first three rows of @xmath717 is the matrix @xmath740 from [ somedecoddsec ] , in agreement with [ allodd2 ] .
we leave the reader to try to work out the pattern in the rest of this matrix .
we can not even see why the entries of @xmath717 are necessarily non - negative !
another point to observe in the matrix @xmath694 is the sixth row , which shows a spin character whose @xmath8-modular reduction just has two equal composition factors .
this is relevant to the discussion of the double cover of the alternating group in [ ansec ] : it shows that if @xmath11 is one of @xmath1041 then the ordinary spin characters of @xmath1015 labelled by @xmath11 remain irreducible in characteristic @xmath8 .
for the reader s convenience we conclude with an index of the notation we use in this paper .
we provide references to the relevant subsections . | for any finite group @xmath0 and any prime @xmath1 one can ask which ordinary irreducible representations remain irreducible in characteristic @xmath1 .
we answer this question for @xmath2 when @xmath0 is a proper double cover of the symmetric group .
our techniques involve constructing part of the decomposition matrix for a rouquier block of a double cover , restricting to subgroups using the brundan kleshchev modular branching rules and comparing the dimensions of irreducible representations via the bar - length formula .
= 3mu plus 1mu minus 1mu 0 |
it has been suggested that primordial magnetic fields can develop large correlation lengths provided they carry nonvanishing `` magnetic helicity '' @xmath2 , a quantity known to particle physicists as the abelian , euclidean chern - simons term . here
@xmath3 is an abelian gauge potential for the magnetic field @xmath4 . if there exists a period of decaying turbulence in the early universe , which can occur after a first - order phase transition , a magnetic field with nonvanishing helicity could have relaxed to a large - scale configuration , which enjoys force - free dynamics ( source currents for the magnetic fields proportional to the fields themselves ) thereby avoiding dissipation@xcite . in this paper
we accomplish two things .
first we show that configurations of @xmath5 , with quantized helicity , arise from vacuum configurations of a non - abelian @xmath0 vector potential .
the quantization occurs because @xmath5 wind and intertwine in an intricate manner .
we relate this `` winding number '' of our abelian fields to the topological properties of the structures in a non - abelian @xmath0 gauge group .
our construction is based on @xmath6 , which is relevant to a non - abelian gauge theory in 3-space .
[ the construction can also be described in terms of hopf maps and @xmath7 , which have already appeared in the literature@xcite ; we show how our expressions appear in that formalism . ]
only integer winding numbers occur in the mathematical setting of @xmath6 or @xmath7 ; yet we find that half - integer windings also lead to interesting abelian gauge fields .
second , we study the time evolution of the magnetic helicity . since @xmath8 as a consequence of the definition for the electric field , @xmath9
, time variation is determined by the specific physical situation that fixes @xmath10 . for plasma physics or magnetohydrodynamics the projection of @xmath9 on @xmath11
is proportional to the resistivity of the medium , and vanishes for infinite conductivity ( zero resistivity ) , see sec . [
sec : conclusion ] . in that approximation
, helicity is conserved .
however , cosmological electromagnetic fields can be expected also to experience evolution in vacuum , where @xmath12 .
therefore , it is interesting to determine what happens to the magnetic helicity under maxwell evolution .
specifically , we posit that at @xmath13 there are magnetic fields with quantized helicity , and no electric fields . with this initial data
, we solve the maxwell equations and find that at @xmath14 the helicity is precisely half its initial value .
the calculation is carried out explicitly for two interesting cases ( integer and half - integer quanta ) and then a general argument is given , which requires a regularity hypothesis . in section [ sec : helicity ] , we describe how to construct from non - abelian vacuum fields abelian gauge fields with quantized helicity . in section [ sec : time ] , we study time evolution . concluding remarks comprise section [ sec : conclusion ] , where we also speculate on the applicability of these mathematical considerations to the @xmath15 standard model " , before and after its electroweak phase transition .
the chern - simons number of a non - abelian gauge potential @xmath16 is given by @xmath17 & = & -\frac{1}{8 \pi^2 } \int { { \rm d\null}}^3 r \ , \epsilon^{ijk } { \mathop{\rm tr\null}}(a_i \partial_j a_k + { { \textstyle\frac{1}{3 } } } a_i [ a_j , a_k ] ) \nonumber \\[1ex ] & = & -\frac{1}{8 \pi^2 } \int { \mathop{\rm tr\null}}(ada + { { \textstyle\frac{2}{3 } } } a^3 ) \label{eq:2.1}\end{aligned}\ ] ] in the second equality , the @xmath18 are elements of the lie algebra with generators @xmath19 @xmath20=f_{abc}t^c \ , \quad { \mathop{\rm tr\null}}t^at^b = -\delta_{ab}/2\ ] ] while form notation is used in the third equality : @xmath21 . the normalization factor @xmath22 is chosen for later convenience , and is also maintained in the abelian limit , where only the bilinear part of ( [ eq:2.1 ] ) survives .
thus our chern - simons quantity is @xmath23 times the magnetic helicity , which entered physics in the work of woltier@xcite and was subsequently further elaborated by many people , including moffatt , berger and field , as well as arnold and khesin@xcite .
particle physicists have considered both abelian and non - abelian chern - simons terms on 3-dimensional minkowski space in studies of @xmath24-dimensional gauge theories@xcite . to show how abelian fields with nonvanishing and quantized helicity arise from non - abelian vacuum configurations , we recall first that under a gauge transformation a a^u u^-1 au + u^-1 du [ eq:2.2 ] where @xmath1 is an element of the gauge group , the chern - simons term transforms as cs(a^u)= cs(a ) + d(duu^-1a ) + ( u^-1 du)^3 [ eq:2.3 ] with sufficiently regular @xmath25 and @xmath1 , the second term on the right side integrates to zero@xcite
. however , the last term , also a total derivative although that is not apparent from its formula ( but see below ) depends only on the group element , and is not damped by any fall - off of @xmath25 . indeed its value is the winding number of @xmath1 , which effects a mapping from 3-space into the group .
w(u ) = dr ( u^-1 du)^3 [ eq:2.4 ] @xmath26 is an integer when @xmath1 is nonsingular for finite @xmath27 and tends to @xmath28 at infinity , for then 3-space can be compactified to the 3-sphere , and we assume that the gauge group contains @xmath0 .
thus we are relying on @xmath29 integers .
it follows therefore that the chern - simons number of a non - abelian vacuum gauge potential , that is , one which is a pure gauge = u^-1 du [ eq:2.5 ] is the winding number of @xmath1 .
let us now specialize to @xmath0 .
then @xmath1 involves the pauli matrices @xmath30 .
@xmath31 & = & \cos f/2 - i \sigma^a { \hat{\omega}}^a \sin f/2 \label{eq:2.6}\end{aligned}\ ] ] and the winding number is explicitly seen to involve a total derivative@xcite w(u ) = - d ( _ abc ^a d^b d^c ( f - f ) ) [ eq:2.7 ] where @xmath32 is a unit vector in isospace : @xmath33 .
consider next the abelian gauge potential 1-form @xmath34 , constructed from a non - abelian vacuum configuration @xmath35 by projecting the latter on a fixed direction in isospace , specified by a constant unit vector @xmath36 a = i ^a ^a u^-1 du [ eq:2.8a ] in components , this reads a_i=
^a ^a_i [ eq:2.8b ] [ eq:2.8 ] where @xmath37 is the vacuum ( pure gauge ) non - abelian potential u^-1 _ i u = ^a_i [ eq:2.9 ] which satisfies ^ijk _ j
^a_k = -^ijk _ abc ^b_j ^c_k [ eq:2.10 ] and carries the winding number w(u ) = - d^3 r ^ijk _ abc ^a_i ^b_j ^c_k [ eq:2.11 ] note that @xmath3 is * not * an abelian pure gauge .
we now show that the magnetic helicity of @xmath3 coincides with @xmath26 .
the magnetic helicity , with our normalization , is given by [ eq:2.12 ] @xmath38 & = & \frac{1}{16 \pi^2 } { \hat{n}}^a { \hat{n}}^b \int { { \rm d\null}}^3 r \ , \epsilon^{ijk } { \mbox{$\cal a\;$}}^a_i \partial_j { \mbox{$\cal a\;$}}^b_k \nonumber \\[1ex ] & = & -\frac{1}{32 \pi^2 } { \hat{n}}^a { \hat{n}}^b \int { { \rm d\null}}^3 r \ , { \mbox{$\cal a\;$}}^a_i \epsilon^{ijk } \epsilon_{bcd } { \mbox{$\cal a\;$}}^c_j { \mbox{$\cal a\;$}}^d_k \label{eq:2.12a}\end{aligned}\ ] ] where ( [ eq:2.8 ] ) and ( [ eq:2.10 ] ) have been used .
but @xmath39 thus h(a ) = - d^3 r ^ijk _ abc ^a_i ^b_j ^c_k = w(u ) [ eq:2.12b ] and
we conclude that the magnetic helicity is quantized by the winding number of the non - abelian vacuum configuration .
the form of the abelian potential @xmath3 may be given explicitly .
parameterizing the unit vector @xmath32 as ^a = ( , , ) [ eq:2.13 ] and taking @xmath36 to point in the third ( electromagnetic ) direction , we find a= df - ( f ) d- ( 1-f ) ^2 d[eq:2.14a ] an alternate formula presents ( [ eq:2.14a ] ) in `` clebsch '' form , which will be useful later .
( the clebsch form for an arbitrary 3-vector @xmath40 is @xmath41 where @xmath42 are three scalar functions . )
a = d(-2 ) + 2(1-(^2 f/2 ) ^2 ) d(+^-1[(f/2 ) ] ) [ eq:2.14b ] [ eq:2.14 ] the magnetic field is determined by the 2-form obtained from ( [ eq:2.14a ] ) [ eq:2.15 ] da=[eq:2.15a ] or from the clebsch expression ( [ eq:2.14b ] ) da=-2 d ( ( ^2 f/2 ) ^2 ) d ( + ^-1 [ ( f/2 ) ] ) [ eq:2.15b ] finally the magnetic helicity becomes , according to ( [ eq:2.14a ] ) and ( [ eq:2.15a ] ) [ eq:2.16 ] h(a ) = - ( 1-f ) df d d[eq:2.16a ] or from ( [ eq:2.14b ] ) and ( [ eq:2.15b ] ) h(a)= d d ( ( ^2 f/2 ) ^2 ) d(^-1 [ ( f/2 ) ] ) [ eq:2.16b ] in the clebsch parameterization , the magnetic helicity is seen to involve integration of a total derivative and is therefore given by a surface integral .
the explicit clebsch expressions for @xmath34 and @xmath43 demonstrate that one may find two magnetic surfaces , @xmath44 @xmath45 , which satisfy @xmath46 @xmath47 s_2 & = & \phi+ \tan^{-1}\big [ ( \tan f/2 ) \cos
\theta \big]=\phi_0 \nonumber \\[1ex ] & & c , \phi_0~{\rm constants } \label{eq:2.17}\end{aligned}\ ] ] the intersection of these surfaces forms magnetic lines , that is , integral curves of @xmath11 that solve the dynamical system = ( ( ) ) [ eq:2.18 ] where @xmath48 is an evolution parameter for the dynamical system .
evidently , for our configuration this problem in integrable , leading to curves given by @xmath49 \sin \theta & = & \frac{c}{\sqrt{\sin^2(\phi-\phi_0 ) + c^2 \cos^2 ( \phi-\phi_0 ) } } \label{eq:2.19}\end{aligned}\ ] ] it is known that the abelian chern - simons term can be related to the degree of the hopf map , which is quantized according to @xmath7@xcite .
the construction of the relevant vector potential proceeds in the following manner . consider a complex spinor @xmath50 with @xmath51 because @xmath52 is divergence - free , it can be written as the curl of a potential @xmath53 given by @xmath54 . by comparison with ( [ eq:2.14b ] )
we find that our potential arises when @xmath55 is chosen as u = ( c e^2i^-1[(f/2 ) ( ) ] + ( f/2 ) e^-2i ) [ eq:2.21 ] and the hopf index coincides with the helicity of the here - constructed @xmath5 . evidently , the present expressions are awkward when compared to those based on @xmath6 , presumably because the latter construction is directly related to gauge fields
. henceforth we work with explicit expressions obtained by identifying @xmath32 in ( [ eq:2.6 ] ) with the radial unit vector @xmath56 and taking @xmath57 to depend only on @xmath58 .
this further means that @xmath59 in @xmath60 in ( [ eq:2.13 ] ) are identified with polar and azimuthal angles of spherical coordinates @xmath61 and @xmath62 , with ranges @xmath63 and @xmath64 respectively . ( a simple generalization would be to identify @xmath59 and @xmath60 with integer multiples of @xmath61 and @xmath62 . )
the magnetic helicity ( [ eq:2.16 ] ) then becomes h=- ^_0 dr ( f - f ) = - ( f - f ) |_r= [ eq:2.22 ] where we have taken @xmath65 to vanish . observe that when @xmath66 is nonvanishing @xmath67 is an irrational / transcendental number .
when @xmath68 is an even integer multiple of @xmath69 , @xmath66 vanishes and @xmath67 is an integer . but note that also an odd integer multiple of @xmath69 for @xmath68 leads to vanishing @xmath66 , and a half - integer value for @xmath67 .
therefore it appears to us that configurations with half - integer magnetic helicity are also singled out .
the half - integer fields share further good properties with the integer - valued configurations : we shall show below that only for these two the winding number does not change under the gauge transformation to the coulomb gauge . when reference is made back to the non - abelian vacuum configuration , which determines @xmath3 , we see from ( [ eq:2.6 ] ) that for the integer windings @xmath70 , as expected .
for the half - integer ones , a `` hedgehog '' asymptote is attained : @xmath71 .
the vector potential @xmath3 and the magnetic field @xmath72 are neatly described by spherical components , which are @xmath62-independent [ eq:2.23 ] @xmath73 a_\theta & = & -(\sin \theta ) \frac{1}{r } \sin f \label{eq:2.23b } \\[1ex ] a_\phi & = & -(\sin \theta ) \frac{1}{r } ( 1-\cos f ) \label{eq:2.23c } \end{aligned}\ ] ] [ eq:2.24 ] @xmath74 b_\theta & = & ( \sin \theta ) \frac{f'}{r } \sin f \label{eq:2.24b } \\[1ex ] b_\phi & = & ( \sin \theta ) \frac{f'}{r } ( 1-\cos f ) \label{eq:2.24c } \end{aligned}\ ] ] ( the dash denotes differentiation with respect to @xmath58 . )
the clebsch representation ( [ eq:2.14b ] ) , ( [ eq:2.15b ] ) and ( [ eq:2.16b ] ) gives a clear picture of the helicity . from ( [ eq:2.16b ] )
we have [ eq : new24 ] h = - d^3 r _
i b^i = - d^3 r ( ) b_[eq : new24a ] since @xmath75 is @xmath62-independent , the @xmath62 integration is trivial , leaving h = - ^_0 d^_0 r dr b_|_=2 [ eq : new24b ] the integral is over the positive-@xmath76 @xmath77 half - plane and @xmath75 is the toroidal magnetic field , perpendicular to that plane .
so @xmath67 measures the flux of the toroidal magnetic field through a half - plane .
note that the functions occurring in our clebsch parameterization are multivalued , owing to the presence of the naked azimuthal angle @xmath62 and the @xmath78 .
this is as it must be , because the helicity is nonvanishing@xcite . indeed
the multivaluedness of @xmath62 is responsible for the nonzero value of the @xmath62 `` surface '' integral ( [ eq : new24 ] ) , which reproduces the helicity .
when using @xmath3 in a description of electromagnetic fields , we are effectively in the @xmath79 , weyl gauge . but also the vector potential must be transverse since there are no sources .
however , the potential in ( [ eq:2.23 ] ) is not transverse for general @xmath57 .
transversality may be achieved in one of two ways .
we may perform a gauge transformation , transforming @xmath3 to a transverse expression @xmath80 . alternatively
, we may choose the function @xmath57 so that @xmath3 is transverse .
we discuss these two possibilities in turn .
the transversality condition affects only the poloidal components of @xmath3 ( @xmath81 and @xmath82 ) , because the toroidal component @xmath83 is @xmath62-independent . by using a gauge function proportional to @xmath84 times a function of @xmath58
, one can choose that function so that the resulting , gauge equivalent potential is transverse .
the components of the transverse potential then read @xmath85 a^t_\theta & = & -(\sin \theta ) \frac{1}{2r } \big(r^2 f\big ) ' \nonumber \\[1ex ] a^t_\phi & = & -(\sin \theta ) \frac{1}{r } \big(1-\cos f\big ) \nonumber \\[1ex ] f(r ) & \equiv & { { \textstyle\frac{2}{3 } } } \int^\infty_r \frac{{{\rm d\null}}r'}{r^{\prime 2 } } ( f-\sin f ) - \frac{4}{3r^3 } \int^r_0 { { \rm d\null}}r ' \ , r ' ( f-\sin f ) \label{eq:2.25}\end{aligned}\ ] ] of course the magnetic field remains unchanged , but we must still check the helicity integral : while the integrand is gauge dependent , the integral changes only be a surface term . it may be that the above gauge transformation contributes from the surface .
indeed this happens for general @xmath57 : one finds h(a^t ) = h(a ) + ( 1-f ) f|_r= [ eq:2.26 ] and the helicity of the gauge - equivalent , transverse configuration differs from the original expression .
however if , and only if , @xmath57 goes at infinite @xmath58 to an even or odd multiple of @xmath69 , the gauge variance vanishes .
thus both integer and half - integer windings are stable against this gauge transformation , which renders the potential transverse , but for other , irrational windings , the `` winding number '' loses its meaning because it is gauge dependent@xcite . the other way to achieve transversality is to impose that condition on the original configuration ( [ eq:2.23 ] ) , thereby determining @xmath57 .
transversality requires that @xmath57 satisfy r^2 f
+ 2rf - 2f=0 [ eq:2.27 ] although this equation can not be solved by elementary functions , it is easily analyzed by analogy to a mechanical problem where `` time '' is @xmath86 .
one finds two solutions that are regular at the origin , vanishing linearly with @xmath58 , and tending to @xmath87 in an oscillatory manner for large @xmath58 .
moreover , the scale of @xmath58 is arbitrary : the solutions are a universal function of @xmath88 and its negative ; the positive solution is plotted in fig .
[ fig:1 ] .
evidently this transverse potential necessarily corresponds to half - integer winding@xcite .
according to maxwell s vacuum equations the transverse vector potential satisfies the wave equation , which we integrate subject to the initial condition that @xmath3 is given at time @xmath13 by ( [ eq:2.25 ] ) , with a definite choice for @xmath57 , and that @xmath90 is zero .
thus at initial time , there is no electric field and only the magnetic field ( [ eq:2.24 ] ) is present , maintained for @xmath91 by a steady current @xmath92 . @xmath93 j_\theta & = & -(\sin \theta ) \frac{1}{r } ( f-\sin f ) ''
\nonumber \\[1ex ] j_\phi & = & ( \sin \theta ) \bigl ( \frac{1}{r^2 } [ r ( 1- \cos f ) ] ' \bigr ) ' \label{eq:3.1}\end{aligned}\ ] ] and carrying energy @xmath94 @xmath95\nonumber\\ & = & \frac{8\pi}3 \int_0^\pi \frac{{{\rm d\null}}r}{r^2 } ( 1-\cos f)(1-\cos f + r^2 f^{\prime 2 } ) \label{3.2.new}\end{aligned}\ ] ] we determine the time asymptote for the helicity in two cases [ eq:3.2 ] @xmath96 f & = & 4 \tan^{-1 } r / r_0 \label{eq:3.2b}\end{aligned}\ ] ] the former , with @xmath97 , corresponds to half - integer winding ; for the latter @xmath98 , one has integer winding , also the energy density is spherically symmetric ; see ( [ 3.2.new])@xcite .
the corresponding magnetic fields are [ eq:3.3 ] @xmath99 b_\theta & = & 4(\sin \theta ) \frac{r_0 ^ 2}{(r^2+r_0 ^ 2)^2 } \nonumber \\[1ex ] b_\phi & = & 4(\sin \theta ) \frac{rr_0}{(r^2+r_0 ^ 2)^2 } \label{eq:3.3a}\\[2ex ]
\hbox{integer winding } \qquad b_r & = & -16(\cos \theta ) \frac{r_0 ^ 2}{(r^2+r_0 ^ 2)^2 } \nonumber \\[1ex ] b_\theta & = & 16(\sin \theta ) \frac{r_0 ^ 2(r_0 ^ 2-r^2)}{(r^2+r_0 ^ 2)^3 } \nonumber \\[1ex ] b_\phi & = & 32(\sin \theta ) \frac{rr_0 ^ 3}{(r^2+r_0 ^ 2)^3 } \label{eq:3.3b}\end{aligned}\ ] ] the time - evolved fields are obtained by standard fourier transform techniques , and the helicity integral is evaluated as function of time .
we find for the two cases @xmath100 h & = & -1 + \half \frac{t^2}{r_0 ^ 2+t^2 } \bigl ( 1 + \frac{r_0 ^ 2}{r_0 ^ 2+t^2 } + \frac{8r_0 ^ 6}{(r_0 ^ 2+t^2)^3 } \bigr ) \label{eq:3.4b}\end{aligned}\ ] ] [ eq:3.4 ] it is seen that for @xmath101 , @xmath67 attains one - half its value at @xmath13 . while one would like to have an understanding how the localization of the helicity changes with time , it does not seem possible to pose such a question in a meaningful way , because the helicity density , namely , the integrand that defines @xmath67 , is gauge dependent , and without invariant meaning
indeed , as we have noted earlier , even the integrated helicity can be gauge dependent when the fields and the gauge function survive on the surfaces bounding the integration region .
( this points to an analogy with the energy in general relativity , whose density is diffeomorphism - dependent .
only the integrated quantity is invariant and a unique value is determined only after asymptotic conditions are prescribed .
furthermore , the energy may be presented as a surface integral and so also can the helicity , when the clebsch parameterization is used for the gauge potential . ) the result that under maxwell evolution @xmath67 decreases to half its value at infinite time can also be understood from a general argument . in the weyl - coulomb gauge , which we are using ,
the vector potential satisfies the wave equation , which is uniquely solved with our posited initial conditions by ^t ( t , ) = kt e^-i ^t ( ) [ eq:3.5 ] here @xmath102 is fourier transform of the initial data ( [ eq:2.25 ] ) and @xmath103 vanishes at @xmath13 .
it follows that @xmath104 & = & \frac1{2i } \int \frac{{{\rm d\null}}^3 k}{(2 \pi)^3 } ( 1 + \cos 2 kt ) \epsilon^{ijk } k^i \tilde{a}_j^t ( { \bbox{\rm k } } ) \tilde{a}^{t^*}_k({\bbox{\rm k } } ) \label{eq:3.67}\end{aligned}\ ] ] by appealing to the riemann - lebesgue lemma , we can argue that the term involving the cosine disappears at large @xmath105 , owing to destructive interference , leaving half the value at @xmath13 . however , this step is justified provided the rest of integrand is well behaved at @xmath106 , which in turn depends on the behavior of @xmath107 at large @xmath58 .
a dimensional estimate shows that a large - distance decrease of @xmath80 faster than @xmath108 is sufficient , which would mean that @xmath11 should decrease faster than @xmath109 .
but ( [ eq:3.3a ] ) exhibits a large @xmath58 behavior for @xmath11 of order @xmath109 , modulated by an angular factor .
since our explicit calculation supports the general argument , the angular factor evidently provides sufficient large-@xmath58 damping .
our investigation is based on the connection between the winding number of non - abelian gauge group elements and the chern - simons number of abelian gauge fields that are obtained from the former by projection .
this mathematical fact suggests a physical scenario for the @xmath110 standard model " . before its ( first - order ) phase transition
, its vacuum could be populated by pure gauge , vacuum configurations @xmath111 of the @xmath0 gauge group , which carry nonvanishing winding numbers .
after the phase transition , one direction in isospace is identified with electromagnetism , and the projection of the vacuum configuration becomes a magnetic field with nonvanishing helicity .
it remains to be shown whether the above scenario is energetically stable and can be justified on physical grounds .
another problem deserving further study concerns nontrivial evolution of an initial configuration with magnetic helicity . rather than using the free maxwell equations
, one would rely on the magnetohydrodynamical ones , which make use of ohm s law to express @xmath9 in terms of @xmath11 and @xmath112 .
its nonrelativistic form is = ( + ) [ eq:4.1 ] where @xmath113 is conductivity and @xmath114 is the fluid velocity , taken to be divergenceless in a fluid of constant density .
inserting this in the maxwell equation + = 0 [ eq:4.2 ] produces an evolution equation for @xmath115 - ( ) = -1 .
[ eq:4.3 ] further , approximating @xmath112 by @xmath116 , that is , ignoring @xmath117 because it is negligible on the relevant time scales , converts the above into - ( ) = 1 ^ 2 [ eq:4.4 ] which can be analyzed either with @xmath114 prescribed externally or determined self - consistently by its euler equation , with initial @xmath11 of the form ( [ eq:2.24 ] ) . an interesting choice for an external @xmath114 could be a transverse form that carries nonvanishing kinetic helicity " @xmath118 ; for example , what one gets by taking @xmath114 in the form @xmath3 of ( [ eq:2.23 ] ) with @xmath57 solving ( [ eq:2.27 ] ) . for zero resistivity ( infinite conductivity ) the right side of ( [ eq:4.4 ] )
is absent there is no dissipation .
then @xmath119 and magnetic helicity is conserved , since @xmath120 vanishes .
cornwall , phys .
d * 56 * 6146(1997 ) ; s. carroll and g. field , astro - ph/9811206 ; d.t .
son , phys . rev .
d * 59 * , 063008(1999 ) . for background , see d. biskamp , _ nonlinear magnetohydrodynamics _ ( cambridge university press , cambridge , uk 1993 ) .
h. moffatt , j. fluid .
mech . * 35 * , 117(1969 ) , ( with a. tsinober ) ann .
fluid mech .
* 24 * , 282(1992 ) ; m. berger and g. field , j. fluid mech .
* 147 * , 133(1984 ) ; v. arnold and b. khesin , _ topological methods in hydrodynamics _
( springer , berlin , 1998 ) .
there are physically interesting situations where this term survives ; see r. jackiw and s .- y .
pi , phys .
b423 * , 364(1998 ) ; s. deser , l. griguolo , and d. seminara , comm .
* 197 * , 443(1998 ) .
h. moffatt , j. fluid . mech . * 35 * , 117(1969 ) ; g. field , in _ magnetospheric phenomena in astrophysics _ ,
r. epstein and w. feldman , eds .
, aip conference proceedings * 144 * , 324 ( american institute of physics , new york , 1986 ) . the response to more general gauge transformations is as follows . choosing the gauge function @xmath121 [ so that the form ( [ eq:2.23 ] ) is preserved ] @xmath122 change into @xmath123 , @xmath124 .
consequently the winding number becomes @xmath125 .
configurations with integer winding remain gauge invariant [ provided @xmath126 does not increase too rapidly at infinity ] . for all other configurations ,
@xmath126 must vanish at infinity to maintain gauge invariance of @xmath67 .
moreover , the gauge function @xmath121 is `` well - defined '' at infinity only for vanishing @xmath127 .
( we thank a. polychronakos for discussion . ) this magnetic field with integer winding number has also been constructed by c. adam , b. muratori , and c. nash , phys .
d * 60 * , 125001(1999 ) and m. giovannini , hep - th/9906241 .
half - integer winding numbers are studied also in adam et al . in the work cited as well as in k. haller , l. chen , and y.s .
choi , phys .
d * 60 * , 125010(1999 ) . | projecting a non - abelian @xmath0 vacuum gauge field a pure gauge constructed from the group element @xmath1 onto a fixed ( electromagnetic ) direction in isospace gives rise to a nontrivial magnetic field , with nonvanishing magnetic helicity , which coincides with the winding number of @xmath1 .
although the helicity is not conserved under maxwell ( vacuum ) evolution , it retains one - half its initial value at infinite time .
amssym.def amssym.tex boxedeps = msbm10 at 12pt = msbm7 at 9pt = msbm5 at 6pt = = = # 1#1 # 1 # 1 = 14pt = 20pt = 12pt = 6.25 in = 6.25 in = 0.25 in = -.5 in |
the choice of the coordinates used to describe a molecule is an important issue if computational considerations are to be taken into account and the efficiency of the simulations is pursued .
this choice also affects the coding of applications .
if cumbersomely defined coordinates are used , an unnecessary complexity may be added to the design of monte carlo movements , the construction and pruning of a database of structures @xcite or the programming of molecular visualization and manipulation tools .
suitable coordinates frequently used to describe arbitrary conformations of molecules are the so - called `` internal '' or `` valence - type '' coordinates @xcite .
their adequacy stems from a number of characteristics : first , they are closely related to chemically meaningful structural parameters , such as bond lengths or bond angles ; second , they are local , in the sense that each one of them involves only a small number of ( normally close ) atoms in its definition ; and finally , there are only @xmath2 of them ( where @xmath3 is the number of atoms in the molecule ) , in such a way that the overall rotation and translation have been naturally removed
. there also exists a family of coordinates @xcite , extensively used in the inner calculations of many quantum chemistry packages ( such as gaussian @xcite or gamess @xcite ) and based on the `` natural internal coordinates '' originally proposed by pulay and coworkers @xcite , which are defined through linear combinations of the original internals .
these coordinates are specially designed to describe normal - mode vibrations in the immediate neighbourhood of energy minima and represent the best choice for accelerating convergence of geometry optimizations in a particular basin of attraction , via diagonal estimation of the hessian matrix @xcite .
accordingly , they maximally separate hard and soft movements in these conditions .
however , if the conformation of the molecule is far from a minimum , this type of coordinates lose great part of their meaning and they introduce many computational difficulties without increasing the efficiency . also , some of the definitions are _ redundant _
@xcite , i.e. , they use a number of coordinates larger than the number of degrees of freedom . in this work , we will only discuss coordinates , such as internals or cartesian , that may be conveniently used to specify an _ arbitrary _ conformation of the system and that can be directly related to simple geometrical variables . in macromolecules , such as proteins , the number of degrees of freedom is the main limiting factor when one tries to predict their behaviour via computer modeling .
therefore , it is also advisable that the set of coordinates chosen allows for a direct implementation of physically meaningful constraints that reduce the dimensionality of the conformational space considered .
most of the expressions used in statistical mechanics or in molecular dynamics are best written in cartesian coordinates , however , the implementation of constraints naturally appearing is far from being straightforward in these coordinates . in internal coordinates , on the contrary , the approximate separation of hard and soft movements of the system allows to easily constrain the molecule @xcite by setting the hard coordinates ( those that require a considerable amount of energy to change noticeably ) to constant values or to particular functions of the soft coordinates .
moreover , in internal coordinates ( and appealing to some reasonable approximations ) , the statistical mechanics formulae for the constrained system may be written in convenient closed form @xcite . still , although the bond lengths and bond angles are customarily regarded as hard and their definition is unproblematic , the same is not true for dihedral angles .
some definitions of dihedrals may lead to difficulties or to worse separation of hard and soft modes .
let us exemplify this with a particular case : consider the definition of z - matrix - like @xcite internal coordinates for the hco - l - ala - nh@xmath1 molecule in fig .
[ fig : num_ala ] .
imagine that we `` position '' ( i.e. , we write the corresponding z - matrix row ) every atom up to the hydrogen denoted by h@xmath4 and that we are now prepared to position the hydrogens in the side chain ( h@xmath5 , h@xmath6 and h@xmath7 ) via one bond length , one bond angle and one dihedral for each one of themwe will denote by @xmath8 the bond length between atoms @xmath9 and @xmath10 ; by @xmath11 , the bond angle between the vectors @xmath12 and @xmath13 ; and by @xmath14 the dihedral angle between the plane defined by the atoms @xmath9 , @xmath10 and @xmath15 and the one defined by @xmath10 , @xmath15 and @xmath16 . ] .
a choice frequently seen in the literature @xcite is the one shown in table [ tab : bad_coord ] .
cccc atom name & bond length & bond angle & dihedral angle + + h@xmath5 & ( 10,8 ) & ( 10,8,5 ) & @xmath17(10,8,5,3 ) + h@xmath6 & ( 11,8 ) & ( 11,8,5 ) & @xmath18(11,8,5,3 ) + h@xmath7 & ( 12,8 ) & ( 12,8,5 ) & @xmath19(12,8,5,3 ) if we now perform the _ gedanken experiment _ that consists of taking a typical conformation of the molecule and slightly moving each internal coordinate at a time while keeping the rest constant , we find that any one of the three dihedrals in the previous definition is a hard coordinate , since moving one of them while keeping the other two constant distorts the internal structure of the methyl group . hence , in these coordinates , the soft rotameric degree of freedom @xmath0 , which we know , for chemical arguments , that must existaccording to our calculations , at the rhf/6 - 31+g(d ) level of the theory , the barrier for crossing from one of the three equivalent minima to any of the other two ranges from 3.1 to 6.8 kcal / mol , depending on the values of the ramachandran angles @xmath20 and @xmath21 . compare with the barriers in @xmath20 or @xmath21 which may be as large as 20 kcal / mol depending on the region of the ramachandran map explored .
] , is ill - represented .
in fact , it must be described as a _ concerted _ movement of the three dihedrals . in references
@xcite , this fact is recognized and the concept of `` related dihedrals '' is introduced , however , no action is taken to change the definition of the coordinates . in this work , using the ideas of r. abagyan and coworkers @xcite , we define a set of rules to uniquely and systematically number the groups , the atoms and define the internal coordinates of organic molecules and , particularly , of polypeptidesiupac conventions only define a numeration system for the groups , for the branches and for some selected dihedral angles .
they focus on functional considerations and not in computational problems . for related documents and references ,
see http://www.chem.qmul.ac.uk/ iupac / jcbn/. ] .
the main difference with other z - matrix - like coordinates normally used in the literature @xcite is that , instead of positioning each atom with a bond length , a bond angle and a dihedral angle , we use normal dihedral angles ( called , from now on , `` principal dihedrals '' ) only to fix the orientation of whole groups and a somewhat non - standard type of dihedrals , termed `` phase dihedrals '' by r. abagyan and coworkers @xcite ( see fig . [
fig : dihedrals ] ) , to describe the covalent structure inside a groupanother option may be to use , as a third internal coordinate for each atom , another bond angle .
this is rather awkward , however , since two bond angles and a bond length do not specify the position of a point in space .
any values of these three coordinates ( except for irrelevant degenerate cases ) are compatible with two different symmetrical positions and a fourth number must be provided to break the ambiguity . ] .
this allows to approximately separate soft and hard movements of the molecule using only topological information ( i.e. , not knowing the exact form of the potential ) and to easily implement constraints by forcing the coordinates that correspond to hard movements to take constant values or ones that depend on the soft coordinatesin reference @xcite they correctly take this approach into account using out - of - plane angles instead of phase dihedrals , however , they do not describe any rules for a general definition and their numeration of the atoms is non - modular , as it proceeds first through the backbone ( see sec . [ sec : definition_pro ] ) . ] .
in addition , the coordinates herein defined , are straightforwardly cast into z - matrix form and may be directly implemented in any quantum chemistry package , such as gaussian @xcite or gamess @xcite .
this is due to the fact that , although they involve atoms whose covalent structure is different , the mathematical construction of the two types of angles in fig .
[ fig : dihedrals ] is exactly the same , and the phase dihedrals are treated like principal ones without any problem by the applications .
a number of perl scripts are provided that number the atoms and generate the coordinates herein defined for polypeptide chains .
the applications read a sequence file in which the different ionization states of the titratable side chains , the tautomeric forms of histidine and several terminal groups may be specified .
then , an output file is generated with the symbolic definition of the z - matrix of the molecule which may be directly pasted into the input files of gaussian @xcite or gamess @xcite ( and , upon slight modifications , of any quantum chemistry package that is capable of reading z - matrix format ) .
these scripts may be found at http://neptuno.unizar.es / files / public / gen_sasmic/. now , if we redo the example in table [ tab : bad_coord ] using phase dihedrals , we must write the rows of the z - matrix for the hydrogens in the side chain as shown in table [ tab : good_coord ] .
cccc atom name & bond length & bond angle & dihedral angle + + h@xmath5 & ( 10,8 ) & ( 10,8,5 ) & @xmath22*(10,8,5,3 ) * + h@xmath6 & ( 11,8 ) & ( 11,8,5 ) & @xmath23(11,8,5 , ) + h@xmath7 & ( 12,8 ) & ( 12,8,5 ) & @xmath24(12,8,5 , ) where the angle * ( 10,8,5,3 ) * is now the principal dihedral @xmath0 describing the relative rotation of the methyl group around the bond length ( 8,5 ) and the other two are phase dihedrals that describe the internal structure of the group and that are _ pure _ hard coordinates ( as far as can be told only from topological information ) .
however , one must point out that , although all bond lengths , bond angles and phase dihedrals may be regarded as hard coordinates , not all the principal dihedrals will be soft .
examples of hard principal dihedrals are the ones that describe the rotation around a double bond ( or a triple one ) or some of the principal dihedrals in cyclic parts of molecules .
the _ physical approach _ described in this section , which should be taken into account when designing internal coordinates , is embodied in a set of rules for any organic molecule in sec .
[ sec : definition_gen ] , with a slightly different prescription for polypeptide chains in sec .
[ sec : definition_pro ] .
the systematic numeration introduced facilitates the computational treatment of this type of systems and the rules given for polypeptide chains ensure modularity @xcite , i.e. , allows to add any residue with minimal modification of the already existing notation and to easily construct databases of structures or of potential energy surfaces ( pes ) . the characteristics aforementioned have led us to term the coordinates herein defined _ systematic , approximately separable and modular internal coordinates _ ( sasmic ) . in this work , we will only deal with the numeration of one isolated molecule , however , the procedure described may be easily generalized ( and will be in future works ) to systems of many molecules ( an important example being a macromolecular solute in a bath of solvent molecules ) . this could be done using _
ghost atoms _ in a similar manner to what is done in ref .
@xcite , to position the center of mass of the system , and in refs .
@xcite , to actually define the coordinates of a system of molecules . finally , in sec .
[ sec : application ] , we use the new coordinates and ab initio quantum mechanical calculations in order to evaluate the approximation of the free energy , obtained from `` integrating out '' the rotameric degree of freedom @xmath0 , via the typical pes in the protected dipeptide hco - l - ala - nh@xmath1 .
this will be relevant to design effective polypeptide potentials .
we also present a small part of the hessian matrix in two different sets of coordinates to illustrate the approximate separation of soft and hard movements when the sasmic defined in this work are used .
[ sec : conclusions ] is devoted to the conclusions .
first , we realize that any molecule may be formally divided in groups such as those in fig . [
fig : groups ] . we will call `` centers '' the shaded atoms in the figure and `` vertices '' the white ones .
in general , there may exist groups with more than four vertices , however , in proteins , only groups with four or less vertices
occura fact that will not be used in the definitions . ] .
examples of tetrahedral groups are the one whose center is the c@xmath25 in the backbone or the c@xmath26 in the side chain of alanine , triangular groups occur , for example , at the n or the c in the backbone , finally , linear groups may be found at the o in the side chain of tyrosine or at the s in methionine ( see fig . [ fig:20dipeptides ] )
. a particular atom may be vertex of different groups but may only be center of one group .
there exist atoms that are only vertices but there do not exist atoms that are only centers , except in the case of molecules with only one group . in the trivial case of diatomic molecules (
in which the only internal coordinate is a bond length ) , neither of the previous definitions are possible , since we can not identify a group .
atoms that are covalently bonded to more than one atom will be called `` internal atoms '' and are indicated as shaded circles in fig .
[ fig : int - ext_his ] . atoms that are covalently attached to only one internal atom will be called `` external atoms '' and are indicated as white - filled circles in fig .
[ fig : int - ext_his ] . in proteins ,
only h and o may be external atoms . in most macromolecular models ( such as the born - oppenheimer approximation used in sec .
[ sec : application ] ) , nuclei are considered point - like particles . hence , rotation around bonds joining external and internal atoms ( termed `` external bonds '' or `` non - dihedral bonds '' ) is neglected , i.e. , there are no internal coordinates associated to this movement . on the other hand , rotation around bonds joining two internal atoms ( called `` internal bonds '' or `` dihedral bonds '' and indicated with curved arrows in fig .
[ fig : int - ext_his ] ) is relevant and there may exist internal coordinates describing it . in order to conform with the _ physical approach _
stated in the introduction , only one _ golden rule _ must be followed when defining the internal coordinates : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ one principal dihedral , at most , must be defined on each internal bond . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the rest of the rules that will be given are mere tidy conventions and systematics . first of all , we will divide the molecule in groups and number them . to do this
we proceed _ by branches _ , i.e. , we choose the next group following a linear sequence of covalently attached groups until there is no possible next one , in which case , we either have finished the numeration process or we start another branch .
every group is numbered one time and it can not be renumbered as the process continues .
this numeration is done for completeness and as a support for the numeration of atoms and coordinates . in fig .
[ fig : groups_his]a , we have implemented these _ general _ rules in a protected histidine dipeptide .
later , in sec . [ sec : definition_pro ] , while stating the rules for polypeptides , we will remark the differences ( which will lead to fig .
[ fig : groups_his]b ) and show the reasons for the special prescriptions using the same example .
the rules are as follows : 1 .
the first group ( @xmath27 ) , is chosen , among those that are linked to the molecule via only one internal bond ( termed `` terminal groups '' ) , as _ the one that has the greater mass_the mass of a group is defined as the sum of the atomic masses of its constituents . ] . if two or more terminal groups have the same mass , we add the mass of their first neighbours to break the tie .
if this does not lead to a decision , we proceed to the second neighbours and so on .
if we run out of neighbours and there is still a tie , we choose a group arbitrarily among the ones that have been selected via this process and we indicate the convention .
if there are no terminal groups , we perform this selection process among those groups that have _ at least one external atom_the rare case in which there are neither terminal groups nor external atoms ( such as c@xmath28 fullerene ) will not be treated here , although it would require only a small number of adjustments to the rules . ] .
if there is only one unnumbered group linked to group @xmath10 , we number it as @xmath29 , set @xmath30 and go to ( ii ) .
if there are two or more unnumbered groups linked to group @xmath10 , we choose _ the one with the greater mass _ as in point ( i ) , we number it as @xmath29 , set @xmath30 and go to ( ii ) .
4 . if there are no unnumbered groups linked to group @xmath10 but there are still unnumbered groups in the molecule , we set @xmath10 to _ the number of the lowest numbered group that has unnumbered neighbours _ ( we prepare to start another branch ) and we go to ( ii ) .
this process terminates when all the groups are numbered . the atoms will be numbered in the order that they will be positioned via internal coordinates in the z - matrix . as in the previous subsection , in fig .
[ fig : num_his ] , these general rules , as well as the special rules for polypeptides , are exemplified in a protected histidine dipeptide . in sec .
[ sec : definition_pro ] , we will remark the differences and show the advantages of slightly modifying the prescription .
the rules are as follows : 1 . the first atom ( @xmath31 ) , is chosen as _ the heaviest of the external atoms in the first group_.
if there are two or more candidates with the same mass , we choose arbitrarily and indicate the convention .
the second atom ( @xmath32 ) is the center of the first group and we set @xmath27 ( the index of the group ) .
if group @xmath29 exists and is covalently attached to group @xmath10 , we number the unnumbered vertices of group @xmath10 _ starting by the center of group @xmath29 and , then , in order of decreasing mass_. if , otherwise , group @xmath29 does not exist or it is not covalently attached to group @xmath10 , _ we simply number the unnumbered vertices of group @xmath10 in order of decreasing mass_. if , at any point , there are two or more candidates with the same mass , we choose arbitrarily and indicate the convention .
exception : if groups @xmath10 and @xmath29 belong to the same cyclic part of the molecule , the vertices of @xmath10 that are centers of groups ( other than @xmath29 ) belonging to the same cycle _ must not be numbered at this step _
( for an example of this rule , see the numeration of c@xmath33 and n@xmath34 in fig .
[ fig : num_his]a , or c@xmath35 and n@xmath36 in fig .
[ fig : num_his]b ) .
4 . if group @xmath29 does not exist , we have finished .
otherwise , we set @xmath30 and go back to ( iii ) . using the numeration for the atoms given in the previous section , we give now a set of rules for defining the internal coordinates that conform with the _ physical approach _ discussed in the introduction of this work .
the coordinates are written in z - matrix form ( see table [ tab : coor_his_gen ] ) for convenience and the rules are applied to the protected dipeptide hco - l - his - nh@xmath37 ( with the side chain in its uncharged @xmath38 tautomeric form ) using the general numeration given in fig .
[ fig : num_his]a . cccc atom name & bond length & bond angle & dihedral angle + + o@xmath39 & & & + c@xmath1 & ( 2,1 ) & & + n@xmath40 & ( 3,2 ) & ( 3,2,1 ) & + h@xmath41 & ( 4,2 ) & ( 4,2,3 ) & ( 4,2,3,1 ) + c@xmath42 & ( 5,3 ) & ( 5,3,2 ) & * ( 5,3,2,1 ) * + h@xmath43 & ( 6,3 ) & ( 6,3,2 ) & ( 6,3,2,5 ) + c@xmath44 & ( 7,5 ) & ( 7,5,3 ) & * ( 7,5,3,2 ) * + c@xmath45 & ( 8,5 ) & ( 8,5,3 ) & ( 8,5,3,7 ) + h@xmath46 & ( 9,5 ) & ( 9,5,3 ) & ( 9,5,3,7 )
+ n@xmath5 & ( 10,7 ) & ( 10,7,5 ) & * ( 10,7,5,3 ) * + o@xmath6 & ( 11,7 ) & ( 11,7,5 ) & ( 11,7,5,10 ) + h@xmath7 & ( 12,10 ) & ( 12,10,7 ) & * ( 12,10,7,5 ) * + h@xmath35 & ( 13,10 ) & ( 13,10,7 ) & ( 13,10,7,12 ) + c@xmath47 & ( 14,8 ) & ( 14,8,5 ) & * ( 14,8,5,3 ) * + h@xmath48 & ( 15,8 ) & ( 15,8,5 ) & ( 15,8,5,14 ) + h@xmath49 & ( 16,8 ) & ( 16,8,5 ) & ( 16,8,5,14 ) + c@xmath33 & ( 17,14 ) & ( 17,14,8 ) & * ( 17,14,8,5 ) * + n@xmath36 & ( 18,17 ) & ( 18,17,14 ) & * ( 18,17,14,8 ) * + h@xmath50 & ( 19,17 ) & ( 19,17,14 ) & ( 19,17,14,18 ) + c@xmath51 & ( 20,18 ) & ( 20,18,17 ) & * ( 20,18,17,14 ) * + h@xmath52 & ( 21,18 ) & ( 21,18,17 ) & ( 21,18,17,20 ) + n@xmath34 & ( 22,20 ) & ( 22,20,18 ) & * ( 22,20,18,17 ) * + h@xmath53 & ( 23,20 ) & ( 23,20,18 ) & ( 23,20,18,22 ) the rules are as follows : 1 .
the positioning of _ the first three atoms _ is special .
the corresponding rows of the z - matrix are _ always _ as the ones in table [ tab : coor_his_gen ] ( except , of course , for the chemical symbol in the first column , which may change ) .
the positioning of the remaining vertices of group number 1 ( if there is any ) is also special , their rows in the z - matrix are : + t@xmath54 @xmath55 @xmath56 @xmath57 + where t is the chemical symbol of the @xmath9-th atom , and @xmath57 is a phase dihedral .
3 . we set @xmath9 to _ the number that follows that of the last vertex of the first group_. 4 .
we choose @xmath10 as _ the lowest numbered atom that is covalently linked to _ @xmath9 .
we choose @xmath15 as _ the lowest numbered atom that is covalently linked to _ @xmath10 .
if no principal dihedral has been defined on the bond @xmath58we say that a principal dihedral @xmath14 _ is _
`` on the bond @xmath58 '' .
] , we choose @xmath16 as _ the lowest numbered atom that is covalently linked to _ @xmath15 .
otherwise , we choose @xmath16 as _ the second lowest numbered atom that is covalently linked to _
@xmath10 ( i.e. , the lowest numbered atom that is covalently linked to @xmath10 and that is different from @xmath15 , or , equivalently , the atom that was used to define the only principal dihedral on the bond @xmath58 ) .
the row of the z - matrix that corresponds to atom @xmath9 is : + t@xmath54 @xmath8 @xmath11 @xmath14 + where t is the chemical symbol of atom @xmath9 , @xmath8 is a bond length , @xmath11 is a bond angle and @xmath14 is a principal dihedral if the first case in point ( vi ) has occurred or a phase dihedral otherwise .
if @xmath59 does not exist , we have finished .
otherwise , we set @xmath60 and go to ( iv ) .
the _ numeration of the groups _ in a polypeptide chain is the same as in sec .
[ subsec : number_groups_gen ] except for some details : * instead of using rule ( i ) for choosing the first group , we selectthese three cases are the most frequent .
if a different species is used to n - protect the polypeptide chain , a convention must be sought that also starts at the n - terminus . ] : * * the _ amino group _ at the n - terminus ( either charged or not ) if the polypeptide is not n - protected . * * the _ formyl group _ at the n - terminus if the polypeptide is formyl - n - protected . * * the _ methyl group _ at the n - terminus if the polypeptide is acetyl - n - protected .
+ this is done because the primary structure of a polypeptide is normally presented from the n- to the c - terminus . in the case of hco - l - his - nh@xmath37 in fig .
[ fig : groups_his ] , the agreement between rule ( i ) and this one is accidental .
* when we must choose the next group to the one whose center is a c@xmath25 in the backbone , instead of applying rule ( iii ) , which would yield the group at the c as the next one ( compare fig .
[ fig : groups_his]a and [ fig : groups_his]b ) , we choose _ the first group in the side chain _ ( for residues that are different from glycine ) .
then , we resume the numeration as usual .
this is done in order to ensure modularity , since , if we numbered following rule ( iii ) , the backbone would be always numbered first and the whole numeration would have to be modified if we added a new residue to the chain . * also for modularity reasons , we want to completely number the side chain before proceeding into the backbone .
hence , if we are numbering side chain atoms and the requirements to apply rule ( iv ) are fulfilled , instead of applying this rule , we set @xmath10 to _ the number of the lowest numbered group that has unnumbered neighbours and that belongs to the side chain of the residue whose groups we are numbering .
_ then , we go to ( ii ) as usual . the _ numeration of the atoms _ in a polypeptide chain is the same as in sec .
[ subsec : number_atoms_gen ] except for some details : * we seek that the principal dihedrals that are to be defined after numbering the atoms conform to the biochemical iupac conventions for the dihedrals @xmath20 , @xmath21 and @xmath61 in the backbone . at the termini
, we want that the atom where the c@xmath25 of the hypothetical residue @xmath62 or @xmath63 would occur is used to define the principal dihedrals .
the general rules ensure this , except in two cases where a special convention must be given : * * if the polypeptide is formyl - n - protected , instead of applying rule ( i ) to choose the first atom , which would give the oxygen at the formyl group , we choose _ the hydrogen at the formyl group _
( compare fig .
[ fig : num_his]a and [ fig : num_his]b ) .
then , we resume the numeration at rule ( ii ) . *
* if the polypeptide is amide - c - protected , instead of applying rule ( iii ) and arbitrarily choosing one of the hydrogens in the terminal amide group before the other , we number _ the trans hydrogen _ before the other ( compare fig . [ fig : num_his]a and [ fig : num_his]b ) . * due to the special rules for the numeration of groups given above , the next group to the one at the c@xmath25 is the first one in the side chain .
if we applied rule ( iii ) for numbering the vertices of the c@xmath25-group , we would number first the center of the first group at the side chain and , then , the c in the backbone .
this would make the only principal dihedral defined on bond ( c@xmath25,n ) different from the conventional ramachandran angle @xmath20 . in order to avoid this , instead of applying rule ( iii ) at this point , we number the c first among the unnumbered vertices of the c@xmath25-group and , then , resume the usual numeration process .
see fig .
[ fig:20dipeptides ] for the numeration of the twenty naturally ocurring amino acids with formyl - n- and amide - c - protection .
once the groups and the atoms have been numbered following these special prescriptions , the _ definition of the internal coordinates _ for polypeptide chains is identical to the one described for the general case in sec .
[ subsec : number_coor_gen ] .
when a number of degrees of freedom are removed from the description of the conformations of a physical system via their integrating out in the partition function , the energy function that remains , which describes the behaviour of the system only in terms of the rest of the degrees of freedom , is a _
free energy_. it depends on the temperature and contains the entropy of the information that has been averaged out as well as the enthalpy . however , it is frequent , when studying the conformational preferences of model dipeptides in order to use the information for designing effective potentials of polypeptides @xcite , that the energy of these molecules be approximated by the potential energy surface ( pes ) in the bidimensional space spanned by the ramachandran angles @xmath20 and @xmath21 @xcite .
if we recognize that the potential energy of the system in the born - oppenheimer approximation ( denoted by @xmath64 ) depends on the @xmath2 internal coordinates , this surface ( denoted by @xmath65 ) may be defined as : @xmath66 where @xmath67 denotes the rest of the internal coordinates .
the use of this surface , instead of a free energy function with the @xmath67 degrees of freedom integrated out , is justified in the approximation that these internal coordinates are _ hard _ and that they are comparably much more difficult to excite at room temperature than @xmath20 and @xmath21 .
if we assume that this is correct , these hard degrees of freedom may be easily eliminated @xcite and the partition function of the system may be written as follows : @xmath68 where @xmath69 .
note however that , in the `` flexible '' picture for the constraints , this expression is correct only if we assume that the jacobian determinant of the change of coordinates from cartesians to @xmath70 and the determinant of the potential second derivatives matrix with respect to the hard coordinates , both evaluated at the equilibrium values , do not depend on @xmath20 and @xmath21 ( see ref .
if , alternatively , we accept the `` rigid '' picture for the constraints , we must ask that the determinant of the induced metric tensor in the constrained sub - manifold do not depend on @xmath20 and @xmath21 @xcite . if these approximations ( which will be reexamined in future works ) do not hold but the hardness of the @xmath67 degrees of freedom is still assumed , the expressions in eq .
[ eq : z_app ] must be modified by adding some correction terms to @xmath71 . in eq .
[ eq : z_app ] for the partition function , one also may see that , apart from the different multiplicative constants @xmath72 and @xmath73 , which do not affect the expected values of observables , the use of the pes @xmath71 as the fundamental energy function of the system is justified because it plays the same role as the whole potential energy of the system in the first integral . however
, although the hardness of the bond lengths , the bond angles and even the dihedral @xmath61 in the peptide bond may be assumed , this is not a good approximation for the rotameric degrees of freedom in the side chains of residues . in the frequently studied @xcite example of hco - l - ala - nh@xmath37 ( see fig . [
fig : num_ala ] ) , as it has already been said in footnote [ foot : rotation_barrier ] , the side chain degree of freedom @xmath0 must be regarded as _
soft_. still , although it is more complex , a soft degree of freedom may also be averaged out if it is considered convenient . in this section
, we will assume that the energy of the formyl - alanine - amide dipeptide may be correctly approximated by a potential energy hypersurface ( peh ) ( denoted by @xmath74 ) that depends on the ramachandran angles @xmath20 and @xmath21 but also on the principal dihedral @xmath0 that describes the rotation of the methyl group in the side chain .
analogously to eq .
[ eq : pes_def ] , its definition in terms of the whole energy of the system is : @xmath75 where @xmath76 represents the internal coordinates that are not @xmath20 , @xmath21 or @xmath0 . note , in addition , that the two definitions are related by the following expression : @xmath77 we will also assume for @xmath78 the aforementioned approximations that lead to eq .
[ eq : z_app ] , in such a way that we can write ( deliberately omitting the irrelevant multiplicative constants ) : @xmath79 where we have defined : @xmath80 this is what must be done in general when a soft degree of freedom is needed to be integrated out in statistical mechanics @xcite and the approximations in ref .
@xcite can not be made .
the function @xmath81 is a free energy because , in general , it depends on the temperature and it contains the entropy of the degree of freedom @xmath0 whose influence has been averaged .
we must remark at this point that , to integrate out the side chain angle @xmath0 could be reasonable if one s aim is to use the ab initio obtained information from a single dipeptide to include it in an effective potential for simulating polypeptides .
there is no point in integrating out the ramachandran angles @xmath20 and @xmath21 , since the conformation of the larger system will depend crucially on their particular values , because they lie in the backbone of the molecule and there are as many pairs @xmath82 as residues in the chain .
the side chain angle @xmath0 , on the contrary , will only influence its immediate surroundings and its importance could be of different magnitude depending on the treatment that the side chains are given in the model for the polypeptide . in this context , if we wanted to use an energy function that does not depend on @xmath0 ( in some circumstances , a computational must ) , we would have to perform the integral in the last term of eq .
[ eq : f_def ] and use @xmath81 instead of @xmath71 , since , as it has already been remarked , @xmath0 is not a hard coordinate and the approximations needed to write eq .
[ eq : z_app ] do not hold . therefore , if we compare the last term in eq .
[ eq : z_app_hyper ] with the last term in eq .
[ eq : z_app ] , we see that , apart from additive constants that do not depend on @xmath20 and @xmath21 and that come from the multiplicative constants omitted , the pes @xmath71 must be understood as a candidate for _ approximating _ the more realistic @xmath81 and saving much computational effort .
to assess the goodness of this approximation in the particular case of formyl - alanine - amide is what will be done in the following subsections .
the ab initio quantum mechanical calculations have been done with the package gamess @xcite under linux .
the coordinates used for the hco - l - ala - nh@xmath37 dipeptide in the gamess input files and the ones used to `` move '' the molecule in the the automatic perl scripts that generated the input files are the sasmic defined in secs .
[ sec : definition_gen ] and [ sec : definition_pro ] .
they are presented in table [ tab : coor_ala_pro ] indicating the name of the conventional dihedral angles ( see also fig . [
fig : num_ala ] for reference ) . in the energy optimizations , on the contrary
, they have been converted to delocalized coordinates @xcite to accelerate convergence .
cccc atom name & bond length & bond angle & dihedral angle + + h@xmath39 & & & + c@xmath1 & ( 2,1 ) & & + n@xmath40 & ( 3,2 ) & ( 3,2,1 ) & + o@xmath41 & ( 4,2 ) & ( 4,2,3 ) & ( 4,2,3,1 ) + c@xmath42 & ( 5,3 ) & ( 5,3,2 ) & @xmath83*(5,3,2,1 ) * + h@xmath43 & ( 6,3 ) & ( 6,3,2 ) & ( 6,3,2,5 ) + c@xmath44 & ( 7,5 ) & ( 7,5,3 ) & @xmath84*(7,5,3,2 ) * + c@xmath45 & ( 8,5 ) & ( 8,5,3 ) & ( 8,5,3,7 ) + h@xmath46 & ( 9,5 ) & ( 9,5,3 ) & ( 9,5,3,7 ) + h@xmath5 & ( 10,8 ) & ( 10,8,5 ) & @xmath22*(10,8,5,3 ) * + h@xmath6 & ( 11,8 ) & ( 11,8,5 ) & ( 11,8,5,10 ) + h@xmath7 & ( 12,8 ) & ( 12,8,5 ) & ( 12,8,5,10 ) + n@xmath35 & ( 13,7 ) & ( 13,7,5 ) & @xmath85*(13,7,5,3 )
* + o@xmath47 & ( 14,7 ) & ( 14,7,5 ) & ( 14,7,5,13 ) + h@xmath48 & ( 15,13 ) & ( 15,13,7 ) & @xmath86*(15,13,7,5 ) * + h@xmath49 & ( 16,13 ) & ( 16,13,7 ) & ( 16,13,7,15 ) first , we have calculated the typical pes @xmath71 defined in eq .
[ eq : pes_def ] in a regular 12x12 grid , with both @xmath20 and @xmath21 ranging from @xmath87 to @xmath88 in steps of @xmath89 .
this has been done by running energy optimizations at the level of the theorysince this is only an exploratory study , the level of the theory is considered enough to show the general trend and to illustrate the procedure to be followed in a more exhaustive assessment with more reliable ab initio methods . ] , freezing the two ramachandran angles at each value on the grid , starting from geometries previously optimized at a lower level of the theory and setting the gradient convergence criterium to opttol=0.0001 and the self - consistent hartree - fock convergence criterium to conv=0.00001 . then , at each grid point , we have defined another one - dimensional grid in the coordinate @xmath0 that ranges from @xmath90 to @xmath91 in steps of @xmath92 , where @xmath93 is one of the three equivalent equilibrium values ( selected arbitrarily ) of this degree of freedom at each point of the original pes .
this partition in 12 points spans one third of the @xmath0-space , but it is enough for computing the integrals because the surface @xmath78 has exact three - fold symmetry in @xmath0 ( note , for example , that the value of @xmath74 at @xmath94 would be equal to the one at @xmath91 ) .
next , we have run energy optimizations , with the same parameters described above and at the same level of theory , at each point of the @xmath0-grid for every grid - value of the pes ( i.e. , freezing the three angles ) .
the starting geometries have been automatically generated via perl scripts taking the final geometries in the @xmath82-grid and systematically changing @xmath0 .
note that this amounts to only changing the principal dihedral ( 10,8,5,3 ) in the in table [ tab : coor_ala_pro ] ; with poorly designed coordinates that did not separate the hard modes from the soft ones , this process would have been more difficult and rather unnatural .
after all the optimizations ( @xmath95 54 days of cpu time in 3.20 ghz piv machines ) , we have 12x12x12=1728 points with grid coordinates of the function @xmath78 and we may approximate the integral defining @xmath81 in eq .
[ eq : f_def ] by a finite sum : @xmath96 } \right ) } - \langle v_3 \rangle ( \phi_i,\psi_j ) \ .\end{aligned}\ ] ] where additive constants arising from the three - fold symmetry in the coordinate @xmath0 have been discarded . the harmless quantity @xmath97 , defined as : @xmath98 has been introduced in order for the values of the exponential function to be in the precision range of the computer .
analogously , the average energy may be computed via : @xmath99 } } { \displaystyle \sum_k \mathrm{e}^{-\beta [ v_3(\phi_k,\psi_j,\chi_k ) -\langle v_3 \rangle(\phi_i,\psi_j ) ] } } = \ .\end{aligned}\ ] ] and , finally , we extract the entropy from : @xmath100 additionally , apart from the calculations needed to integrate out @xmath0 , we have also performed an unconstrained geometry optimization in the basin of attraction of the local minima of the pes normally known as @xmath101 or c7@xmath102 depending on the author @xcite .
this calculation was done at the mp2/6 - 31++g(d , p ) level of the theory and with the same values of the variables opttol and conv than the ones used in the pes case .
the starting geometry was the final structure corresponding to the point @xmath103 of the pes calculations at the lower level of the theory described in the preceding paragraphs .
in the local minimum found , we have computed the hessian matrix ( also at mp2/6 - 31++g(d , p ) ) in two different sets of coordinates : the properly defined sasmic shown in table [ tab : coor_ala_pro ] and an ill - defined set in which the lines corresponding to the hydrogens h@xmath5 , h@xmath6 and h@xmath7 in the side chain have been substituted by those in table [ tab : bad_coord ] .
this is done to numerically illustrate the better separation of the hard and soft modes achieved by the internal coordinates defined in this work . in order to assess
if @xmath71 could be considered a good approximation of @xmath81 , we have used a statistical quantity , defined in @xcite , which measures the typical error that one makes in the energy differences between arbitrary pairs of conformations of the system if one effective potential is used instead of the other .
if we measure this _ distance _ between @xmath81 and @xmath71 , using the 144 points in the @xmath82-grid , we obtain : @xmath104 we present the result in units of @xmath105 ( at @xmath106 k , where @xmath107 kcal / mol ) because it has been argued in @xcite that , if the distance between two different approximations of the energy of the same system is less than @xmath105 , one may safely substitute one by the other without altering the relevant physical properties . _ in this case , this criterium is widely satisfied .
_ moreover , if one assumes that the effective energy studied will be used to construct a polypeptide potential and that the latter will be designed as simply the sum of mono - residue ones ( making each term suitably depend on different pairs of ramachandran angles ) , then , the number @xmath108 of residues up to which one may go keeping the distance between the two approximations of the the @xmath3-residue potential below @xmath105 is ( see ref .
@xcite ) : @xmath109 the goodness of the approximation in this case is much due to the simplicity and small size of the side chain of the alanine residue and also to the fact that the dipeptide is isolated . for bulkier residues included in polypeptides , we expect the difference between @xmath81 and @xmath71 to be more important
although the essential result is the one stated in the previous paragraphs , we wanted to look in more detail at the origin of the differences between @xmath81 and @xmath71 . for this
, we have first subtracted from @xmath81 , @xmath110 and @xmath71 the same constant reference ( @xmath111)at the level of the theory used in the calculations , the minimum of @xmath81 in the grid is -414.7985507934 hartree . ] in order to render the numerical values more manageable and to minimize the statistical error of the @xmath112-intercept in the linear fits @xcite that will be made in the following .
then , fitting @xmath110 against @xmath71 , we have found that they are more correlated than @xmath81 and @xmath71 ( compare the pearson s correlation coefficient , @xmath113 vs. @xmath114 , and the aforementioned distance , @xmath115 vs. @xmath116 ) , and that they are separated by an almost constant offset : @xmath71 is @xmath117 kcal / mol lower that @xmath110 ( on the other hand , @xmath71 is @xmath118 kcal / mol higher than @xmath81 ) .
hence , the three ramachandran surfaces @xmath81 , @xmath110 and @xmath71 are very similar , except for an offset . in fig . [
fig : rama_plots]a , @xmath81 is depicted graphically and , in fig [ fig : offsets ] , the relative offsets among the three energies are schematically shown .
contrarily , the entropy ( we use @xmath119 in order to deal with quantities that have units of energy ) , which may be found in fig . [
fig : rama_plots]b , and whose average magnitude is @xmath120 kcal / mol , is almost uncorrelated with @xmath81 , @xmath110 and @xmath71 , being the correlation coefficients @xmath121 , @xmath122 and @xmath123 , respectively .
hence , given that @xmath124 is almost an order of magnitude lower than @xmath125 , it is reasonable to conclude that the greatest part of the ( little ) noise between @xmath81 and @xmath71 comes from the entropic term @xmath126 .
this is supported by the fact that the difference @xmath127 is highly correlated with @xmath119 , being the correlation coefficient @xmath128 .
finally , and in order to illustrate the better separation of the hard and soft modes achieved by the internal coordinates defined in this work , we have calculated the hessian matrix in the minimum @xmath101 ( also c7@xmath102 ) in two different sets of coordinates .
they are described at the end of sec .
[ subsec : app_methods ] and they correspond to the sasmic set , defined according to the rules given in sec .
[ sec : definition_pro ] , and a set in which the coordinates that position the hydrogens in the side chain have been ill - defined . in fig .
[ fig : hessians ] , we present the sub - boxes of the two hessian matrices corresponding to the coordinates defined in tables [ tab : good_coord ] and [ tab : bad_coord ] . from the values shown
, one can conclude that , in the `` properly defined coordinates '' , some convenient characteristics are present : on one side , the relatively low values of the elements @xmath129 and @xmath130 ( and their symmetric ones ) indicate that the soft degree of freedom @xmath0 and the hard ones @xmath131 and @xmath132 , which describe the internal structure of the methyl group , are uncoupled to a reasonable extent ; on the other side , the relatively low value of @xmath133 compared to @xmath134 and @xmath135 ( a difference of almost an order of magnitude ) proves that @xmath0 may be regarded as soft when compared to the hard degrees of freedom @xmath131 and @xmath132 . on the contrary , in the `` ill - defined coordinates '' , the three dihedrals are hard , considerably coupled and equivalent .
extending the approach of refs . @xcite and the ideas stated in @xcite , we have defined a systematic numeration of the groups , the atoms and the internal coordinates ( termed sasmic ) of organic molecules and , particularly , of polypeptide chains .
the advantages of the rules herein presented are many - fold : * the internal coordinates may be easily cast into conventional z - matrix form and they can be directly implemented into quantum chemical packages . * the algorithm for numbering allows for automatizing and facilitates the coding of computer applications . *
the modularity of the numeration system in the case of polypeptides permits the addition of new residues without essentially changing the already numbered items .
this is convenient if databases of peptide structures need to be designed .
* the set of internal coordinates defined reasonably separate the hard and soft movements of organic molecules for arbitrary conformations using only topological information .
a number of perl scripts that automatically generate these coordinates for polypeptide chains are provided as supplementary material .
they may be found at http://neptuno.unizar.es / files / public / gen_sasmic/. in addition , we have used the coordinates herein defined and ab initio quantum mechanics to assess the approximation of the free energy obtained from averaging out the rotameric degree of freedom @xmath0 via the conventional pes in the protected dipeptide hco - l - ala - nh@xmath1 . applying the criterium in ref .
@xcite , we have found that approximating @xmath81 by @xmath71 is justified up to polypeptides of medium length ( @xmath95 100 residues ) and much computational effort may be saved using the pes instead of the more realistic free energy . however , the small size of the side chain of the alanine residue and the fact that the dipeptide is isolated do not allow to extrapolate this result . for bulkier residues included in polypeptides , we expect the difference between @xmath81 and @xmath71 to be more important .
, m. a. sahai , j. m. s. law , s. lovas , .
farkas , a. perczel , j .- l .
rivail , and i. g. csizmadia , toward a computed peptide structure database : the role of a universal atomic numbering system of amino acids in peptides and internal hierarchy of database , , 933 ( 2002 ) , s. lovas , g. a. chass , p. botond , and i. g. csizmadia , a modular numbering system of selected oligopeptides for molecular computations : using pre - computed amino acid building blocks , , 169 ( 2003 ) , , john wiley & sons , chichester , 2nd edition , 2004 , a. kessi , and b. delley , the generation and use of delocalized internal coordinates in geometry optimization , , 192 ( 1996 ) and r. ahlrichs , geometry optimization in generalized natural internal coordinates , , 9183 ( 1999 ) , j. baker , s. suhai , and p. pulay , geometry optimization of large biomolecules in redundant internal coordinates , , 6566 ( 2000 ) and m. challacombe , the quasi - independent curvilinear coordinate approximation for geometry optimization , , 2877 ( 2004 ) , g. w. trucks , h. b. schlegel , g. e. scuseria , m. a. robb , j. r. cheeseman , j. a. montgomery , jr . , t. vreven , k. n. kudin , j. c. burant , j. m. millam , s. s. iyengar , j. tomasi , v. barone , b. mennucci , m. cossi , g. scalmani , n. rega , g. a. petersson , h. nakatsuji , m. hada , m. ehara , k. toyota , r. fukuda , j. hasegawa , m. ishida , t. nakajima , y. honda , o. kitao , h. nakai , m. klene , x. li , j. e. knox , h. p. hratchian , j. b. cross , v. bakken , c. adamo , j. jaramillo , r. gomperts , r. e. stratmann , o. yazyev , a. j. austin , r. cammi , c. pomelli , j. w. ochterski , p. y. ayala , k. morokuma , g. a. voth , p. salvador , j. j. dannenberg , v. g. zakrzewski , s. dapprich , a. d. daniels , m. c. strain , o. farkas , d. k. malick , a. d. rabuck , k. raghavachari , j. b. foresman , j. v. ortiz , q. cui , a. g. baboul , s. clifford , j. cioslowski , b. b. stefanov , g. liu , a. liashenko , p. piskorz , i. komaromi , r. l. martin , d. j. fox , t. keith , m. a. al - laham , c. y. peng , a. nanayakkara , m. challacombe , p. m. w. gill , b. johnson , w. chen , m. w. wong , c. gonzalez , and j. a. pople , gaussian 03 , revision c.02 , aussian , inc . , wallingford , ct , 2004 , k. k. baldridge , j. a. boatz , s. t. elbert , m. s. gordon , h. j. jensen , s. koseki , n. matsunaga , k. a. nguyen , s. su , t. l. windus , m. dupuis , and j. a. montgomery , eneral atomic and molecular electronic structure system , , 1347 ( 1993 ) , g. fogarasi , f. pang , and j. e. boggs , systematic ab initio gradient calculation of molecular geometries , force constants , and dipole moment derivatives , , 2550 ( 1979 ) and g. fogarasi , geometry optimization in redundant internal coordinates , , 2856 ( 1992 ) , x. zhou , p. w. taylor , and p. pulay , the calculation of ab initio molecular geometries : efficient natural internal coordinates and empirical correction by offset forces , , 8191 ( 1992 ) and r. a. abagyan , new methodology for computer - aided modelling of biomolecular structure and dynamics . 1 .
non - cyclic structures , , 815 ( 1989 ) and a. k. mazur , new methodology for computer - aided modelling of biomolecular structure and dynamics .
2 . local deformations and cycles , , 833 ( 1989 ) , m. m. totrov , and d. a. kuznetsov , : a new method for protein modeling and design : applications to docking and structure prediction from the distorted native conformation , , 488 ( 1994 ) and h. a. scheraga , on the use of classical statistical mechanics in the treatment of polymer chain conformation , , 535 ( 1976 ) and j. n. kushick , method for estimating the configurational entropy of macromolecules , , 325 ( 1980 ) , , prentice hall , upper saddle river , 5th edition , 1999 , j. coffin , s. q. newton , m. ramek , v. k. w. cheng , f. a. momany , and l. schfer , importance of correlation - gradient geometry optimization for molecular conformational analyses , , 5369 ( 1992 ) and s. suhai , -acetyl - l - alanine n-methylamide : a density functional analysis of the vibrational absorption and vibrational circular dichroism spectra , , 81 ( 1996 ) , r. m. nieminen , m. knapp - mohammady , and s. suhai , vibrational analysis of various isotopomers of l - alanyl - l - alanine in aqueous solution : vibrational circular dichroism , raman , and raman optical activity spectra , , 239 ( 2002 ) , m. a. norman , l. schfer , m. ramek , a. peeters , and c. van alsenoy , ab initio conformational analysis of n - formyl l - alanine amide including electron correlation , , 361 ( 2001 ) , g. a. chass , i. g. csizmadia , p. echenique , and a. tarancn , do theoretical physicists care about the protein folding problem ? , in _ meeting on fundamental physics ` alberto galindo ' _ , edited by r. f. lvarez - estrada et al . , aula documental , madrid , 2004 , ( arxiv : q - bio.bm/0407024 ) , merck molecular force field . i. basis , form , scope , parametrization , and performance of mmff94 , , 490 ( 1996 ) , merck molecular force field .
mmff94 van der waals and electrostatica parameters for intermolecular interactions , , 520 ( 1996 ) , merck molecular force field .
iii . molecular geometrics and vibrational frequencies for mmff94 , , 553 ( 1996 ) , merck molecular force field .
iv . conformational energies and geometries for mmff94 , , 587 ( 1996 ) , merck molecular force field . v. extension of mmff94 using experimental data , additional computational data , and empirical rules , , 616 ( 1996 ) , d. s. maxwell , and j. tirado - rives , development and testing of the opls all - atom force field on conformational energetics and properties of organic liquids , , 11225 ( 1996 ) , p. cieplak , c. i. bayly , i. r. gould , j. merz , k. m. , d. m. ferguson , d. c. spellmeyer , t. fox , j. w. caldwell , and p. a. kollman , a second generation force field for the simulation of proteins , nucleic acids , and organic molecules , , 5179 ( 1995 ) and d. a. case , force fields for protein simulations , , 27 ( 2003 ) , r. e. bruccoleri , b. d. olafson , d. j. states , s. swaminathan , and m. karplus , : a program for macromolecular energy , minimization , and dynamics calculations , , 187 ( 1983 ) , b. brooks , c. l. brooks iii , l. nilsson , b. roux , y. won , and m. karplus , : the energy function and its parameterization with an overview of the program , in _ the encyclopedia of computational chemistry _ , edited by p. v. r. schleyer et al . , pp . 217277 , john wiley & sons , chichester , 1998 , m. feig , and c. l. brooks iii , extending the treatment of backbone energetics in protein force fields : limitations of gas - phase quantum mechanics in reproducing protein conformational distributions in molecular dynamics simulations , , 1400 ( 2004 ) , h. a. baldoni , f. suvire , r. nieto vzquez , g. zamarbide , r. d. enriz , .
farkas , a. perczel , m. a. mcallister , l. l. torday , j. g. papp , and i. g. csizmadia , characteristics of ramachandran maps of l - alanine diamides as computed by various molecular mechanics , semiempirical and ab initio mo methods . a search for primary standard of peptide conformational stability , , 275 ( 1998 ) , o. farkas , i. jakli , i. a. topol , and i. g. csizmadia , peptide models .
extrapolation of low - level hartree - fock data of peptide conformation to large basis set scf , mp2 , dft and ccsd(t ) results .
the ramachandran surface of alanine dipeptide computed at various levels of theory , , 1026 ( 2003 ) and w. j. briels , free energy from molecular dynamics with multiple constraints , , 773 ( 2000 ) and m. karplus , thermodynamics of protein folding : a microscopic view , , 367 ( 2003 ) and p. echenique , a physically meaningful method for the comparison of potential energy functions , ( 2005 ) , ( arxiv : q - bio.bm/0504029 ) and d. k. robinson , , mc . graw hill , new york , 3rd edition , 2003 , s. a. teukolsky , w. t. vetterling , and b. p. flannery , , cambridge university press , new york , 2nd edition , 2002 | a set of rules is defined to systematically number the groups and the atoms of organic molecules and , particularly , of polypeptides in a modular manner . supported by this numeration
, a set of internal coordinates is defined .
these coordinates ( termed systematic , approximately separable and modular internal coordinates , sasmic ) are straightforwardly written in z - matrix form and may be directly implemented in typical quantum chemistry packages .
a number of perl scripts that automatically generate the z - matrix files for polypeptides are provided as supplementary material . the main difference with other z - matrix - like
coordinates normally used in the literature is that normal dihedral angles ( `` principal dihedrals '' in this work ) are only used to fix the orientation of whole groups and a somewhat non - standard type of dihedrals , termed `` phase dihedrals '' , are used to describe the covalent structure inside the groups .
this _ physical approach _ allows to approximately separate soft and hard movements of the molecule using only topological information and to directly implement constraints . as an application
, we use the coordinates defined and ab initio quantum mechanical calculations to assess the commonly assumed approximation of the free energy , obtained from `` integrating out '' the side chain degree of freedom @xmath0 , by the potential energy surface ( pes ) in the protected dipeptide hco - l - ala - nh@xmath1 .
we also present a sub - box of the hessian matrix in two different sets of coordinates to illustrate the approximate separation of soft and hard movements when the coordinates defined in this work are used . +
* pacs : * 87.14.ee , 87.15.-v , 87.15.aa , 87.15.cc , 89.75.-k + |
the role of starbursts in active galactic nuclei ( agn ) is an issue which permeates the history of agn literature .
theoretical scenarios for a `` starburst - agn connection '' vary from those in which starbursts are a key piece of the agn machinery , either through symbiotic or evolutionary processes ( e.g. , perry & dyson 1985 ; sanders 1988 ; norman & scoville 1988 ; rees 1989 ; daly 1990 ; terlevich 1992 ; williams , baker & perry 1999 ) , to those in which these are fundamentally decoupled phenomena , but which are likely to coexist merely because both are triggered by and live on gas fueling ( e.g. , byrd 1986 ; byrd , sundeluis & valtonen 1987 ; lin , pringle & rees 1988 ; heller & shlosman 1994 ; hernquist & mihos 1995 ; mihos & hernquist 1996 ) .
given the enormous breadth of possibilities in between these extremes , it is likely that progress in understanding this connection will be driven more by input from observations than from theoretical considerations .
observational work during the past decade has indeed substantially advanced our ability to recognize starbursts in the complex inner environment of agn .
optical and uv studies of seyfert 2 galaxies have been particularly enlightening in this respect . in a series of papers ( heckman 1997 ; cid fernandes , storchi - bergmann & schmitt 1998 ; gonzlez delgado 1998 ; storchi - bergmann , cid fernandes & schmitt 1998 ; schmitt , storchi - bergmann & cid fernandes 1999 ; storchi - bergmann 2000 ; gonzlez delgado , heckman & leitherer 2001 ) we have detected unambiguous signatures of young massive stars within @xmath9 pc of the nucleus in 30 to 50% of seyfert 2 s by means of high quality optical and , whenever possible , uv spectroscopy .
this result fits well with the near ir studies of oliva ( 1995 , 1999 ) , which find a comparable incidence of starbursts in seyfert 2 s through measurements of the stellar mass - to - light ratio .
the starbursts in these composite starburst@xmath0agn systems can make a significant contribution to the total luminosity output .
in fact , studies specifically aimed at candidate composite galaxies , as selected by emission line ratios intermediate between hii regions and agn on classical diagnostic diagrams , demonstrate that their far ir and radio properties are dominated by star - forming activity at the 90% level ( hill 2001 , 1999 ) .
even the presence of compact radio cores , a more classical indicator of agn activity ( condon 1991 ; sramek & weedman 1986 ; norris 1990 ) , can be accounted for by star - formation ( hill 2001 ; kewley 2000 ; smith 1998a , b ; lonsdale , smith & lonsdale 1993 ) .
other recent results are covered in the review by veilleux ( 2001 ) and the volume edited by aretxaga , kunth & mugica ( 2001 ) .
evidence is also steadily accumulating that associates agn with star - formation on galactic scales .
the ubiquity of super - massive black holes in the nuclei of normal galaxies in the local universe ( ho 1999 ) , the proportionality between bulge and black - hole masses ( magorrian 1998 ; gebhart 2000 ; ferrarese & merrit 2000 ) , the link between @xmath10 and the age of the last major star - formation episode in the spheroid ( merrifield , forbes & terlevich 2000 ) , all imply that the creation of these black holes and the ensuing qso activity were an integral part of the formation of ellipticals and galactic bulges .
this fossil evidence indirectly traces a much more active past , in which both copious star - formation and nuclear activity coexisted .
since the technology to directly study this high redshift era in detail is not available , we must guide our study of the starburst - agn connection by data gathered on nearby objects in which traces of this connection are caught _ in fraganti_. this means refining our knowledge of circum - nuclear starbursts in seyfert galaxies . in the optical - uv range , as hinted above , this is best accomplished studying type 2 seyferts . their favorable geometry , in which the blinding glare of the nucleus is blocked away by a dusty torus , facilitates the detection of features from circum - nuclear starbursts , thus making them the best - suited local laboratories to study the starburst - agn connection .
advancing our empirical understanding of circum - nuclear starbursts in seyfert galaxies can be described as a 3-stage process : ( i ) identifying starbursts in these systems ; ( ii ) characterizing the properties of the starburst ( age and mass , or star - formation rate ) and the agn ( black - hole mass , accretion rate ) ; and ( iii ) investigating possible connections between these properties .
once we get to this third stage fundamental questions can start to be addressed .
for instance : is star - formation inextricably associated with agn activity ?
are the starburst and agn powers related , say , by a @xmath11 proportionality between their masses ? or , analogously , is the star - formation rate connected to the rate of accretion onto the nuclear super - massive black - hole ? does the central engine evolve in parallel to the starburst around it ?
much work remains to be done before tackling these key questions .
the work on seyfert 2 s reviewed above pertains mostly to stage i of this process , as it essentially demonstrates how certain spectral features of massive stars can be used as sign - posts of starburst activity .
estimates of the recent history of star - formation in the circum - nuclear regions were also presented , thus touching the domain of stage ii .
these estimates come associated with known difficulties in accurately retrieving the detailed star - formation history of stellar systems ( e.g. , leitherer 1999 ; cid fernandes 2001 ; kennicutt 1998 ) , but they represent the current state - of - the - art regarding the characterization of the basic starburst properties .
this paper deals mainly with stage i , that is , the identification of starbursts in seyfert 2 s , using the data sets of cid fernandes ( 1998 ) and gonzlez delgado ( 2001 , hereafter gd01 ) as a training set .
this is probably the largest sample whose optical spectra have been systematically and carefully screened for starburst features , so it represents a good starting point .
the analysis is geared towards verifying whether composite seyfert 2/starburst systems , those in which stellar features produced by young populations have been directly detected , exhibit other properties which differentiate them from systems not showing signs of starburst activity ( `` pure '' seyfert 2 s ) . unlike the detailed analysis reported by the previous work on these data , we here limit ourselves to easily measurable quantities .
specifically , we will use equivalent widths of strong absorption bands and continuum colors in the near uv , emission line equivalent widths , fluxes and line ratios , far ir luminosities and near uv surface brightness .
the purpose of this analysis is two - fold : * to verify whether the optical continuum and emission line properties of composite systems are consistent with the expected impact of starbursts upon these observables . * to investigate alternative ways to infer the presence of circum - nuclear starbursts in seyfert 2 s that may provide more straightforward diagnostics applicable to larger and/or more distant samples .
the first of these points serves both as a consistency check and to give a sense of the role played by starbursts in the phenomenology and energetics of agn .
the relevance of the second point is that increasing the statistics is crucial to improve upon the still sketchy overall census of circum - nuclear starbursts in seyfert 2 s , to extend these studies to higher redshifts and to provide the raw material for further stage ii and iii studies . in [ sec : sample ] we describe the data for the seyfert 2 s and comparison samples used in this study . in [ sec : synthesis ] we deal with the problem of estimating the `` featureless continuum '' ( fc ) strength .
we apply a method based on population synthesis , but which can be parameterized in terms of a few easily measurable quantities . essentially , we are able to estimate the fc contribution to the optical light with little more than the equivalent width of the caii k absorption line , a method which may be of wide applicability in seyfert 2 studies . in [ sec : emission_lines ] we present an analysis of the emission line equivalent widths , line ratios and profiles , and discuss how they are affected by the presence of circum - nuclear starbursts .
emission line - fc correlations are also presented and discussed . in [ sec : iras ] we investigate how the composite systems behave in terms of far ir luminosity and how they fare in comparison to normal and interacting galaxies . in [
sec : surfbrightness ] we compare the near uv surface brightness of seyfert 2 s in our sample to that of normal and starburst galaxies . in
[ sec : discussion ] we collect the results obtained in a set of empirical criteria which may be used to diagnose the presence of circum - nuclear starbursts in seyfert 2 s .
we also discuss the meaning of the composite/``pure '' seyfert 2 classification and present tentative evidence of evolutionary effects in our sample .
section [ sec : conclusions ] summarizes our conclusions .
in this study we merge the seyfert 2 s in the southern sample studied by cid fernandes ( 1998 ) with those in the northern sample of gd01 , totaling 35 galaxies .
details of the observations , sample selection and previous analysis of these data can be found in the original papers and heckman ( 1997 ) , gonzlez delgado ( 1998 ) , storchi - bergmann ( 1998 ) , schmitt ( 1999 ) and storchi - bergmann ( 2000 , hereafter sb00 ) .
we concentrate our analysis on the nuclear spectra , extracted through @xmath12 apertures for the northern sample and @xmath13 for the southern one . at the distances of our galaxies
this covers a region 60860 pc in radius , with a median of 300 pc ( @xmath14 km@xmath15s@xmath16mpc@xmath17 is adopted throughout this paper ) .
we will also focus on the near uv to optical region of the spectrum , between 3500 and 5100 .
these 35 galaxies will hereafter be referred to as the `` seyfert 2 sample '' .
they constitute a good data set to investigate the starburst - agn connection because the above papers have scrutinized these sources in search for signs of a starburst component .
specifically , the following signatures of a starburst were examined : ( 1 ) the presence of far - uv stellar wind lines ( nv @xmath181240 , siiv@xmath181397 and civ@xmath181549 ) due to o stars ; ( 2 ) high order balmer absorption lines of hi and hei in the near uv ; ( 3 ) the wr bump underneath heii@xmath184686 .
the far - uv features are the cleanest signatures of recent star formation , but , unfortunately , there are only a handful of seyfert 2 s for which hst far - uv spectroscopy is feasible .
all four galaxies in the sample for which we acquired such data ( mrk 477 , ngc 5135 , ngc 7130 and ic
3639 ) have their far - uv spectrum dominated by young stars . for the remaining galaxies ,
detection of a starburst component relies on either high order balmer lines , which originate in the photosphere of o , b and a stars , or the wr bump , a tracer of very recent ( 36 myr ) star formation .
although more subtle than the far - uv features , these are reliable and more easily accessible indicators of the presence of a starburst .
this is confirmed by the fact that either of these two features is also detected in the galaxies for which we have identified stellar wind lines in the far - uv .
the least certain of these diagnostics is the wr bump , because independent evidence of the presence of wr stars is hard to obtain ( gd01 ; shaerer 2001 ) .
nevertheless , this is the simplest interpretation of this feature , given the similarity to the feature sometimes seen in star - forming galaxies ( `` wr galaxies '' ) , the lack of convincing alternatives and the simultaneous detection of other starburst features in some of the sources .
the information to be used in this paper comprises fluxes and equivalent widths ( @xmath19 ) of the [ oii]@xmath203726,3729 , heii@xmath184686 , h@xmath3 and [ oiii]@xmath185007 emission lines ; the @xmath19 s of caiik@xmath183933 ( @xmath21 ) , cn@xmath184200 ( @xmath22 ) and g - band@xmath184301 ( @xmath23 ) absorption bands , the continuum fluxes at 3660 , 4020 and 4510 , as well as the 12 @xmath24 m and far ir luminosities as measured by iras .
these are listed in tables [ tab : tabelao1 ] and [ tab : tabelao2 ] , along with heliocentric velocities extracted from ned .
table [ tab : tabelao2 ] also lists the h@xmath25/h@xmath3 ratio for the seyfert 2 sample , mostly compiled from the literature , since the kitt peak spectra do not cover the h@xmath25 region .
these measurements come from spectra obtained with apertures larger than those used in our observations . for this reason , and also because of the intrinsic uncertainties associated with reddening corrections , we will concentrate on reddening insensitive diagnostics inasmuch as possible .
galaxies which had hidden seyfert 1 nuclei revealed through spectropolarimetry are indicated by the corresponding reference in the last column of table [ tab : tabelao2 ] . all absorption line @xmath19 s and continuum fluxes used in this paper were measured according to the system explained in cid fernandes ( 1998 ) .
emission line and continuum fluxes were corrected for galactic extinction following the extinction law of cardelli , clayton & mathis ( 1989 , with @xmath26 ) , and @xmath27 values from schlegel , finkbeiner & davis ( 1998 ) as listed in ned .
lrrrrrrrrrrrrrr cgcg420 - 015 & 14.0 & 9.7 & 8.4 & 0.53 & 1.67 & 32.5 & 4.7 & 16.9 & 208.5 & 7.9@xmath28 & 3.6@xmath28 & 1.7@xmath29 & 1.4@xmath30 eso 417-g6 & 14.6 & 11.3 & 9.4 & 0.57 & 1.29 & 42.1 & & 2.8 & 18.1 & 1.8@xmath30 & & 2.3@xmath30 & 2.8@xmath28 eso 362-g8 & 7.3 & 6.6 & 5.9 & 0.35 & 1.21 & 3.2 & & & 13.1 & 4.2@xmath28 & & 5.3@xmath30 & fairall 316 & 17.2 & 14.2 & 11.3 & 0.61 & 1.59 & 30.1 & & 1.4 & 22.2 & 8.8@xmath28 & & 2.0@xmath30 & 1.2@xmath28 ic 1816 & 11.8 & 10.0 & 8.2 & 0.65 & 1.29 & 59.8 & 4.7 & 15.7 & 244.7 & 3.0@xmath30 & 4.9@xmath28 & 2.6@xmath29 & 1.8@xmath30 ic 3639 & 8.0 & 4.6 & 5.4 & 0.67 & 1.07 & 40.3 & 2.8 & 16.2 & 122.6 & 4.4@xmath30 & 4.7@xmath28 & 2.1@xmath29 & 2.7@xmath30 iras 11215 - 2806 & 13.1 & 7.5 & 7.3 & 0.54 & 1.28 & 19.9 & 1.4 & 2.9 & 52.9 & 6.1@xmath28 & 1.1@xmath28 & 4.3@xmath30 & 2.2@xmath28 mcg -05 - 27 - 013 & 15.4 & 10.9 & 9.7 & 0.71 & 1.54 & 114.2 & 7.2 & 22.9 & 244.8 & 1.2@xmath30 & 2.0@xmath28 & 8.1@xmath30 & 7.0@xmath28 mrk 1 & 7.6 & 6.0 & 4.4 & 0.83 & 1.11 & 131.4 & 12.2 & 53.5 & 765.9 & 3.4@xmath30 & 4.8@xmath28 & 2.7@xmath29 & 2.2@xmath30 mrk 3 & 11.3 & 10.2 & 6.7 & 1.04 & 1.38 & 263.4 & 12.3 & 67.8 & 752.7 & 4.9@xmath29 & 3.6@xmath30 & 3.0@xmath31 & 2.1@xmath29 mrk 34 & 9.9 & 6.7 & 7.0 & 0.81 & 1.20 & 157.2 & 11.2 & 42.1 & 475.0 & 2.7@xmath30 & 2.7@xmath28 & 1.0@xmath29 & 1.0@xmath30 mrk 78 & 9.6 & 7.6 & 6.2 & 0.48 & 1.25 & 96.0 & 3.3 & 9.3 & 131.2 & 2.5@xmath30 & 2.3@xmath28 & 9.0@xmath30 & 6.6@xmath28 mrk 273 & 5.6 & 5.2 & 3.9 & 0.58 & 1.03 & 46.4 & & 11.9 & 35.8 & 7.6@xmath28 & & 7.9@xmath28 & 2.8@xmath28 mrk 348 & 12.5 & 8.4 & 7.7 & 0.76 & 1.30 & 118.6 & 3.7 & 22.1 & 274.7 & 4.7@xmath30 & 2.8@xmath28 & 2.2@xmath29 & 1.7@xmath30 mrk 463 & 3.0 & 5.6 & 0.9 & 1.05 & 1.04 & 135.0 & 11.6 & 88.1 & 978.5 & 5.5@xmath30 & 4.7@xmath28 & 3.3@xmath29 & 3.8@xmath30 mrk 477 & 0.3 & 2.4 & 0.2 & 1.10 & 0.98 & 168.1 & 13.2 & 92.5 & 743.2 & 7.6@xmath30 & 6.4@xmath28 & 4.5@xmath29 & 4.5@xmath30 mrk 533 & 4.1 & 3.1 & 3.1 & 0.68 & 1.06 & 49.8 & 9.2 & 36.3 & 429.2 & 2.2@xmath30 & 6.1@xmath28 & 2.2@xmath29 & 2.1@xmath30 mrk 573 & 14.1 & 11.8 & 9.0 & 0.67 & 1.44 & 83.2 & 9.2 & 20.6 & 282.7 & 4.9@xmath30 & 1.1@xmath30 & 3.1@xmath29 & 2.5@xmath30 mrk 607 & 13.8 & 12.3 & 9.9 & 0.61 & 1.43 & 19.1 & 4.0 & 7.1 & 78.1 & 8.6@xmath28 & 4.5@xmath28 & 9.5@xmath30 & 8.0@xmath28 mrk 1066 & 5.6 & 4.3 & 3.5 & 0.75 & 1.03 & 43.9 & 2.3 & 24.3 & 93.1 & 5.9@xmath30 & 4.1@xmath28 & 1.6@xmath29 & 4.2@xmath30 mrk 1073 & 5.7 & 5.4 & 4.7 & 0.74 & 1.06 & 40.2 & 5.1 & 19.9 & 122.1 & 3.2@xmath30 & 5.7@xmath28 & 1.3@xmath29 & 2.2@xmath30 mrk 1210 & 7.5 & 7.6 & 5.5 & 0.83 & 1.20 & 59.3 & 8.8 & 57.0 & 614.0 & 4.1@xmath30 & 9.5@xmath28 & 5.6@xmath29 & 5.5@xmath30 ngc 1068 & 7.4 & 7.7 & 5.1 & 0.72 & 1.26 & 33.4 & 11.6 & 35.0 &
343.7 & 3.2@xmath29 & 1.9@xmath29 & 7.3@xmath31 & 5.6@xmath29 ngc 1358 & 17.8 & 16.1 & 11.5 & 0.54 & 1.50 & 17.8 & & 0.9 & 19.7 & 9.4@xmath28 & & 3.2@xmath30 & 1.4@xmath28 ngc 1386 & 13.2 & 9.0 & 8.1 & 0.54 & 1.56 & 25.5 & 2.5 & 3.4 & 58.8 & 2.3@xmath30 & 6.8@xmath28 & 1.6@xmath29 & 9.2@xmath28 ngc 2110 & 13.5 & 12.8 & 8.9 & 0.78 & 1.31 & 177.6 & 2.0 & 15.8 & 78.3 & 8.6@xmath30 & 1.9@xmath28 & 8.1@xmath30 & 1.6@xmath30 ngc 3081 & 13.7 & 10.5 & 8.7 & 0.69 & 1.38 & 36.6 & 4.9 & 9.3 & 128.0 & 7.9@xmath28 & 2.3@xmath28 & 6.6@xmath30 & 4.7@xmath28 ngc 5135 & 2.5 & 1.7 & 2.5 & 0.77 & 0.87 & 20.6 & 1.8 & 13.2 & 51.3 & 4.5@xmath30 & 4.1@xmath28 & 1.1@xmath29 & 2.7@xmath30 ngc 5643 & 9.4 & 4.2 & 4.6 & 0.50 & 1.24 & 56.4 & 2.8 & 8.7 & 115.0 & 5.7@xmath30 & 6.8@xmath28 & 2.7@xmath29 & 1.9@xmath30 ngc 5643 & 9.4 & 4.2 & 4.6 & 0.50 & 1.24 & 56.4 & 2.8 & 8.7 & 115.0 & 3.0@xmath29 & 3.5@xmath30 & 1.4@xmath31 & 1.0@xmath29 ngc 5929 & 14.3 & 11.0 & 9.6 & 0.71 & 1.39 & 107.0 & 1.6 & 14.5 & 58.5 & 3.9@xmath30 & 1.2@xmath28 & 5.0@xmath30 & 1.2@xmath30 ngc 6300 & 15.9 & 4.0 & 5.8 & 0.49 & 1.64 & 15.9 & & & 27.9 & 3.3@xmath28 & & 2.4@xmath30 & ngc 6890 & 12.6 & 1.0 & 4.8 & 0.72 & 1.42 & 16.5 & 2.4 & 3.9 & 74.2 & 4.5@xmath28 & 1.3@xmath28 & 4.4@xmath30 & 2.2@xmath28 ngc 7130 & 3.7 & 2.2 & 2.4 & 0.73 & 1.00 & 24.8 & 2.3 & 14.8 & 104.0 & 1.4@xmath30 & 1.8@xmath28 & 7.3@xmath30 & 1.1@xmath30 ngc 7212 & 10.4 & 8.1 & 6.9 & 0.92 & 1.33 & 211.2 & 13.8 & 55.2 & 725.5 & 7.0@xmath30 & 7.0@xmath28 & 3.5@xmath29 & 3.0@xmath30 ngc 7582 & 3.7 & 3.1 & 3.0 & 0.55 & 1.23 & 19.2 & 1.6 & 15.2 & 33.0 & 1.3@xmath30 & 2.6@xmath28 & 5.6@xmath30 & 2.4@xmath30 [ tab : tabelao1 ] lrrrrrrrr cgcg420 - 015 & 0.37 & 8811 & 5.5 & 10.43 & 9.76 & 40.37 & dg92 & eso 417-g6 & 0.09 & 4884 & 4.7 & 9.29 & 8.45 & 39.16 & s98 & eso 362-g8 & 0.14 & 4785 & & 9.34 & 9.16 & & & fairall 316 & 0.60 & 4950 & 6.4 & & & 38.80 & s98 & ic 1816 & 0.12 & 5080 & 6.0 & 9.75 & 10.03 & 39.98 & dg92 & ic 3639 & 0.30 & 3275 & 6.1 & 9.98 & 10.35 & 39.79 & dg92 & h97 iras 11215 - 2806 & 0.31 & 4047 & 3.8 & 9.56 & 9.08 & 38.88 & dg92 & mcg -05 - 27 - 013 & 0.28 & 7162 & 5.0 & 9.75 & 9.99 & 39.88 & dg92 & mrk 1 & 0.26 & 4780 & 5.9 & 10.47 & 10.18 & 40.03 & m94 & mrk 3 & 0.81 & 4050 & 6.6 & 10.21 & 10.18 & 40.86 &
m94 & mg90 , t95 mrk 34 & 0.04 & 15140 & 10.5 & 10.33 & 10.66 & 40.69 & m94 & mrk 78 & 0.15 & 11137 & 6.5 & 10.34 & 10.39 & 40.24 & m94 & mrk 273 & 0.04 & 11326 & 9.3 & 10.62 & 11.84 & 39.88 & m94 & mrk 348 & 0.29 & 4507 & 6.0 & 9.94 & 9.84 & 39.86 & m94 & mg90 , t95 mrk 463e & 0.13 & 14895 & 5.6 & 11.19 & 11.07 & 41.25 & m94 & mg90 , t95 mrk 477 & 0.05 & 11332 & 5.4 & 10.35 & 10.67 & 41.09 & m94 & t95 mrk 533 & 0.25 & 8713 & 5.0 & 10.85 & 11.08 & 40.54 & m94 & mg90 , t95 , h97mrk 573 & 0.10 & 5174 & 4.2 & 9.66 & 9.88 & 40.16 & m94 & k94 mrk 607 & 0.20 & 2716 & 4.9 & 9.53 & 9.63 & 39.10 & d88 & mrk 1066 & 0.57 & 3605 & 8.5 & 9.90 & 10.56 & 40.06 & m94 & mrk 1073 & 0.69 & 6998 & 6.3 & 10.48 & 11.04 & 40.35 & m94 & mrk 1210 & 0.13 & 4046 & 5.2 & 10.05 & 9.85 & 40.28 & b99 & t95 ngc 1068 & 0.14 & 1137 & 4.5 & 10.85 & 10.79 & 40.18 & m94 & am85 ngc 1358 & 0.27 & 4028 & 3.4 & 8.95 & 9.34 & 38.69 & m94 & ngc 1386 & 0.05 & 868 & 4.9 & 8.71 & 9.09 & 38.17 & m94 & ngc 2110 & 1.62 & 2335 & 8.1 & 9.42 & 9.79 & 39.28 & m94 & ngc 3081 & 0.24 & 2385 & 4.5 & & & 38.75 & m94 & m00 ngc 5135 & 0.26 & 4112 & 6.1 & 10.17 & 10.93 & 39.99 & m94 & ngc 5643 & 0.73 & 1199 & 6.2 & 9.34 & 9.95 & 38.77 & m94 & ngc 5929 & 0.10 & 2492 & 6.2 & 9.57 & 10.21 & 39.21 & m94 & ngc 6300 & 0.42 & 1110 & & 9.20 & 9.80 & & & ngc 6890 & 0.17 & 2419 & 4.2 & 9.44 & 9.87 & 38.43 & m94 & ngc 7130 & 0.12 & 4842 & 6.6 & 10.28 & 11.05 & 39.73 & s98 & ngc 7212 & 0.31 & 7984 & 5.0 & 10.24 & 10.74 & 40.61 & m94 & t95 ngc 7582 & 0.06 & 1575 & 8.3 & 9.74 & 10.54 & 39.11 & m94 & [ tab : tabelao2 ] fifteen of our seyfert 2 s have unambiguous evidence for starbursts as revealed by at least one of the 3 diagnostics outlined above .
these are eso 362-g8 , ic 3639 , mrk 1 , mrk 78 , mrk 273 , mrk 477 , mrk 463e , mrk 533 , mrk 1066 , mrk 1073 , mrk 1210 , ngc 5135 , ngc 5643 , ngc 7130 and ngc 7582 ( observed prior to the type transition reported in aretxaga 1999 ) , which will henceforth be tagged _ composite _ systems .
the circum - nuclear starbursts in these galaxies cover a wide range of properties , from very young bursts still in the wr phase ( e.g. , mrk 477 , mrk 1 ) to more mature systems in a `` post - starburst '' phase with deep high order balmer absorption lines ( e.g. , mrk 78 , eso 362-g8 ) , and systems which have a mixture of both young and intermediate age stars ( e.g. , ngc 5135 , ngc 7130 ) .
the remaining 20 sources will be labeled `` _ pure _ '' seyfert 2 s .
this denomination is used simply to indicate that no clear signs of starburst activity have been detected in our previous studies .
the spectra among our `` pure '' seyfert 2 s vary from galaxies whose optical light is entirely dominated by an old , red stellar population ( e.g. , ngc 1358 and ngc 5929 ) to galaxies which show clear signs of an uv - excess ( e.g. , mrk 3 and mrk 348 ) .
it is nearly impossible to tell on the basis of optical spectra alone whether this fc is associated with scattered light , a starburst or a combination of both .
this point was discussed at length by sb00 , who argue that as many as 30% of seyfert 2 s can fall in this ambiguous category .
gd01 also finds that several galaxies are equally well described in terms of a scattered power - law or a starburst component , depending basically on what is chosen to represent the underlying old stellar population .
far - uv spectroscopy offers the best hope to break this degeneracy , but this is not practical with currently available instrumentation , so we will have to live with this ambiguity for some time . indeed , this ambiguity pervades all discussions in this paper .
we emphasize that the composite and `` pure '' categories are _ not _ meant as new elements in the agn taxonomy .
it is clear that this classification reflects a _
effect , since we define composites as systems in which stars with ages of @xmath32 yr have been conclusively detected .
this definition therefore introduces an inevitable bias towards starbursts that are powerful in contrast with the other components of seyfert 2 spectra , a bias which will become evident in several of the results presented in this paper . while it is likely that some `` pure '' sources eventually be `` up - graded '' to composites with deeper or different observations ( e.g. , when far - uv spectra are obtained , or when near - ir mass to light ratios are measured ) , it is unlikely that any of the composites will ever be reclassified as a `` pure '' seyfert 2 . until such further notice
, all galaxies not in the list of 15 composites above will be designated `` pure '' seyfert 2 s .
the composite/``pure '' classification will be used throughout this paper as a guide to interpret our results , even though we shall _ not _ make explicit use of any of the properties on which this classification is based .
instead , we seek to find where composite sources fit into more general and easily accessible properties of seyfert 2 s . as comparison samples we draw data from three different observational studies .
the first is the spectroscopy of 61 starburst and narrow line active galaxies ( nlagn ) by storchi - bergmann , kinney & challis ( 1995 ) and mcquade , kinney & calzetti ( 1995 , hereafter collectively referred to as sbmq95 ) .
the starburst galaxies in their study will be referred to as the `` starburst sample '' , which will be extensively used as a benchmark when gauging the results for our seyfert 2 s .
ugca410 was removed from our analysis because of apparently incompatible values of @xmath21 and @xmath23 , leaving a total of 36 starburst galaxies .
the remaining 24 sources in sbmq95 contain a mixture of seyfert 2 s , liners and starburst / agn composites , so we give them a generic denomination of `` nlagn sample '' . in principle
these galaxies serve as a test - bed to evaluate the efficacy of the diagnostics of compositeness developed from our seyfert 2 sample . in practice , we will make limited use of this sample because of the mixed nature of its sources , which complicates the interpretation .
galaxies in the starburst and nlagn samples were observed with very large @xmath33 apertures to match the iue observations , sampling physical scales of typically @xmath341 kpc in radius at the distances of the sources .
a third comparison sample used here is the `` merger sample '' from the study of liu & kennicutt ( 1995a , b , hereafter lk95 ) .
this sample comprises 28 nuclear optical spectra integrated over a region corresponding to @xmath352 kpc in radius .
the emission line properties of these galaxies are quite heterogeneous , comprising 5 seyfert 2 s , 9 hii nuclei and 14 liners .
lutz , veilleux & genzel ( 1999 ) have used iso data to show that ir - luminous galaxies with hii or liner classifications are predominantly powered by starbursts , while dusty agn are energetically significant in those classified as seyfert 2 s .
thus , we will use the merger sample to empirically define the properties of powerful merger - induced dusty starbursts . the data which will be used from these samples consists of fluxes and @xmath19 s of [ oii ] , h@xmath3 and [ oiii ] ; @xmath21 , @xmath22 and @xmath23 , plus the continuum fluxes at 3660 , 4020 , and 4510 and iras fluxes .
both the sbmq95 and lk95 data offer more information than this , but we shall limit our analysis to the properties above for consistency with the seyfert 2 sample . fig .
[ fig : aperture_x_wk ] shows the projected aperture radii , defined as @xmath36slit area@xmath37 , against @xmath21 for the seyfert 2 , nlagn , starburst and merger samples .
though the physical region sampled in the comparison samples extend well beyond that covered by the nuclear spectra of the seyfert 2 sample , there is some overlap at small @xmath38 which guarantees the fairness of the comparisons presented in this paper , and in any case aperture effects will be discussed whenever relevant .
[ fig : aperture_x_wk]a also shows that the composite classification is _ not _ related to the physical size of the region sampled by our spectra . in principle
one would expect that systems observed through larger apertures are more prone to exihibit starburst features due to the inclusion of light from off - nuclear star - forming regions .
this effect has been quantified by heckman ( 1995 ) in their analysis of the aperture dependent properties of ngc 1068 , whose 1 kpc ring of hii regions dominates the light for observations taken through kpc - scale apertures .
such rings are in fact rather common ( e.g. , pogge 1989 ; wilson 1991 ; storchi - bergmann , wilson & baldwin 1996 ; colina 1997 ) , and thus can play a role in defining a composite nature for more distant objects , as also discussed by heckman etal ( 1995 ) . if such aperture effects were pronounced in the seyfert 2 sample we would expect composites to be concentrated towards large @xmath38 , whereas fig .
[ fig : aperture_x_wk]a shows that composites and `` pure '' seyfert 2 s are similarly distributed as a function of @xmath38 .
furthermore , most of the apertures for the seyfert 2 sample are much smaller than the 1 kpc typical of star - forming rings .
this effect does not affect the seyfert 2 sample , as our circum - nuclear starbursts are relatively compact ( @xmath34 few hundred pc ) .
these dimensions have in fact been confirmed by hst uv imaging for mrk 477 ( heckman 1997 ) , ngc 5135 , ngc 7130 and ic 3639 ( gonzlez delgado 1998 ) , and are comparable to the inner star - forming structures identified in the circinus galaxy ( wilson 2000 ) .
a further example of how compact circum - nuclear starbursts around agn can be is given in shields ( 2000 ) .
another dataset that will be used in this paper is that of jansen ( 2000 ) .
this sample contains multicolor surface photometry for 198 galaxies , most of which classified as normal .
we used their data to calculate the 3660 surface brightness of these galaxies inside radii of 250 and 500 pc , similar to the apertures used for our galaxies .
these values are used to compare the uv surface brightness of seyfert and normal galaxies .
the 4 seyfert 1 and 1 bl lac in their sample were excluded from our analysis .
determining the fc contribution to the optical spectrum of seyfert 2 s is an important but difficult task , for which there is no definitive recipe .
the traditional and still most widely employed method , which dates back to at least koski ( 1978 ) , is to remove the contribution of the old red stars from the host s bulge by adopting an elliptical or early type spiral normal galaxy as a spectral template .
the spectral decomposition is accomplished by assuming a @xmath39 power - law for the `` non - stellar '' fc , whose strength is then adjusted to minimize the residuals , yielding typically @xmath402 . naturally , this procedure precludes the detection of young stellar populations , whose most conspicuous optical absorption features are masked by emission lines even in starburst galaxies . furthermore , there is no simple way to tell apart a power law fc from that produced by a reddened starburst ( e.g. , cid fernandes & terlevich 1995 ) . only upon careful scrutiny of high quality spectra
is it possible to discern the presence of a starburst from its high order balmer lines or , sometimes , the wr bump ( see discussion in gd01 ) .
therefore , _ the fc strength will in general include a contribution from young stars as well as a _
bona fide non - stellar fc _ _ , composed of scattered light from the hidden nucleus plus nebular continuum .
`` featureless '' is of course a misnomer for a spectral component at least partly associated with starlight .
we nonetheless retain this terminology both for historical reasons and because to first order a young stellar population does produce weak spectral features in the optical range .
when referring to scattered light from the hidden nucleus we will use the term ` fc1 ' introduced by tran ( 1995c ) . as argued by cid fernandes & terlevich ( 1995 ) , heckman ( 1995 ) and tran ( 1995c ) , strong fc seyfert 2 s must also contain another source of fc photons , ` fc2 ' in tran s notation , to account for the absence of conspicuous broad lines and the larger polarizations observed in the reflected broad lines than in the reflected nuclear continuum .
at least for the composite systems , there is little doubt that fc2 is associated with circum - nuclear starbursts ( see gd01 ) .
nebular continuum and emission from the scattering region itself ( tran 1995c ) are other possible contributors to fc2 . in this paper
we explore a different way to determine the fc strength .
we decompose the total light into a base of 12 simple stellar population components represented by star clusters spanning the @xmath41 to @xmath42 yr age range and different metallicities ( schmidt 1991 ; bica 1988 ) , plus a @xmath43 power - law to represent fc1 and any other agn component .
this is done by means of a population synthesis analysis similar to that employed by schmitt ( 1999 ) , but performed with the probabilistic formalism described in cid fernandes ( 2001 ) , extended to include the power - law component .
note that this method is analogous , but not technically equivalent to a full spectral decomposition as only a few @xmath19 s and colors are actually synthesized . as input data we use the @xmath19 s of caii
k , cn and g - band , plus the @xmath44 and @xmath45 colors .
the errors on these quantities were set to 0.5 for @xmath21 and @xmath23 , 1 for @xmath22 and 0.05 for the colors ; these errors are consistent with the quality of the spectra , and in any case they do not affect our conclusions . the output is a 13-d population vector @xmath46 which contains the expected values of the fractional contribution of each component to the total light at a reference normalization wavelength . in
what follows we shall use 4861 as our reference wavelength unless noted otherwise .
uncertainties in @xmath46 are also provided by the code .
it is clear from the outset that with so little input information we will not be able to recover all 13 components of @xmath46 accurately , but , as discussed by cid fernandes ( 2001 ) , we may rightfully hope to achieve a coarse yet useful description of the population mixture by grouping similar elements of the population vector . in what follows
we shall make use of only three grouped @xmath46 components : @xmath47 , made up from the sum of all base components with age @xmath48 gyr , @xmath49 , corresponding to the `` post - starburst '' @xmath50 yr intermediate age bin , and , most importantly for our purposes , @xmath51 , containing the total contribution of @xmath52 yr stars plus the power - law component .
these quantities are linked by the @xmath53 normalization constraint , so in practice our method characterizes the spectral mixture in terms of _ 2 parameters_. this bi - parametric description of the data is in many ways analogous to a principal component analysis ( sodr & cuevas 1997 ; rodrigues - lacerda 2001 ) , except that by construction we have _ a priori _ knowledge of the physical meaning of our components .
note that _ we do not attempt to disentangle the starburst and power - law components _ explicitly ; these are merged in @xmath54 , where @xmath55 and @xmath56 stand for the light fractions due to young ( @xmath52 yr ) stars and the power - law respectively . as shown below ,
both the fc strength and the emission line properties give us strong hints as to which of these two components dominates @xmath51 , thus helping to break this degeneracy .
lrrr mrk 477 & 10.5 @xmath57 3.4 & 2.8 @xmath57 2.0 & 86.6 @xmath57 4.2 mrk 463e & 27.7 @xmath57 4.3 & 3.8 @xmath57 2.5 & 68.5 @xmath57 4.0 ngc 5135 & 27.2 @xmath57 7.3 & 26.9 @xmath57 6.2 & 45.9 @xmath57 5.8 ngc 7130 & 36.3 @xmath57 8.0 & 23.6 @xmath57 6.1 & 40.1 @xmath57 5.9 mrk 533 & 40.9 @xmath57 7.7 & 25.3 @xmath57 6.1 & 33.8 @xmath57 5.5 mrk 1066 & 51.5 @xmath57 6.8 & 16.3 @xmath57 5.4 & 32.3 @xmath57 4.9 ngc 7582 & 35.6 @xmath57 7.8 & 33.8 @xmath57 6.5 & 30.6 @xmath57 5.7 mrk 1 & 63.1 @xmath57 5.6 & 7.1 @xmath57 3.6 & 29.8 @xmath57 4.0 mrk 1073 & 55.3 @xmath57 6.2 & 15.6 @xmath57 5.1 & 29.1 @xmath57 4.6 mrk 1210 & 65.1 @xmath57 4.7 & 5.9 @xmath57 3.1 & 29.0 @xmath57 3.5 ngc 1068 & 63.2 @xmath57 5.5 & 11.1 @xmath57 4.3 & 25.7 @xmath57 4.0 mrk 3 & 77.8 @xmath57 2.7 & 1.1 @xmath57 0.8 & 21.1 @xmath57 2.0 ngc 7212 & 77.3 @xmath57 3.6 & 2.0 @xmath57 1.4 & 20.8 @xmath57 2.6 mrk 273 & 47.3 @xmath57 7.1 & 34.8 @xmath57 6.2 & 17.9 @xmath57 4.7 mrk 34 & 78.1 @xmath57 4.6 & 4.1 @xmath57 2.5 & 17.7 @xmath57 3.1 ic 3639 & 69.9 @xmath57 6.8 & 14.4 @xmath57 5.0 & 15.7 @xmath57 4.1 mrk 348 & 87.2 @xmath57 4.4 & 3.2 @xmath57 2.0 & 9.7 @xmath57 2.6 ngc 2110 & 87.8 @xmath57 3.4 & 2.5 @xmath57 1.6 & 9.6 @xmath57 2.0 ic 1816 & 84.5 @xmath57 4.8 & 7.2 @xmath57 3.3 & 8.3 @xmath57 2.6 ngc 5643 & 73.2 @xmath57 7.2 & 19.5 @xmath57 5.3 & 7.3 @xmath57 3.2 mrk 78 & 72.0 @xmath57 6.5 & 21.2 @xmath57 4.9 & 6.8 @xmath57 2.9 ngc 3081 & 90.1 @xmath57 4.3 & 3.4 @xmath57 2.0 & 6.4 @xmath57 2.1 mrk 573 & 90.9 @xmath57 4.3 & 3.4 @xmath57 2.0 & 5.6 @xmath57 1.9 ngc 5929 & 91.7 @xmath57 3.7 & 2.7 @xmath57 1.7 & 5.6 @xmath57 1.9 ngc 6890 & 93.1 @xmath57 4.5 & 1.9 @xmath57 1.4 & 5.0 @xmath57 2.2 ngc 1386 & 89.9 @xmath57 5.4 & 5.1 @xmath57 2.7 & 5.0 @xmath57 2.1 mrk 607 & 90.4 @xmath57 4.4 & 4.7 @xmath57 2.4 & 4.9 @xmath57 1.8 eso 362-g8 & 56.6 @xmath57 5.9 & 38.5 @xmath57 4.9 & 4.9 @xmath57 2.4 cgcg 420 - 015 & 91.7 @xmath57 5.0 & 4.1 @xmath57 2.3 & 4.2 @xmath57 1.8 mcg -05 - 27 - 013 & 93.9 @xmath57 3.7 & 1.9 @xmath57 1.3 & 4.2 @xmath57 1.6 iras 11215 - 2806 & 90.0 @xmath57 5.7 & 5.9 @xmath57 3.0 & 4.0 @xmath57 1.9 eso 417-g6 & 92.7 @xmath57 4.6 & 4.0 @xmath57 2.2 & 3.3 @xmath57 1.5 fairall 316 & 96.6 @xmath57 3.8 & 1.5 @xmath57 1.0 & 1.9 @xmath57 0.9 ngc 6300 & 96.5 @xmath57 4.6 & 1.7 @xmath57 1.2 & 1.8 @xmath57 1.0 ngc 1358 & 97.6 @xmath57 3.8 & 1.2 @xmath57 0.8 & 1.2 @xmath57 0.7 [ tab : synthesis ] the results of this analysis are listed in table [ tab : synthesis ] in order of decreasing @xmath51 .
a first noticeable result in this table is that composites are concentrated towards the top of the list : all 10 @xmath58 sources are confirmed seyfert 2/starburst composites , and 12 of the 15 composites are among the 16 sources with @xmath59 at 4861 .
the identification of @xmath51 with @xmath55 ( i.e. , @xmath60 ) is thus safe for these objects .
an even more remarkable result is that _ all _ 15 composites but _ only one _ `` pure '' seyfert 2 ( ngc 1068 ) have @xmath61 .
this is illustrated in fig .
[ fig : fig_xfc_x_xint]a , where we condense the results of the synthesis in the @xmath62 plane , with dotted lines indicating lines of constant @xmath63 . _
the diagram segregates composite from `` pure '' systems very effectively _ , and it does so in agreement with what one would expect : composites have larger proportions of intermediate age and young stars , the latter being included in @xmath51 . as a check on our method ,
we have applied it to the spectra in the comparisons samples , which , for consistency , were processed in exactly the same way as our seyfert 2 s .
if our method works , we would expect _ bona fide _ starburst galaxies to be located within the @xmath61 zone , probably more towards even lower values of @xmath47 , overlapping with our more extreme composites .
this is confirmed in fig .
[ fig : fig_xfc_x_xint]b , which shows the synthesis results for the starburst , merger and nlagn samples .
there is not a single starburst galaxy ( filled triangles ) for which @xmath64 , whereas most of the nlagn are located there .
furthermore , most of the nlagn which intrude into the starburst region are known starburst / seyfert 2 composites ! the four empty triangles in the @xmath65 zone , for instance , are mrk 477 , ngc 5135 , ngc 7130 ( all also in our seyfert 2 sample ) and ngc 7496 , whose composite nature is discussed in vron , gonalves & vron - cetty ( 1997 ) .
the @xmath66 zone also contains known composites , both from our sample ( ic 3639 and ngc 7582 ) , and from independent work , such as ngc 4569 ( keel 1996 ; maoz 1998 ) , ngc 1672 ( vron 1997 ) and ngc 6221 ( levenson 2001b ) .
as expected , most of the sources in the merger sample ( squares in fig .
[ fig : fig_xfc_x_xint]b ) behave like starbursts in our stellar population diagram .
note that our synthesis indicates a strong contribution of @xmath67 yr populations for most sources in this sample .
this is consistent with the results of lk95 , who find a high incidence of systems in a post - starburst phase based on the detection of the balmer absorption series .
those mergers which are dominated by old stars ( @xmath68 ) are predominantly liners ( e.g. , ngc 942 , ngc 3656 and 3c293 ) .
it is thus clear that , despite its limitations , our bi - parametric description of the data provides a very efficient empirical diagnostic of compositeness in seyfert 2 s . since our operational definition of composites
is based on the detectability of starburst features , which is certainly facilitated when they make a strong contribution to the continuum , the result that sources with large @xmath51 and/or @xmath49 are mostly composites is somewhat redundant .
the importance of this result is twofold .
first , it shows a striking empirical similarity between the _ bona fide _ starburst galaxies ( whose optical / near - uv continuum is definitely produced by young and intermediate - age stars ) and the starburst / seyfert 2 composites .
second , it demonstrates an excellent _ consistency _ between the two techniques , which are based on different observables .
the synthesis process did _ not _ use any of the information used in the original identification of starbursts in these composites , namely , far - uv stellar wind features , high order balmer lines , and/or the wr bump , all of which are much more subtle or hard to obtain than the observables used as input for the synthesis code .
the separation between composites and `` pure '' seyfert 2 s in fig .
[ fig : fig_xfc_x_xint]a strongly suggests that , as anticipated in [ sec : compositesandpureseyfert2s ] , the composite/``pure '' classification is mostly driven by a _ contrast _ effect : systems with circum - nuclear starbursts residing in galaxies where the old stellar population dominates the spectrum would simply not be recognized as composite systems .
this contrast effect is nicely quantified by the value of @xmath47 , which can therefore be read as a measure of the difficulty to detect any non - trivial spectral component and to interpret its origin .
[ fig : fig_xfc_x_xint ] shows that our current techniques establish a threshold of @xmath69 , beyond which we are not able to recognize circum - nuclear starbursts .
for instance , if we scale down the circum - nuclear starburst in ic 3639 by a factor @xmath70 , it would move from @xmath71 to more than 82% , where , judging by fig .
[ fig : fig_xfc_x_xint ] , we would classify it as a `` pure '' seyfert 2 .
it is therefore likely that at least some `` pure '' seyfert 2 s are just that : composites with weak starbursts .
the presence of strong hi emission lines introduces another contrast effect , since they dilute the high order balmer absorption lines , thus hindering the use of this diagnostic of starbursts .
therefore _ the combination of strong emission lines and a large @xmath47 is highly unfavorable to the detection of circum - nuclear - starbursts_. we will return to this issue in [ sec : difficulty_to_detect_starbursts ] , after we analyze the emission line data . besides the contrast between the starburst and the underlying old stellar population , _ evolution _ is the main property defining the location of a source in fig .
[ fig : fig_xfc_x_xint ] .
one can easily imagine an evolutionary sequence which starts at some large value of @xmath51 in the bottom right part of the diagram , with massive young starts dominating the spectrum and no intermediate age population , and gradually moves towards large @xmath72 ( i.e. , to the top left of the diagram ) on @xmath67 yr time scales .
the exact shape of evolutionary tracks in our @xmath62 diagram is dictated by the detailed star - formation history , but for the purposes of the qualitative discussion below one can imagine that such tracks broadly follow the dotted lines traced in fig .
[ fig : fig_xfc_x_xint ] up to a few hundred myr , and then collapse to the origin , after the starburst ends and its stars move into our old population bin . here too
we find a remarkable agreement between the present synthesis analysis and the results of our previous work on the composites in the seyfert 2 sample .
the four composites with larger @xmath73 are mrk 477 , mrk 463e , mrk 1210 and mrk 1 ( going upwards in fig .
[ fig : fig_xfc_x_xint]a ) .
these are precisely the 4 systems where we have identified wr stars ( heckman 1997 ; storchi - bergmann 1998 ; gd01 ) , signposts of very recent star - formation .
they therefore rank as the youngest circum - nuclear starbursts among our composites , in agreement with their location in fig .
[ fig : fig_xfc_x_xint]a . at the other extreme ,
the three composites with @xmath74 are ngc 5643 , mrk 78 and eso 362-g8 , all of which present spectroscopic signatures of a dominant ` post - starburst ' population ( schmitt 1999 ; sb00 ; gd01 ) , and rank as our oldest composites , in agreement with the large @xmath49 but small @xmath51 we obtain . given this agreement at the young and old ends of a starburst evolutionary sequence
, we would expect that systems located at @xmath73 values in between these extremes present a mixture of young and intermediate age stars .
this is indeed the case . the cluster of 4 composites towards the top right in fig .
[ fig : fig_xfc_x_xint]a , for instance , contains ngc 5135 , ngc 7130 , ngc 7582 and mrk 533 , all of which reveal spectroscopic signatures of this mixture ( sb00 ; gd01 ) , such as pronounced high order balmer absorption lines simultaneous with far - uv stellar wind lines .
what makes this agreement remarkable is the fact that while the characterization of the starburst component in our earlier work involved a detailed spectroscopic analysis , the synthesis performed here used just three @xmath19 s of metallic absorption bands plus a couple of near - uv colors .
our quicker and cheaper method is therefore not only able to recognize composite systems from a handful of easily measured observables , but it allows us to go one step further to provide a rough description of evolutionary state of the starburst component .
we will return to the issue of evolution in [ sec : discussion ] . before that
, we have to deal with the contribution of the non - stellar fc in @xmath51 .
this issue was deliberately omitted from the discussion above because we do _ not _ have a recipe to separate the stellar and non - stellar parts of the fc , neither do we know how or whether the non - stellar fc evolves on time - scales comparable to the starburst lifetime .
these difficulties limit considerations about evolution to composites alone , for which we know that the starburst component is the dominant contributor to the fc .
our inability to disentangle the starburst and non - stellar components of the fc is currently the major obstacle to a full characterization of the starburst - agn connection in seyfert 2 s .
unfortunately , these are also the most relevant spectral components in seyfert 2 s . the starburst portion of @xmath51 traces the ongoing star - formation , being therefore associated with young massive stars which may have a significant impact on the ionization of the gas in the circum - nuclear environment . on the other hand ,
the non - stellar component ( which is presumably mainly scattered light , i.e. , fc1 ) is the only continuum feature associated with the agn in these galaxies . for these reasons we concentrate on the fc component throughout most of the rest of this paper .
whereas little can be said about the starburst and agn shares of @xmath51 in `` pure '' seyfert 2 s , composite systems have their fc dominated by the starburst ( gd01 ; sb00 ) , and hence provide a useful guide as to what can be deduced from the fc strength alone .
table [ tab : synthesis ] shows that strong fc sources ( @xmath75 ) are likely to be composite systems . within the context of the unified model
, it is not surprising to find that a starburst component dominates over fc1 whenever the total fc is strong , since scattered light can not exceed a fraction of @xmath76 , otherwise reflected broad lines should become easily discernible in the direct spectrum and the galaxy would no longer be classified as a type 2 seyfert . for a typical @xmath19 of @xmath77 for broad h@xmath3 in seyfert 1 s
( goodrich 1989 ; binette , fosbury & parker 1993 ) , more than @xmath78 scattered light would imply a broad h@xmath3 stronger than 30 in the direct total spectrum .
cid fernandes & terlevich ( 1995 ) argue that broad lines should be seen even for smaller @xmath79 , but we figure 30% is a reasonable limit in practice because ngc 1068 has @xmath80 , and broad lines have not been conclusively detected there without the hindsight of spectropolarimetry to guide the eye ( miller & goodrich 1990 ; tran 1995b ; malkan & filippenko 1983 ) . in support of this conclusion , we note that several cases of strong fc sources reported in the literature turned out to harbor starbursts not accounted for in the classical starlight template decomposition .
for instance , miller & goodrich ( 1990 ) find @xmath81 in mrk 463e , while gd01 suggest it actually contains a powerful starburst in the wr phase .
the same can be said about mrk 477 , for which tran ( 1995a , b ) finds @xmath82 at 5500 , and mrk 1066 , for which miller & goodrich ( 1990 ) find a 72% fc at 5300 . even in cygnus
a , whose blue fc has remained a mystery for more than a decade ( e.g. , goodrich & miller 1989 ; tadhunter , scarrott & rolph 1990 ) , there now seems to be evidence for young stars ( fosbury 1999 ) .
indeed , inspection of the keck spectrum in ogle ( 1997 ) even indicates the presence of a possible wr bump ! this is consistent with our conclusion , since estimates of the fc strength for cygnus a place it above the 30% level ( osterbrock 1983 ) .
this high `` success''-rate suggests that one can _ predict _ the existence of circum - nuclear starbursts based only on the fc strength . despite the non uniformity of fc estimation methods , one can be fairly confident that a search for sources with @xmath83 reported in the literature should identify other composite systems .
kay ( 1994 ) , for instance , reports the following seyfert 2 s with an fc stronger than 30% at 4400 : mrk 266sw ( 30% ) , mrk 1388 ( 45% ) , ngc 591 ( 46% ) , ngc 1410 ( 38% ) , ngc 1685 ( 38% ) , ngc 4922b ( 53% ) , ngc 7319 ( 35% ) , ngc 7682 ( 30% ) .
koski ( 1978 ) finds a 34% fc at 4861 for 3c184.1 and 35% for mrk 198 .
evidence for young and intermediate age stars in the spectra of mrk266sw was reported by wang ( 1997 ) . according to gonalves , vron - cetty & vron ( 1999 )
, this galaxy also has a composite emission line spectrum , with emission line ratios intermediate between those of seyferts and starbursts .
the same description is given for the interacting galaxy ngc 4922b by alonso - herrero ( 1999 ) .
we could not find information regarding the compositeness or otherwise of the remaining galaxies .
our prediction is that most of them should present detectable signatures of circum - nuclear starburst activity .
while all @xmath84 sources are composites , _ the converse is not true_. for ic 3639 , a galaxy whose composite nature has been conclusively established by both optical and uv spectroscopy , we find a 16% fc .
the range between 15 and 30% also contains 4 `` pure '' seyfert 2 nuclei : ngc 1068 , mrk 3 , ngc 7212 and mrk 34 .
ngc 1068 is sometimes listed as an example of a starburst - agn link because of its bright ring of hii regions at @xmath85 kpc from the nucleus , which contributes roughly half of the bolometric ( ir ) luminosity of this prototype seyfert 2 ( telesco 1984 ; lester 1987 ) .
if observed from further away , the ring would dominate the optical spectrum , imprinting signatures which would cast distant ngc 1068 s in the composite category ( heckman 1995 ; colina 1997 ) , but , as argued in [ sec : sample ] , this is not the effect behind the starburst features identified in our composites ( fig .
[ fig : aperture_x_wk ] ) .
the spectrum of ngc 1068 analyzed here pertains to a much smaller region , corresponding to @xmath86 pc .
the only indications of star - formation within this central region are the strong caii triplet at @xmath87 ( terlevich , daz & terlevich 1990 ) and the small @xmath88 ratio at 1.6@xmath24 m ( oliva 1995 , 1999 ) , both of which point to an intermediate - age population , qualitatively consistent with the @xmath89 found in our synthesis
. there does not seem to be substantial ongoing star - formation associated with this `` post - starburst '' population , since neither far - uv ( caganof 1991 ) nor optical ( gd01 ; miller & antonucci 1983 ) spectroscopy reveal signs of young stars close to the nucleus .
furthermore , from the work of antonucci & miller ( 1985 ) , miller & goodrich ( 1990 ) and tran ( 1995c ) we know that ngc 1068 does _ not _ suffer from the `` fc2-syndrome '' ( lower polarization in the continuum than in the broad lines ) , so most , if not all , of its nuclear fc is indeed fc1 . since our @xmath51 estimate is in fair agreement with previous determinations ( e.g. , the 22 % at 4600 found by miller & antonucci 1983 and the 16% at 5500 derived by tran 1995 ) , there is little doubt that @xmath90 in the central region of ngc 1068 . we therefore keep ngc 1068 in the `` pure '' seyfert 2 category .
the nature of the fc is not so clear for the 3 other moderate fc `` pure '' seyfert 2 s in our sample : mrk 3 , ngc 7212 and mrk 34 . in their detailed analysis of the same data used here
, gd01 find that the dilution of the metal absorption lines in these galaxies is better modeled by an elliptical galaxy plus power - law spectral decomposition than using an off - nuclear template , which , together with the lack of starburst features , lead them to favor a scattered light origin for their fc .
the extended blue continuum aligned with the radio axis and ionization cone in mrk 3 ( pogge & de robertis 1993 ) supports this idea .
kotilainen & ward ( 1997 ) find a similar feature in ngc 7212 , though not aligned along the [ oiii ] emission .
these indications of a fc1 component are confirmed by the spectropolarimetry observations of tran ( 1995a , b , c ) , which revealed their hidden seyfert 1 nuclei .
however , tran also finds that mrk 3 and ngc 7212 suffer from acute `` fc2-itus '' ! according to his analysis , only 4% ( mrk 3 ) and 2% ( ngc 7212 ) of the flux at 5500 is attributable to fc1 ; the rest of their fc is ` fc2 ' !
the total fc = fc1 + fc2 fractions at 5500 found by tran are 12% for mrk 3 and 17% for ngc 7212 , which are in very good agreement with our estimates of @xmath51 when translated to the same wavelength .
therefore , from the point of view of their polarization spectra , most of their fc still have to be explained .
below @xmath91 we therefore enter a `` gray zone '' : the power - law / starburst degeneracy sets in and the identity of the fc becomes ambiguous .
extra information ( spectropolarimetry , uv spectroscopy , imaging ) may help disentangling these two components in a few sources , but the situation becomes even fuzzier for weak fc sources ( @xmath92 ) .
the case of mrk 348 is illustrative in this respect . using an elliptical galaxy template , tran ( 1995a , b )
finds a 27% fc at 5500 , which , when combined with his spectropolarimetry data , propagates to the conclusion that 22% of the total light is associated with something else , ie .
storchi - bergmann ( 1998 ) , on the other hand , find a much weaker fc by using an off - nuclear spectrum as a starlight template , a conclusion corroborated by gd01 on the basis of independent observations .
in fact , they find that the nuclear spectrum is well matched with no fc at all , but , guided by the results of tran , they favor a model in which 5% of the light comes from the fc1 component detected via spectropolarimetry .
these inconsistencies undoubtedly emerge due to the intrinsic weakness of the fc , which , combined with the dispersion of stellar populations in active galaxies ( cid fernandes 1998 ; schmitt 1999 ; boisson 2000 ; gd01 ) , boosts the differences between different fc estimation methods . whereas in mrk 348
the combination of spectropolarimetry and long - slit spectra allowed considerations about the nature of its fc , we do not have such information for most sources below mrk 348 in table [ tab : synthesis ] , and even if such data existed , it is clear that any attempt to perform a detailed spectral decomposition would be highly uncertain , since it is simply unrealistic to expect accuracies better than @xmath93% in any @xmath51 determination method . from the point of view of this paper , we consider all such sources ambiguous .
we hope to shed light on the nature of these moderate and weak fc seyfert 2 s by considering information other than that contained in the optical continuum and absorption lines ( [ sec : emission_lines ] ) . as in any other study dealing with optical spectra of seyfert 2 s
, the fc strength plays a major role throughout this paper .
the results above already indicate that , regardless of the ambiguities involved in the interpretation of @xmath51 , our method was able to recover a meaningful component of the near uv optical continuum , a conclusion which will become stronger in the analysis that follows .
this in itself is an important result , since we based our fc estimation on just a handful of easily measurable quantities .
though the synthesis process is rather elaborate , we may use its results to derive an _ a posteriori _ calibration of @xmath51 in terms of the observables it is based upon .
in fact , as shown in fig . [
fig : xfc_callibration ] , we find that the expression @xmath94 where @xmath21 is in , recovers @xmath51 within @xmath95 for all galaxies . since the measurements of both @xmath21 and the @xmath96 color do not require large s / n spectra , this calibration can be applied to data sets not meeting the s / n @xmath97 standard of our sample ( which was necessary to identify weak starburst features ) .
this provides a much more straight - forward and well defined way to estimate @xmath51 .
although old stars are of no direct interest in the starburst - agn connection , we have seen that they play a central role in defining the detectability of circum - nuclear starbursts , so it is useful to have an equation analogous to ( [ eq : xfc_calibration ] ) for @xmath47 .
we find that @xmath98 reproduces the synthetic @xmath47 to better than @xmath99 for all galaxies .
equations ( [ eq : xfc_calibration ] ) and ( [ eq : xold_calibration ] ) also do a good job for the merger , nlagn and starburst samples .
this is illustrated in fig .
[ fig : xfc_callibration]b for the galaxies in the starburst sample .
the difference , of course , is that for these _ bona fide _ starbursts the fc is undoubtedly produced by young stars .
a corollary of @xmath51 and @xmath47 being so closely related to @xmath21 is that this quantity can be used as a first order measure of the stellar population mix and the spectral dilution of metallic features of old stars caused by young stars , a power - law or a combination of both .
indeed , in several of the plots below @xmath21 is used with this function , since , for most practical purposes , @xmath21 and @xmath51 are equivalent ( fig .
[ fig : xfc_callibration ] ) .
circum - nuclear starbursts in seyfert 2 s act as a second source of ionizing photons besides the hidden agn , and as such are bound to contribute at some level to their emission line spectrum . in this section
we examine our sample in search of signs of this contribution , analyzing the equivalent widths of strong emission lines ( [ sec : ws ] ) , emission line ratios ( [ sec : he2_and_o3 ] ) and line profiles ( [ sec : lineprofiles ] ) .
an investigation of line - fc relations is also presented ( [ sec : line_fc_relation ] and [ sec : line_x_fc_interpretation ] ) .
when referring to emission line equivalent widths measured with respect to the fc we will use the notation @xmath100 to distinguish them from the observed ones , denoted by @xmath101 .
the effect of a circum - nuclear starburst on the @xmath102 of seyfert 2 s is easily predicted by considering the inverse situation : a pure starburst to which we add a seyfert 2 .
the latter will surely add more to the @xmath103 flux than to its underlying continuum , since the non - stellar fc from the active nucleus is only seen periscopicaly and hence strongly suppressed in the scattering process , whereas photons from the narrow line region ( nlr ) do not suffer such suppression because they are not obscured by the pc scale dusty torus which surrounds the nucleus .
the net effect of this superposition is that seyfert 2 s should have a larger ratio of line per fc photons than in pure starbursts .
this expectation is totally confirmed in the left panels of fig .
[ fig : w_hist ] , where we compare the distribution of @xmath102 for our seyfert 2 sample with that of the 35 starburst galaxies in the starburst sample .
_ these histograms reveal a clear offset in @xmath102 between seyfert 2 s and stellar powered emission line sources_. the median @xmath102 is a factor of 5 larger in the seyfert 2 sample than in pure starbursts ( 129 compared to 26 ) .
given their dual nature , we intuitively expect composite systems to exhibit @xmath102 in between those of `` pure '' seyfert 2 s and starbursts . this expectation is also confirmed .
the composite seyfert 2/starburst systems in our sample , marked by the filled region in fig .
[ fig : w_hist]a , tend to populate the low end of the @xmath102 distribution in seyfert 2 s , overlapping with the high @xmath102 starbursts ( fig . [ fig : w_hist]b ) . the same considerations , of course ,
apply to other emission lines .
in fact , given that [ oiii]/h@xmath3 is much larger in seyferts than in stellar powered systems , we expect an even clearer separation between these two types of objects in terms of @xmath104}$ ] .
this is confirmed in figs .
[ fig : w_hist]c and d. the difference between median @xmath105 s , which was a factor of 5 for h@xmath3 , is now 32-fold , with 44 for starburst galaxies and 1431 in the seyfert 2 sample .
as for h@xmath3 , composites are skewed towards low @xmath104}$ ] s ( fig .
[ fig : w_hist]c ) .
the same effects are identified in [ oii ] ( figs .
[ fig : w_hist]e and f ) .
the large offset in @xmath105 s between seyfert 2 s and starburst galaxies is not an artifact of aperture differences between the two samples .
one does not expect drastic changes in the line per continuum photon ratio between nuclear and off - nuclear star - forming regions , so the distribution of @xmath105 s for starburst galaxies should remain roughly the same for small and large apertures .
this is demonstrated by the shaded portions of the histograms in fig .
[ fig : w_hist]b , d and f , which mark starburst galaxies observed through physical apertures @xmath106 kpc .
aperture effects become important for seyfert 2 s , since there the contribution of off - nuclear hii regions affects the line / fc proportion , diluting the higher @xmath105 of the nucleus .
this effect is detected for the objects in common between the seyfert 2 and nlagn samples .
mrk 477 , for instance , has @xmath107 for the observations of sbmq95 , whereas with our nuclear spectra we obtain @xmath108 .
the same effect occurs comparing the nuclear with the `` whole aperture '' spectra of lk95 , which integrate over the entire galaxy . since in
[ sec : synthesis ] we found that composites tend to have a strong fc , we may expect them to be well separated from `` pure '' seyfert 2 s in diagrams involving @xmath51 and these emission line @xmath105 s .
this is illustrated in figs .
[ fig : ws_x_wk]a c , where we plot the @xmath105 s of h@xmath3 , [ oiii ] and [ oii ] against @xmath21 , here representing the fc strength . note that all composites are located to the left of the @xmath109 line indicated in the plots .
( a slightly more stringent constraint of @xmath110 can be adopted by restraining our definition of composites to systems which exhibit signs of young stars , thus excluding the post - starbursts ngc 5643 , mrk 78 and eso 362-g8 . )
the @xmath105 s by themselves do not segregate composite from `` pure '' seyfert 2 s as efficiently as @xmath21 , but sources with the weakest emission lines are predominantly composites , so the vertical axis provides some extra diagnostic power .
the `` pure '' seyfert 2 at @xmath111 , trespassing the @xmath1 zone of composites , is ngc 1068 , which has @xmath112 } = 68 $ ] , @xmath104 } = 1334 $ ] and @xmath113 .
this prototype seyfert 2 , which shows no signs of recent star - formation close to the nucleus , is located either among or close to the composites in all diagrams presented in this paper .
we will return to this issue in [ sec : discussion ] .
the panels on the right side of fig .
[ fig : ws_x_wk ] show the results for sources in the starburst and merger samples for comparison .
once again we must draw attention to contrast effects which may lead us to classify seyfert 2 s with weak circum - nuclear starbursts as `` pure '' systems .
part of the difficulty in identifying starburst features in seyfert 2 nuclei with strong emission - lines comes from the fact that as @xmath102 increases the high order hi balmer absorption lines of massive stars become increasingly filled up by emission . in the absence of far - uv spectra ( available for very few sources ) and the short - lived wr bump
, this effect effectively prevents us from seeing the starburst in the near - uv to optical range .
this contrast effect comes on top of the difficulty in identifying circum - nuclear starbursts in systems dominated by the old stellar population ( [ sec : constrast_and_evolution ] )
. the combination of these two effects may be at least partly responsible for the horizontal and vertical separations between composites and `` pure '' seyfert 2 s seen in fig .
[ fig : ws_x_wk]a c .
the location of a seyfert 2 in the @xmath21-@xmath102 plane can therefore be read as an empirical measure of the _ `` degree of difficulty '' _ in the recognition of circum - nuclear starbursts in seyfert 2 s : sources at the bottom left are easily identified as composites , whereas as one progresses to the top right it becomes increasingly difficult to discern the starburst features .
the usefulness of this concept is illustrated by the location of mrk 3 , mrk 34 and ngc 7212 in figs .
[ fig : ws_x_wk]a c , all are around the @xmath109 ``dividing line '' but with @xmath102 and @xmath104}$ ] larger than for composites .
it is no accident , therefore , that the nature of their moderately strong fc remains unclear ( [ sec : moderate_and_weak_fc_sources ] ) .
it is important to emphasize that the offset in emission line @xmath19 s between seyfert 2 s and starbursts only appears when these are measured with respect to the fc , as hinted in fig .
[ fig : ws_x_wk ] .
our estimate of @xmath51 , which takes into consideration continuum colors and metal absorption lines through our semi - empirical synthesis analysis , is thus the key behind this significant off - set .
the role of the synthesis process can be better appreciated in fig .
[ fig : hb_x_continuum ] . in fig .
[ fig : hb_x_continuum]a we plot @xmath114 against the total observed continuum flux at 4861 .
what is seen is a scatter diagram , which is not surprising since the continuum in most sources is dominated by old stars and thus has little relation to the line emission .
this is confirmed by the fact that the scatter increases even further if the continuum fluxes are multiplied by @xmath47 ( fig .
[ fig : hb_x_continuum]b ) , thus isolating the flux due to @xmath48 gyr stars . the line versus continuum plot changes dramatically when we isolate the fc contribution ! using @xmath115 in the abscissa ( fig .
[ fig : hb_x_continuum]c ) has the remarkable effect of uncovering an underlying order immersed in the scatter plot on the left .
it is therefore clear that _ our fc determination method isolated a component of the optical continuum which is directly linked to the line emission _ , lending further credibility to the inferred @xmath51 values .
the same result is obtained for the sources in our comparison samples , i.e. , @xmath51 organizes the data along a @xmath114-@xmath116 correlation which is not evident in the @xmath114-@xmath117 plane because of the contribution of the old stellar population .
lrrrrrrrrrrrr @xmath114 & @xmath118 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath124 & @xmath127 & @xmath128 @xmath129 & @xmath130 & @xmath131 & @xmath132 & @xmath133 & @xmath134 & @xmath119 & @xmath135 & @xmath136 & @xmath137 & @xmath125 & @xmath138 & @xmath139 @xmath140}$ ] & @xmath126 & @xmath122 & @xmath138 & @xmath141 & @xmath142 & @xmath143 & @xmath135 & @xmath143 & @xmath126 & @xmath124 & @xmath144 & @xmath145 @xmath146}$ ] & @xmath138 & @xmath147 & @xmath148 & @xmath128 & @xmath124 & @xmath149 & @xmath122 & @xmath130 & @xmath150 & @xmath125 & @xmath145 & @xmath145 @xmath151 & @xmath152 & @xmath119 & @xmath145 & @xmath148 & @xmath131 & @xmath124 & @xmath152 & @xmath122 & @xmath126 & @xmath124 & @xmath153 & @xmath132 [ tab : correlations ] possible interpretations of the h@xmath3-fc relation identified in fig .
[ fig : hb_x_continuum]c will be discussed in [ sec : line_x_fc_interpretation ] , after we gather other related results . to substantiate the discussion it is useful to compare it to the equivalent relation in starburst systems and to investigate line - fc relations for other transitions .
this is done in figs .
[ fig : hb_and_o2_x_fc ] and [ fig : he2_and_o3_x_fc ] , which contain luminosity - luminosity versions of fig .
[ fig : hb_x_continuum]c for the seyfert 2 , starburst and nlagn samples and for h@xmath3 , [ oii ] , [ oiii ] and heii . lines of constant @xmath154 , 100 and 1000 run diagonally across these plots to facilitate their intercomparison . also to aid
the discussion , table [ tab : correlations ] summarizes the results of a spearman rank correlation analysis for flux - flux line - continuum relations .
the values listed are the probabilities of chance correlation ( @xmath155 ) , so small values indicate significant correlations .
significant line - fc correlations are also obtained for [ oii ] , [ oiii ] and heii in the seyfert 2 sample ( figs .
[ fig : hb_and_o2_x_fc]c , [ fig : he2_and_o3_x_fc]a and [ fig : he2_and_o3_x_fc]c , table [ tab : correlations ] ) .
note , however , that the scatter in the [ oiii ] and heii plots is visibly larger than for h@xmath3-fc ( fig .
[ fig : hb_and_o2_x_fc]a )
. it is also clear that a substantial part of the scatter in these plots is induced by the offset location of the composites towards smaller @xmath105 s .
statistical confirmation of this effect is provided in table [ tab : correlations ] , where these correlations are examined for the whole sample and for subsets containing only `` pure '' or composite systems .
starburst galaxies also follow a tight h@xmath3-fc relation , with @xmath156100 ( filled triangles in figs . [
fig : hb_and_o2_x_fc]b ) .
this is a long known ( terlevich 1991 ) and well understood behavior of stellar powered systems , for which @xmath116 actually represents the contribution of young stars and hence traces the ionizing flux .
current models ( e.g. , leitherer 1999 ) predict @xmath157 between 400 and 10 for constant star - formation , different imf s and ages up to @xmath50 yr .
these are broadly consistent with the values we obtain , but a detailed comparison with such evolutionary models will have to await a cross calibration of our semi - empirical synthesis decomposition in terms of the theoretical spectra . as for the seyfert 2 sample ,
the second best line - fc relation for starburst galaxies is that with the [ oii ] line ( fig .
[ fig : hb_and_o2_x_fc]d ) .
this too is a well known relation , which has in fact been used as an alternative estimator of star - formation rates in galaxies ( gallagher 1989 ; kennicutt 1992 ; tresse 1999 ; jansen , franx & fabricant 2001 ) .
the location of the nlagn s in these plots ( empty triangles ) reflects the mixed nature of this sample . some of them behave just like `` pure '' seyfert 2 s , most notably ngc 3393 and ngc 5506 , which stand out in figs . [
fig : hb_and_o2_x_fc]b , d and [ fig : he2_and_o3_x_fc]b as the sources with highest @xmath105 s .
most of the other nlagn s are mixed among starburst galaxies in these plots .
this happens due to a combination of the aperture effects discussed in
[ sec : ws ] and the fact that this sample contains several composites and some objects which are more consistent with a liner classification ( e.g. , ngc 1097 , ngc 1433 ) .
it is thus clear that these galaxies do not constitute a well defined comparison sample .
for these reasons , we concentrate on the comparison between the seyfert 2 and starburst sample .
one would like to believe that extinction does not play a relevant role in these line - fc diagrams since both axes correspond to similar wavelengths .
in starburst galaxies , however , it is known that the continuum suffers @xmath158 times less extinction than the emission line region ( calzetti , kinney & storchi - bergmann 1994 ) .
this effect can reshuffle the data points in the plots above , potentially changing the overall appearance and strength of the line - fc correlations .
an estimate of the effects of extinction is presented in fig .
[ fig : hb_x_fc_reddening ] for the h@xmath3-fc relation using the h@xmath25/h@xmath3 values listed in table [ tab : tabelao2 ] . fig .
[ fig : hb_x_fc_reddening]a presents the data with @xmath159 and @xmath160 dereddened by the same amount , whereas in [ fig : hb_x_fc_reddening]b we halved the optical depth to the fc .
two effects are clear when comparing these plots with their uncorrected counterpart ( fig .
[ fig : hb_and_o2_x_fc]a ) : first , with the extinction correction , composites are more segregated from `` pure '' seyfert 2 s in terms of @xmath160 , with median values which differ by an order of magnitude ( median @xmath161 and @xmath162 erg@xmath15s@xmath17 for composites and `` pure '' seyfert 2 s respectively ) .
this preference of composites for luminous systems will appear again when we discuss their far ir luminosities in [ sec : iras ] .
second , a differential line / fc extinction correction only tightens the h@xmath3-fc correlation !
the effect is such that it practically erases the offset in @xmath102 between composite and `` pure '' seyfert 2 s .
the caveats in these results are that the h@xmath25/h@xmath3 values used were not obtained from the same spectra we analyze ( [ sec : sample ] ) and that we do not know whether the @xmath163 result applies to agn . at any rate ,
the larger h@xmath25/h@xmath3 of composites , which is responsible for this regrouping , is consistent with uv ( heckman 1995 ) and x - ray ( levenson , heckman & weaver 2001a ) indications that circum - nuclear starbursts in seyfert 2 s are substantially reddened , and may significantly contribute to the extinction of the agn component in such composite systems .
this `` excess '' of dust is probably associated with the same gas which feeds the starburst activity , but more work will be required to address this issue quantitatively .
the results reported so far show how the presence of circum - nuclear starbursts manifest itself in the fc strength and the equivalent widths of emission lines .
these results _ per se _ do not necessarily imply that the starburst contributes to the ionization of the line emitting gas in seyfert 2 s , since it may act mostly as a source of fc , diluting an agn powered h@xmath3 emission and thus shifting points to the right , but not much upwards in figs .
[ fig : hb_x_continuum][fig : hb_x_fc_reddening ] . one way to establish whether the starburst participates in the ionization of the gas is to investigate the ratio between heii@xmath184686 and h@xmath3 . unlike agn
, starbursts do not produce significant radiation above 54 ev , so they show little heii emission .
if starbursts in agn contribute to h@xmath3 this ratio will tend to be smaller in composites than in `` pure '' seyfert 2 s .
this is confirmed in figs .
[ fig : he2hb]a b , where we plot heii / h@xmath3 against @xmath21 and @xmath47 , both empirical indicators of compositeness ( [ sec : synthesis ] ) .
whereas the whole sample spans the 0.10.8 interval in heii / h@xmath3 , all composites have heii / h@xmath164 . many `` pure '' seyfert 2 s also have heii this weak , but their stronger ca ii k and hence larger @xmath47 clearly separates them from composites .
starburst galaxies would be located in the same horizontal range spanned by composites in fig .
[ fig : he2hb]a b , but down to weaker heii ( a line which is often not detected in starbursts ) .
composites also tend to be less excited in terms of [ oiii]/h@xmath3 , as shown in figs .
[ fig : he2hb]c d .
the effect of the stellar absorption feature underneath the h@xmath3 emission - line was not corrected for in this analysis .
since composites are more likely to have stronger balmer absorption lines than objects dominated by old stars , this correction would only strengthen our conclusion , shifting the composites to even smaller heii / h@xmath3 and [ oiii]/h@xmath3 ratios .
either line ratio can therefore be used as an auxiliary diagnostic of compositeness .
this is illustrated by the dotted lines in fig .
[ fig : he2hb ] , which isolate composites from `` pure '' seyfert 2 s almost completely ! ( again , the odd ball is ngc 1068 . ) this shows a welcome consistency between diagnostics of compositeness based on continuum and stellar features and the expected impact of circum - nuclear starbursts upon the emission line ratios in seyfert 2 s .
note , however , that while all known composites have low excitation , the converse is not true , since many `` pure '' seyfert 2 s lie in the heii / h@xmath164 and [ oiii]/h@xmath165 range defined by composites . in terms of emission line ratios , these galaxies could well have circum - nuclear starbursts , but which do not make it to the composite class due to the contrast effect discussed above .
it is hard to see how to account for these results in a `` pure agn '' scenario .
the shape of the ionizing spectrum , the proportion of matter to ionization bounded clouds and the ionization parameter are just some of the factors which define the excitation level of the nlr , and source - to - source variations of these properties can account for a wide spread in heii / h@xmath3 and [ oiii]/h@xmath3 ( e.g. viegas - aldrovandi 1988 ) . yet
, none of these intrinsic properties explains the `` zone of avoidance '' for strong fc and high excitation in fig .
[ fig : he2hb ] .
if anything , a strong fc would suggest a higher ionization parameter and perhaps a harder ionizing spectrum , which would produce higher excitation , contrary to what is observed .
overall , the simplest interpretation of these data is that part of the emission lines in composites originates from photoionization by a circum - nuclear starburst .
this contribution is strongest in h@xmath3 , since starbursts are inefficient producers of heii and [ oiii ] compared to agn .
it is clear that in a starburst @xmath0 agn mixture of emission lines the starburst component can dominate h@xmath3 without moving the source outside of the agn region in diagnostic diagrams ( e.g. , hill 2001 ; levenson 2001b ) .
for instance , using typical [ oiii]/h@xmath3 ratios of 1020 for seyfert 2 s and 0.31 for metal - rich starbursts , the total ratio still lies within the [ oiii]/h@xmath166 domain of seyferts for as much as 7090% of h@xmath3 powered by the starburst ! therefore our conclusion that photoionization by massive stars provides a substantial part of @xmath160 in composites is not in conflict with their seyfert 2 classification .
in fact , this conclusion was already implicit in the off - set in @xmath102 between composites and starburst galaxies ( fig .
[ fig : w_hist ] ) .
the @xmath102 values in composites cover the range from 30 to 200 , which may be represented by its geometric mean of @xmath3480 .
the corresponding value in starburst galaxies is @xmath3430 , representing the @xmath156100 interval . since we know that the fc in composites is dominated by the starburst component
, these values imply a typical starburst contribution of @xmath167 to h@xmath3 .
these results imply that h@xmath3 is partially powered by the starburst in composite systems . unfortunately , as for @xmath159
, it is not possible to unambiguously separate the starburst and agn shares of @xmath160 without extra information , such as assuming typical line ratios or equivalent widths .
for instance , taking the mean [ oiii]/h@xmath3 of 13 for our `` pure '' seyfert 2 s and 8 for composites , one obtains a starburst contribution to h@xmath3 of 40% ( precisely what we have derived above comparing composites and starburst galaxies in terms of @xmath102 ) .
illustrative as these global estimates are , it is clear that both composite and `` pure '' systems present a _ range _ of properties . for the composites below ngc 5135 ( [ oiii]/h@xmath168 ) in fig .
[ fig : he2hb]c and d , for instance , the same exercise would yield more than 70% of h@xmath3 powered by the starburst .
much therefore remains to be learned from a more detailed object by object analysis .
one prospect for disentangling the starburst and agn contributions to the emission lines is to study their profiles in search of differential kinematical signatures of these two components .
we have thus performed a line profile analysis for the seyfert 2 sample , but only for the kitt peak data , which have higher spectral resolution .
most of the composites have [ oiii ] and h@xmath3 profiles which can each be described in terms of two components .
interestingly , the narrower component often has a lower excitation than the broader one ( table [ tab : profiles ] ) , qualitatively consistent with them being powered by a starburst and agn respectively .
in fact , the excitation of the broad component ( mean [ oiii]/h@xmath169 , standard deviation @xmath170 ) is very similar to the excitation of `` pure '' seyfert 2 s .
however , the excitation of the narrow components ranges between 2.7 and 11.6 .
meanwhile , the excitation in mrk 1066 , mrk 1073 , ngc 5135 , ngc 7130 and ic 3639 is similar to the excitation in prototypical nuclear starbursts ( e.g. ngc 7714 , gonzlez - delgado 1995 ) , while in mrk 1 , mrk 463e , mrk 477 and mrk 533 the excitation is higher and similar to the excitation in young and very highly excited hii regions ( e.g. ngc 2363 , gonzlez - delgado 1994 ) .
the difference in excitation between these two sub - groups within the composites is suggestive of evolutionary effects , where the youngest systems are the ones with highest excitation .
this is consistent with other properties of composites , such as @xmath102 , the fraction of total optical light provided by intermediate and young stars ( @xmath72 ) , and the detection of wolf - rayet features ( see [ sec : evolution ] ) .
if this interpretation is correct , i.e. , the narrow and broad emission line components are produced by the starburst and agn respectively , the ratio of the narrow to total line flux gives an estimate of the starburst contribution to the ionization of the gas .
for h@xmath3 , this ratio ranges between 37% ( in mrk 463e ) and 80% ( in mrk 1073 and ngc 5135 ) , in agreement with our coarser estimates in [ sec : he2_and_o3 ] .
these values suggest a significant impact of the circum - nuclear starbursts upon the emission line ratios in seyfert 2s .
lrrrr mrk1 & 11.6 & 11.5 & 38 & 38 mrk273 & 3.1 & & & mrk463e & 8.9 & 8.3 & 37 & 39 mrk477 & 8.8 & 8.7 & 58 & 59 mrk533 & 9.4 & 12.5 & 54 & 46 mrk1066 & 2.9 & 5.7 & 70 & 55 mrk1073 & 4.6 & 15.4 & 80 & 54 ngc5135 & 2.7 & 10.5 & 80 & 54 ngc7130 & 2.7 & 11.7 & 60 & 25 ic3639 & 5.4 & 12.0 & 66 & 47 [ tab : profiles ] the fc strength , @xmath171 s , emission line ratios and profiles all consistently reflect the presence of circum - nuclear starbursts in seyfert 2 s and can therefore be used as the empirical diagnostics of compositeness which we set out to identify in this paper . the line - fc correlations presented in figs . [
fig : hb_and_o2_x_fc ] and [ fig : he2_and_o3_x_fc ] do not play any direct part in such diagnostics .
nevertheless , because of their high statistical significance , potential relevance and controversial history , in this section we speculate on what might be driving them .
we do this by contrasting two extreme views on the origin of the fc in `` pure '' seyfert 2 s .
since we have established that circum - nuclear starbursts make substantial contributions both to the fc and h@xmath3 emission , and given that these two quantities are causally linked in starburst systems , _ the presence of circum - nuclear starbursts in seyfert 2 s naturally leads to a h@xmath3-fc correlation_. even if the agn components of @xmath116 and @xmath114 are uncorrelated , the starburst portions should suffice to maintain a h@xmath3-fc correlation in the combined agn @xmath0 starburst data . one extreme interpretation of fig .
[ fig : hb_and_o2_x_fc]a is thus that the h@xmath3-fc correlation is essentially driven by the presence of circum - nuclear starbursts .
if this is to apply to _ all _ sources , the fc must be dominated by the starburst _ even in `` pure '' seyfert 2s_. varying proportions of starburst to agn ionizing power can be invoked to account for the few `` pure '' seyfert 2 s whose h@xmath3/fc ratios exceed the maximum @xmath172 reached by composites .
this can be explained either by scaling up the nuclear ionizing source , thus increasing @xmath114 while keeping @xmath116 constant , or by scaling down the starburst , which also increases @xmath102 since @xmath116 would be more affected than @xmath114 .
such adjustments can even be dispensed with if one allows for differential extinction to the line and fc , since we have seen in fig .
[ fig : hb_x_fc_reddening]b that this alone substantially reduces the @xmath102 differences between composite and `` pure '' systems . but how do the [ oiii ] ( fig .
[ fig : he2_and_o3_x_fc]a ) and heii - fc ( fig . [ fig : he2_and_o3_x_fc]c ) correlations fit into this model ?
table [ tab : correlations ] shows that these correlations , while poorer than the one between @xmath114 and @xmath116 , are still significant at the @xmath173 confidence level , and that the reason for the larger scatter is the small @xmath104}$ ] and @xmath174 of some 5 or 6 composites .
such a scatter is in fact expected in a starburst dominated fc model , since the [ oiii ] and ( especially ) the heii line fluxes trace the agn power , whereas @xmath116 is a measure of the starburst power , so that @xmath104}$ ] and @xmath174 are indicators of the varying contrast between these components .
decreasing the power of the starbursts in composite systems by factors of 23 while keeping the agn constant would move composites along @xmath34 horizontal lines to the left in figs .
[ fig : he2_and_o3_x_fc]a and c , placing them among `` pure '' seyfert 2 s .
while it is easy to see why composites deviate in @xmath104}$ ] and @xmath174 , it is not clear why `` pure '' seyfert 2 s should define line - fc correlations as good as those revealed in figs .
[ fig : hb_and_o2_x_fc ] and [ fig : he2_and_o3_x_fc ] ( see also table [ tab : correlations ] ) .
if their fc is indeed dominated by a weak starburst , then the fact that their agn - dominated @xmath140}$ ] and @xmath129 scale with @xmath116 would indirectly imply the existence of a _ proportionality between the starburst and agn powers _
, i.e. , that more powerful starbursts occur in more powerful agn ! further evidence for this scaling is presented in [ sec : iras ] , [ sec : luminosity_connection ] and gd01 .
an opposite model would be to postulate that the fc is dominated by starbursts _ only _ in composites , while in `` pure '' seyfert 2 s the fc is dominated by scattered light , so these two classes should not be mixed when discussing line - fc relations .
this interpretation would lead us to the intriguing conclusion that the scattering efficiency , @xmath175 , which links the observed fc1 to the intrinsic nuclear fc via @xmath176 , does not vary substantially among seyfert 2 s .
the surprise comes from the fact that @xmath175 depends on the geometry of the mirror and the optical depth of scattering particles ( miller , goodrich & mathews 1991 ; cid fernandes & terlevich 1995 ) .
since there is no _ a priori _ reason why @xmath175 should be similar for all seyfert 2 s , we would expect a _ wide spread _ of @xmath105 s , contrary to the relatively narrow ranges observed , particularly if composites are excluded .
this argument has in fact been pursued by mulchaey ( 1994 ) to explain the _ absence _ of line - fc correlations in their data , from which they claim a consistency with the unified model . as acknowledged by mulchaey etal .
, however , it is possible that , just as we find here , a substantial part of the scatter in their line - fc plots is associated with contamination by starburst activity . that this is more than a possibility is illustrated by the fact that the seyfert 2 list of mulchaey ( 1994 ) includes 12 of our 15 composites !
it also includes other prime composite suspects , most notably ngc 6221 , a source for which levenson ( 2001b ) find a complete dominance of the starburst , with the agn shining through only in hard x - rays . as a whole ,
their sample is similar to the nlagn sample of sbmq95 in terms of the mixed types of sources . indeed , as seen in figs .
[ fig : hb_and_o2_x_fc]b , d and [ fig : he2_and_o3_x_fc]b , no line - fc correlations are present for this heterogeneous sample . furthermore , contamination by starbursts is certainly larger in the mulchaey data than in ours , since they trace the fc by the far - uv flux collected through the large aperture of the iue spectrometer , a procedure which follows the premise that starlight contamination is irrelevant in the uv .
this premise was proven wrong by heckman ( 1995 ) , who have demonstrated that , on the contrary , scattered light from a hidden seyfert 1 is a minor contributor to the iue spectra of seyfert 2 s , which are more typical of reddened starbursts . at least in relative terms
, it is clear that our observations are better suited to investigate line - fc relations .
our results essentially recover the classic yee ( 1980 ) and shuder ( 1981 ) correlation between line and fc luminosity , which extends over more than 5 decades in luminosity from seyfert 2 s to qsos .
looking at their original plots one sees that seyfert 2 s show a larger scatter around the correlations than type 1 sources , but they still broadly follow the correlations . while it is true that the interpretation of their line - fc correlations was left orphan with the advent of the unified model , their reality has not been convincingly challenged as yet .
similarly , barring unidentified selection effects , it is hard to deny the reality of the line - fc correlations seen in figs .
[ fig : hb_and_o2_x_fc ] and [ fig : he2_and_o3_x_fc ] . in a @xmath177 model
@xmath102 equals @xmath178 , where @xmath179 is the equivalent width which would be measured from non - obscured lines of sight to the seyfert 1 nucleus .
observationally , the narrow h@xmath3 in seyfert 1 s roughly spans the range between @xmath180 to 75 ( goodrich 1989 ; rodrguez - ardila , pastoriza & donzelli 2000 ) .
this spread is already comparable to the @xmath181 interval span by `` pure '' seyfert 2 s in our sample , which does not leave much room for a spread in @xmath175 . combining these intervals we obtain an acceptable range for @xmath175 of roughly 0.010.1 .
( as a reference , for ngc 1068 , the only source for which the scattering has been modeled in detail , @xmath182 ; miller 1991 ) .
a `` well behaved ''
@xmath175 is thus required in order to explain the h@xmath3-fc correlation for `` pure '' seyfert 2 s .
this interpretation is analogous to , and just as puzzling as , the interpretation of the constancy of the @xmath19 of broad h@xmath3 among type 1 seyferts and qsos as due to a roughly constant covering fraction of broad line clouds in all sources ( yee 1980 ; shuder 1981 ; osterbrock 1989 ; binette 1993 ) . while viable , this model obviously requires some fine tuning .
yet another possibility is that the fc carries a sizable _ nebular _ component in `` pure '' seyfert 2 s , which would naturally produce line - fc correlations .
a purely nebular fc , however , would produce @xmath102 of order 2000 for electron temperatures between 10000 and 20000 k , much larger than the values we find .
a hotter plasma , however , can enhance the continuum emission and lower the @xmath105 s .
this is the essence of tran s ( 1995c ) model for fc2 : emission from the @xmath183 k electron - scattering mirror . if this emission is somehow tied to the line emission which is reasonable since both arise in circum - nuclear gas heated by the active nucleus then a line - fc correlation follows .
while our personal prejudices tend to favor a starburst - dominated fc interpretation , we are not in a position to rule out a role for a non - stellar fc for `` pure '' seyfert 2 s . indeed ,
the fact that ngc 1068 mingles so well among composite starburst / seyfert 2 s is a clear demonstration that our analysis can not resolve the ambiguous nature of the fc in `` pure '' seyfert 2 s .
since we know that several composites have a scattered component ( table [ tab : tabelao2 ] ) , the correct interpretation of the line - fc relation will certainly involve a mixture of the extreme possibilities discussed above .
in figs . [ fig : hb_and_o2_x_fc][fig : hb_x_fc_reddening ] we have seen that , besides having smaller emission line @xmath105 s , the tendency of composites to live preferentially among luminous sources allows them to be more clearly separated from `` pure '' seyfert2 s in line - fc diagrams . to evaluate this tendency in a reddening independent fashion , fig .
[ fig : iras]a shows the far ir luminosity , computed with the 60 and 100 @xmath24 m iras bands ( cf . sanders & mirabel 1996 ) , against @xmath21 for the seyfert 2 sample .
the tendency identified by gd01 for composites to be powerful ir emitters is clearly seen in this figure . except for mrk 1210 , eso 362-g8 and ngc 5643 , all composites have @xmath184 l@xmath5 .
the difference in median @xmath185 between composites and `` pure '' seyfert 2 s is a factor of 5 .
comparison with fig .
[ fig : iras]b shows that composites are also more powerful far ir sources than typical starburst galaxies , being more comparable to the merger systems studied by lk95 .
normal galaxies with little star - formation would populated the large @xmath21 and low @xmath185 portion of the plot ( telesco 1988 ; soifer , houck & neugbauer 1987 ) .
ir luminous galaxies have a well know preference to be interacting systems ( sanders & mirabel 1996 ) .
[ fig : iras]a therefore suggests that interactions and the associated gas fueling of the nuclear regions may be a key ingredient in triggering circum - nuclear starbursts in agn .
indeed , a substantial fraction of our composites ( 9 of 15 ) are associated with interacting systems or groups ( gd01 ; levenson 2001a ; sb00 ) : ngc 7130 , mrk 1 , mrk 273 , mrk 463e , mrk 477 , mrk 533 and ic 3639 are all in interacting systems , while ngc 5135 and ngc 7582 belong to groups .
the analogous fraction for the the `` pure '' seyfert 2 s is only 4 out of 20 ( mrk 348 , mrk 607 , ngc 5929 and ngc 7212 ) .
this may explain why the composites resemble the mergers in fig .
[ fig : iras ] . since @xmath21 works essentially as a measure of the starburst strength , fig .
[ fig : iras ] indicates a link between the far ir and the amount of recent star - formation in seyfert 2 s .
this link is investigated in fig .
[ fig : iras_x_fc ] .
the abscissa in fig .
[ fig : iras_x_fc]a is the total observed luminosity at 4861 , whereas in fig .
[ fig : iras_x_fc]b we filter out the light from old stars by using @xmath186 . the post - starburst population , which was not included in the line - fc analysis because it does not ionize the gas , is now included because it can heat the dust
furthermore , table [ tab : correlations ] shows that the far ir flux is somewhat better correlated with @xmath187 than with @xmath116 alone .
comparison of figs .
[ fig : iras_x_fc]a and b show how @xmath188 has the property of unveiling a correlation immersed in the scatter when the whole continuum is used .
an estimate of the effects of extinction is presented in fig .
[ fig : iras_x_fc]c , where we deredden @xmath189 using the h@xmath25/h@xmath3 values listed in table [ tab : tabelao2 ] .
the bottom panels in fig .
[ fig : iras_x_fc ] show the corresponding results for starburst galaxies and mergers .
comparison of the top and bottom panels shows that seyfert 2 s have @xmath190 ratios larger than galaxies in the starburst sample , but compatible with the lk95 mergers . despite the mixed emission line properties of the sources in the merger sample , it is reasonable to attribute most of their fir and fc emission to star - formation in a dusty environment ( lutz 1999 ; 1998 ) .
this similarity is thus suggestive of the presence of powerful dusty circum - nuclear starbursts in seyfert 2 s , consistent with our knowledge of composites .
a usual caveat when dealing with iras fluxes is that one has to allow for the fact that a substantial fraction of the far ir light originates well outside the nuclear region .
aperture effects may therefore be partially responsible for the larger @xmath190 found for objects in the seyfert 2 than in the starburst sample .
this effect was detected for sources in common between the seyfert 2 and nlagn samples , the latter having systematically larger @xmath187 and thus smaller @xmath190 . at any rate
, it is clear that circum - nuclear starbursts can be responsible for the bulk of the far ir emission in seyfert 2 s .
this was in fact confirmed in the few cases where this issue has been examined more closely ( e.g. , mrk 477 heckman 1997 ; see also rigopoulou 1999 ) .
we have shown that the near - uv nuclear light in the seyfert 2 composites is dominated by young and intermediate - age stars . in order to make some causal connection between these young stars and the seyfert phenomenon
, we must establish that the properties of this stellar population are unusual compared to the nuclei of normal galaxies .
young and intermediate - age stars dominate the near - uv continuum in the nuclei of normal galaxies of hubble type sbc and later ( e.g. heckman 1980 ) .
seyfert galaxies are generally classified as early type disk galaxies ( e.g. nelson & whittle 1996 ; ho , filippenko & sargent 1997 ) , whose nuclei are dominated by old stars
. however , the hubble types of seyfert galaxies based on ground - based images are quite uncertain , since they are generally much more distant than the galaxies used to define the hubble sequence .
indeed , the hst imaging survey by malkan , gorjian & tam ( 1998 ) reveals a significant number of seyfert 2 s ( including some of our composites ) in which the hst - based hubble type is much later than that assigned from earlier ground - based images ( storchi - bergmann 2001 )
. could it be that the seyfert 2 composites are simply otherwise normal late - type galaxies containing an agn ? the abnormally large far - ir luminosities of the composites argues against this . in this section
we also show that the observed uv surface - brightnesses of the seyfert 2 composite nuclei are much larger than the nuclei of normal galaxies ( of any hubble type ) , but rather similar to typical starbursts .
we have compared the surface brightnesses in the near uv ( @xmath191 ) of all our seyfert 2 s with those of normal galactic nuclei based on the imaging data published by jansen ( 2000 ) . for the normal nuclei ( spanning the full range of hubble types )
we have computed the mean surface brightness through two metric apertures with radii of 250 and 500 pc , chosen to represent typical values of the spectroscopic apertures used for the seyfert 2 nuclei ( see fig . [
fig : aperture_x_wk ] ) .
this was done converting the u - band surface photometry of jansen ( 2000 ) to monochromatic surface brightnesses at 3650 , including a correction for galactic extinction ( from ned ) .
we have also measured near - uv surface brightnesses from the spectra of the starburst and merger samples . in these samples
we have restricted our comparison to galaxies in which the projected aperture size of the spectra had a radius less than 1 kpc ( see fig .
[ fig : aperture_x_wk ] ) .
thus , in all cases we are comparing uv surface - brightnesses on similar physical scales ( radii of several hundred pc ) .
the results are shown in fig .
[ fig : sbu1 ] , from which we draw the following conclusions .
first , the seyfert 2 composites have much higher central uv surface brightnesses than normal galaxies of any hubble type ( by roughly an order - of - magnitude on average ) . a correction for dust extinction would undoubtedly increase the disparity still further .
second , the seyfert 2 composites have rather similar central uv surface - brightnesses to the members of the starburst and merger samples . when the uv surface - brightness is corrected for extinction ( see fig .
[ fig : sbu2 ] ) , the seyfert 2 composites reach values similar to those which would be exhibited by the `` maximal '' starbursts studied by meurer ( 1997 ) . third ,
the seyfert 2 composites have tend to have higher central uv surface - brightness than the `` pure '' seyfert 2 s ( by an average factor of about 2 ) . as we show in fig .
[ fig : sbu2 ] , the difference between the two classes is larger ( factor of @xmath34 4 ) when an extinction correction is made .
fourth , the `` pure '' seyfert 2 s overlap with the normal galactic nuclei having the highest uv surface brightnesses .
these latter tend to be the nuclei of normal early - type ( sb and earlier ) galaxies whose high uv surface brightness is due to their prominent bulge component .
this is consistent with `` pure '' seyfert 2 s being otherwise normal early - type galaxies with only a modest contribution of light from an agn or young stars ( as shown by our spectral synthesis ) .
note that the outlying seyfert 2 with extremely high surface brightness is the nucleus of ngc 1068 ( due to the small physical size of the spectroscopic aperture and the strong contribution by scattered agn light ) .
finally , fig .
[ fig : sbu2 ] also shows the expected trend between increasing uv surface brightness and an increasing relative contribution to the near - uv light by young and intermediate - age stars ( as parameterized by @xmath21 ) .
our experiments have consistently shown that seyfert 2 s with circum - nuclear starbursts detected by means of either far - uv stellar wind lines , high order balmer absorption lines or the wr bump also present distinctive optical / near - uv continuum , optical emission line , far ir continuum , and near - uv surface brightness properties , all in agreement with the expected impact of starbursts upon such observables . given this remarkable consistency , one can use the results of our analysis as alternative tools to identify composite sources .
the following empirical criteria can be used to characterize a seyfert 2 nucleus as composite : 1
. weak ca ii k : @xmath1 . or , equivalently , a small contribution of old stars : @xmath192 .
2 . weak emission lines with respect to the fc : @xmath193 , @xmath104 } < 2500 $ ] , @xmath112 } < 600 $ ] . 3 .
low excitation : heii / h@xmath164 , [ oiii]/h@xmath165 .
large far - ir luminosity : @xmath194 l@xmath5 .
5 . composite line profiles , with a narrow component typical of hii - nuclei and a broader one with nlr - like excitation . 6 .
high near - uv surface - brightnesses on circum - nuclear scales : @xmath195 l@xmath7 pc@xmath8 uncorrected for extinction and @xmath196 l@xmath7 pc@xmath8 after extinction correction .
these criteria are laid roughly in order of importance , the ca ii k ( or @xmath47 ) limit being by far the most useful . the upper limits for the @xmath105 s and emission line - ratios quoted above were chosen as to encompass all composites in our sample , most of which have substantially smaller values than those listed above .
as argued above , one can not base a classification on these properties alone , but they are useful to confirm the compositeness for sources satisfying the first criterion .
taken together , these empirical rules separate composite from `` pure '' systems very efficiently . naturally , there are border line cases .
the moderate fc ( @xmath197 ) galaxies mrk 3 , mrk 34 and ngc 7212 , which we discussed in [ sec : moderate_and_weak_fc_sources ] , lie either within or close to the limits outlined above .
the clearest distinction between these galaxies and our composites is in terms of their emission line @xmath19 s , @xmath198321 and @xmath104 } = 2667$]3565 , typical of `` pure '' seyfert 2 s but larger than in any of our composites , and well away from starburst galaxies .
note , however , that these high @xmath105 s , coupled with the dominant contribution of old stars ( @xmath199 ) are precisely the conditions under which the identification of starburst features are most difficult ( [ sec : difficulty_to_detect_starbursts ] ) , so we can not rule out the possibility that their fc is dominated by relatively weak circum - nuclear starbursts .
the other moderate fc source in our sample is ngc 1068 , for which there is overwhelming evidence that the nuclear optical fc is predominantly scattered agn light .
the similarities between this prototype seyfert 2 and our composites , whose fc are dominated by starbursts , is a reminder of how similar starbursts and agn can be , and a demonstration of the limitations of our diagnostics for border line cases .
overall , the empirical criteria listed above fulfill our objective of delineating a region in a space of easily obtained observables which can be used to indirectly infer the presence of circum - nuclear starbursts in seyfert 2 s .
we therefore anticipate that their application to large samples of near - uv spectra will reveal many more composite systems .
in fact , since these criteria mix several kinds of properties , it is not at all obvious whether they will hold for larger samples . future applications of these diagnostics will therefore not only serve as a means of identifying composite candidates but to verify the consistency of the scheme as a whole .
it will also be interesting to incorporate other multi - wavelength data in order to help disentagling the starburst and agn components .
valuable tools in this sense were developed for the ir part of the spectrum by miley , neugebauer & soifer ( 1985 ) , lutz ( 1998 ) , genzel & cesarsky ( 2000 ) , laurent ( 2000 ) and hill ( 1999 , 2001 ) , among others . in the radio part of the spectrum ,
some of the diagnostics and techniques presented by kewley ( 2000 ) , smith ( 1998a , b ) , lonsdale ( 1993 ) and condon ( 1991 ) can be used , while x - rays signatures of circumnuclear starbursts have been recently investigated by levenson ( 2001a ) .
we have shown that starburst / seyfert 2 composites have large h@xmath3 , fc and far ir luminosities within our sample .
these are all quantities which may carry a significant , even dominant contribution from the starburst itself . in order to investigate whether these powerful starbursts are associated with correspondingly powerful agn
we compared the [ oiii ] , heii and 12 @xmath24 m luminosities of composites and pure seyfert 2 s .
these properties are less affected by the starburst component and thus provide an empirical measure of the agn intensity .
we find that composites are also skewed towards large luminosities in all these properties .
for instance , 53% of the sources above but just 31% of those below the median reddening - corrected @xmath201}$ ] of @xmath202 erg@xmath15s@xmath17 are composites .
nearly identical numbers apply to @xmath203 , while for @xmath204 , which arguably measures the hot dust heated by the agn ( pier & krolik 1992 ; spinoglio & malkan 1989 ) , the composite percentages are 65 and 25% above and below the median respectively .
these asymmetries are comparable to those in starburst - sensitive properties . for @xmath160 , for instance , we find 65% of composites above the median @xmath205 erg@xmath15s@xmath17 , but just 19% bellow it .
the corresponding numbers for @xmath185 are 71 and 19% .
circum - nuclear starbursts in seyfert 2 s , at least those we can detect , are therefore powerful ( bolometric luminosities of @xmath206 l@xmath7 , judging by @xmath185 ) , and surround correspondingly powerful agn .
luminosity link _ between agn and circum - nuclear starburst activity can be interpreted in several ways .
perhaps the starburst activity stirs up the ism in a way which induces a large rate of gas accretion by the central super - massive black - hole ( colina & wada 2000 ) , while less powerful agn can be feed with more modest accretion rates , dispensing the helping hand of a circum - nuclear starburst . while feasible ,
this does not seem to be the only mechanism to produce powerful agn , since there are also some high luminosity `` pure '' seyfert 2 s in our sample . or perhaps the `` @xmath207 '' relation extends all over the @xmath208 spectrum , but is only detected for the most luminous agn because of our difficulties to identify weak circum - nuclear starbursts .
whatever the correct interpretation , our results strongly suggest that powerful seyfert 2 s have at least a 1 in 2 chance of harboring detectable circum - nuclear starbursts .
prediction _ can be readily tested by applying the diagnostics developed in this and our previous work to near - uv spectra of such systems , some of which have luminosities rivaling those of quasars ( de grijp 1992 ) .
extending the starburst - agn connection studies towards the low luminosity end , on the other hand , will require alternative techniques .
high spatial resolution ( i.e. , hst ) near - uv spectroscopy offers the best hope of maximizing the contribution of circum - nuclear star - forming regions with respect to the old stellar population background . this technique was recently applied to the liner ngc 3507 by shields ( 2000 ) , where young stars previously undetected from the ground were clearly revealed with hst .
the composite _ vs.
_ `` pure '' seyfert 2 classification which runs throughout this paper does not necessarily reflect an intrinsic dichotomy among type 2 seyferts . in the spirit of unification , one would rather prefer to look at these two classes as representing extremes of a continuous distribution of some underlying physical parameter .
this view is warranted by the continuity and overlap in observed properties between composites and `` pure '' seyfert 2 s found in this and other studies ( levenson 2001a ; veilleux 2001 and references therein ) .
at least two ingredients must be involved in any physical description of a starburst - agn connection , one of which is some measure of the relative starburst to agn intensities , such as the ratio of star - formation to accretion rate , or the black - hole to stellar mass ratio . of the properties investigated in this paper the @xmath105 s , mainly of [ oiii ] and heii , and the excitation ratios [ oiii]/h@xmath3 and heii / h@xmath3 are in principle the ones
more closely associated with this `` starburst - to - agn ratio '' .
the fact that as a whole `` pure '' seyfert 2 s behave more like agn in terms of these indicators than composites is thus consistent with them having a smaller starburst - to - agn ratio . the strongest contrast effect , however , is between the starburst and old stellar population , as deduced from the fact that _ all _ composites but only one `` pure '' seyfert 2 have @xmath192 , a limit which can be expressed in a more empirical way by @xmath1 .
the ca ii k line originates in bulge stars , being neither a tracer of starburst nor agn activity , so this threshold is clearly an observational limitation which prevents us from identifying circum - nuclear starbursts not powerful enough to dilute @xmath21 below 10 .
the `` pure '' category ought to contain at least some such systems , and in fact several `` pure '' seyfert 2 s in our sample , particularly the less luminous ones ( see fig .
[ fig : hb_and_o2_x_fc ] ) , have @xmath105 s and excitation ratios within the range spanned by composites .
the other ingredient is _
evolution_. the combination of starburst - to - agn ratios and evolution can in principle account for the whole range between composite and `` pure '' systems . while our coarse , 3-component population synthesis analysis does not provide a detailed description of the evolutionary history of circum - nuclear starbursts in seyfert 2 s , in [ sec : constrast_and_evolution ] we have seen that the relative strengths of @xmath51 and @xmath49 in composites bear a very good correspondence with the evolutionary status deduced from a much more detailed spectral analysis ( sb00 ; gd01 ) .
the ratio @xmath209 can thus be used as an empirical measure of the time elapsed since the onset of star - formation , allowing us to broadly assess evolutionary trends in the data .
[ fig : evolution ] shows that younger systems have stronger h@xmath3 , heii and [ oiii ] emission lines , which we tentatively interpret as evidence for evolutionary effects .
only composites are plotted , since only for them we can safely interpret the fc component as due to young stars and hence read @xmath73 as an age indicator .
also , note that @xmath19 s in this figure are measured with respect to the whole `` starburst flux '' @xmath210 .
this choice of normalization facilitates the comparison with spectral evolution models such as starburst99 ( leitherer 1999 ) , particularly for systems undergoing star - formation over a period of order @xmath50 yr . since h@xmath3 contains a strong contribution from gas ionized by the starburst , fig .
[ fig : evolution]a roughly traces the evolution of the starburst .
the heii and [ oiii ] to @xmath187 ratios , on the other hand , are measures of the relative power of the agn and the starburst , and so fig .
[ fig : evolution]b and c can be used to assess the evolution of the agn with respect to that of the starburst . as two extremes , we consider both an instantaneous burst and constant star - formation to describe the starburst . in the former case ,
the optical continuum due to the young stars peaks in luminosity about 3 myr after the burst , and then fades by a factor of 3 by 10 myr and by another factor of 10 by 50 myr .
the systematic decrease in @xmath211 and @xmath212}$ ] in the more evolved starbursts would therefore require that the agn fades more rapidly than this . in this instantaneous burst scenario
, the agn lifetime must be very short in the composite seyfert 2 s ( @xmath213 few @xmath214 years ) . since the agn should disappear before the intermediate - age stars
, the descendants of the seyfert 2 nuclei would be part of the post - starburst population ( the `` e+a '' galaxies ) .
thus , the only plausible evolutionary link between the composite and `` pure '' seyfert 2 s would seemingly require that latter are the evolutionary _ precursors _ to the former .
this does not seem physically appealing , among other reasons , because the agn s in the `` pure '' seyfert 2 s are significantly less powerful on - average than in their supposed descendants ( the youngest composites ) .
moreover , an instantaneous burst description is clearly not appropriate for sources like ngc 5135 and ngc 7130 , which contain a mix of young and intermediate age stars .
alternatively , we can consider a more long - lived ( @xmath215 years ) event in which the star - formation remains roughly constant .
the @xmath157 values predicted by starburst 99 for constant star formation vary from 150450 to 1060 between 0 and @xmath50 yr for different initial mass functions and solar metallicity .
this range covers the values in fig .
[ fig : evolution]a .
the @xmath216-fold decrease in @xmath157 during this period is also in agreement with the data . in this model
, the luminosity of the optical continuum steadily grows with time ( e.g. by a factor of 3 to 4 between 10 and 300 myr after the start of the burst ) . the systematic decrease in @xmath211 and @xmath217 in the more evolved starbursts
would therefore imply that the agn s luminosity remains relatively constant over this time - scale .
long time - scale bursts are indeed seen in pure starburst galaxies , where ground based spectra usually integrate over regions containing several young , but non - coeval , super - associations ( meurer 1995 ; gonzlez delgado 1999 ; lanon 2001 ) .
individual starburst knots have been identified in some of our composites ( gonzlez delgado 1998 ) , so we expect this averaging effect to happen in seyfert 2 s too . in this scenario ,
both the starburst and agn are relatively long - lived , and the agn could well outlive the starburst . as the star - formation ceases , the starburst will fade first in h@xmath3 , then in the optical continuum , eventually reaching the point where only scattered fc and agn - powered lines survive . with the drop in the starburst - continuum
the fc becomes weaker and @xmath21 less diluted .
galaxies like eso 362-g8 , ngc 5643 and mrk 78 , with their weak fc but pronounced post - starburst populations may represent this intermediate stage .
[ oiii ] and heii will be much less affected than h@xmath3 by the death of massive stars , so excitation ratios will go up .
starburst - heated dust emission will subside , decreasing @xmath185 .
all these suggest that composites will end up as `` pure '' seyfert 2 s .
clearly , the above is quite speculative . eventually , it may be possible to test these ideas empirically with improved age dating techniques which will allow us to evolve detected starbursts back and forth in time and thus to evaluate whether the age of circum - nuclear starbursts may be used as a clock for nuclear activity . as in any other astrophysical context where evolution is important ,
statistically significant samples will be necessary to trace all relevant stages .
we have examined a sample of 35 galaxies previously studied in a series of papers which addressed the occurrence of circum - nuclear starbursts in seyfert 2 s and implications for the `` starburst - agn connection '' . the detailed spectral analysis performed in these previous papers
was used to classify the nuclei into 15 composite and 20 `` pure '' seyfert 2 s , with composites defined as nuclei in which large populations of stars with ages of @xmath41@xmath50 yr have been conclusively detected either through uv stellar winds features , high order balmer absorption lines in the near uv or the wr bump . using this classification as a guide
, we investigated continuum colors , absorption line equivalent widths , emission line properties , far - ir luminosities and near - uv surface brightnesses for the sample galaxies and comparison samples of normal and starburst galaxies , with the primary goal of verifying whether these properties reflect the presence or otherwise of circum - nuclear starbursts . since this consistency was verified , these properties , all of which are more easily measured than those originally used to detect these starbursts , offer alternative empirical diagnostics of compositeness which can be applied to larger and more distant samples .
* we have applied a semi - empirical population synthesis method which effectively provides a bi - parametric description of each galaxy in terms of old ( @xmath218 yr ) , post - starburst ( @xmath50 yr ) populations and a `` featureless continuum '' which represent the combined contribution of @xmath52 yr stars and a non - stellar fc .
in practice , this representation requires little more than the equivalent width of the ca ii k line , and empirical calibrations were presented which allow an estimation of the ( @xmath219 ) vector from @xmath21 and the @xmath220 color , both easy to measure observables .
this very simple scheme proved instrumental in unveiling a number of properties and gives results which are consistent with a more detailed spectral analysis of the stellar population mixture . * _ all _ starburst / seyfert 2 composites but _ only one _ `` pure '' seyfert 2 ( ngc 1068 ) have @xmath61 at 4861 .
this threshold can be expressed in a more empirical way by @xmath1 . given that our operational definition of composite/``pure '' systems is based on the detection or otherwise of stellar features from massive stars ,
the most natural interpretation for this result is that it is due to a _ contrast _ effect : seyfert 2 s whose circum - nuclear starbursts are overwhelmed by the light from the old stars in the host galaxy are not recognized as composite systems .
at least some `` pure '' seyfert 2 s must fit this description , i.e. , composites with weak starbursts .
this may in fact apply to most , if not all , `` pure '' seyfert 2 s , since we can not determine the nature of their weak uv - excess with the data discussed here . *
a related result is that strong fc sources , those with @xmath84 , are invariably associated with certified starburst / seyfert 2 composites .
we have argued that this is expected , since if scattered light was dominant in such sources then reflected broad lines would render them classified as type 1 seyferts . *
the equivalent widths of emission lines ( h@xmath3 , [ oiii ] , [ oii ] ) , when measured with respect to the fc , are larger in `` pure '' seyfert 2 s than in starburst galaxies .
composites have @xmath105 s intermediate between these two extremes .
this fits with the idea that circum - nuclear starbursts in seyfert 2 s contribute proportionately more to the fc than to the line flux , since , while the agn - powered line emission is seen directly , only a small fraction of the corresponding non - stellar fc reaches the observer after scattering . * part of the separation in @xmath102 between `` pure '' seyfert 2 s and composites may be due to another contrast effect , in which the high order balmer lines , the main diagnostic of massive stars in the optical , are diluted and eventually filled by nebular emission in strong lined sources , preventing them from being recognized as composites .
* composite systems tend to be less excited in terms of the heii / h@xmath3 and [ oiii]/h@xmath3 ratios .
this indicates that the circum - nuclear starbursts have a direct impact on the emission line properties of seyfert 2 s , with typically @xmath2 of h@xmath3 being powered by ob stars .
* the dual nature of the ionizing source in composites also leaves a kinematical imprint on the emission line profiles , which can be described in terms of a broad component with nlr - like line ratios and a narrower and ( usually ) less excited one .
these components are presumably associated with the gas ionized by the agn and starburst respectively . *
composite galaxies are luminous in the far - ir , with @xmath221 to @xmath222 l@xmath5 and are more luminous than `` pure '' seyfert 2 s and normal galaxies . as a whole , composites are rather similar to merger systems in terms of their far - ir properties . about 50% of the composites , but only 20% of `` pure '' seyfert 2 s are in interacting systems , suggesting that galaxy interactions spur circum - nuclear starbursts . *
composite galaxies have near - uv surface - brightnesses in their central - most few hundred pc that are on - average an order - of - magnitude higher than in normal galaxies of any hubble type .
these high surface brightnesses are similar to those in pure starburst galaxies .
they are also several times higher on - average than in the `` pure '' seyfert 2 s ( whose central surface brightnesses in most cases overlap those of normal early - type galaxies ) . *
these properties were condensed into a set of empirical criteria which separate composite from `` pure '' systems .
most of the diagnostic power resides in the @xmath1 limit .
other properties ( small emission line @xmath105 , low excitation and large @xmath185 ) serve to confirm the @xmath21-based diagnostic of compositeness and to judge border line cases .
we predict that many more starburst / seyfert 2 composites will be uncovered by applying these diagnostics to large samples of near uv spectra . *
the tendency of composites with older starbursts to have smaller h@xmath3 , heii and [ oiii ] emission - line equivalent widths was tentatively interpreted as a sign of the evolution of circum - nuclear starbursts in seyfert 2 s .
circum - nuclear starbursts which are powerful today may fade away and , barring evolution of the active nucleus , appear as `` pure '' seyfert 2 s . by improving the dating techniques and enlarging the databases we hope to trace this evolutionary sequence and thereby put strong constraints on physical scenarios for the starburst - agn connection .
it is a pleasure to thank the organizers and participants of the guilllermo haro 2000 ( inaoe ) workshop on `` the starburst - agn connection '' , for a stimulating meeting in which this work was conceived .
we would like to thank r. a. jansen and c. t. liu for making their data available .
rcf thanks the hospitality of johns hopkins university , where this work was developed , and the support provided by the national science foundation through grant # gf-1001 - 99 from the association of universities for research in astronomy , inc .
, under nsf cooperative agreement ast-9613615 . partial support from cnpq and pronex
are also acknowledged .
hrs work was supported by nasa under grant
nag5 - 9343 .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . | we examine a representative sample of 35 seyfert 2 nuclei .
previous work has shown that nearly half ( 15 ) of these nuclei show the direct ( but difficult - to - detect ) spectroscopic signature at optical / near - uv wavelengths of the hot massive stars that power circum - nuclear starbursts . in the present paper
we examine a variety of more - easily - measured quantities for this sample , such as the equivalent widths of strong absorption features , continuum colors , emission - line equivalent widths , emission line ratios and profiles , far - ir luminosities , and near - uv surface brightness .
we compare the composite starburst @xmath0 seyfert 2 nuclei to `` pure '' seyfert 2 nuclei , starburst galaxies and normal galactic nuclei .
our goals are to verify whether the easily - measured properties of the composite nuclei are consistent with the expected impact of a starburst , and to investigate alternative less - demanding methods to infer the presence of starbursts in seyfert 2 nuclei , applicable to larger or more distant samples .
we show that starbursts do indeed leave clear and easily quantifiable imprints on the near - uv to optical continuum and emission line properties of seyfert 2 s .
composite starburst @xmath0 seyfert 2 systems can be recognized by : ( 1 ) a strong `` featureless continuum '' ( fc ) , which dilutes the caii k line from old stars in the host s bulge to an equivalent width @xmath1 ; ( 2 ) emission lines whose equivalent widths are intermediate between starburst galaxies and `` pure '' seyfert 2 s ; ( 3 ) relatively low excitation line - ratios , which indicate that part of the gas ionization in these seyfert 2 s ( typically @xmath2 of h@xmath3 ) is due to photoionization by ob stars ; ( 4 ) large far ir luminosities ( @xmath4 l@xmath5 ) ; ( 5 ) high near - uv surface brightness ( @xmath6 l@xmath7 pc@xmath8 ) .
these characteristics are all consistent with the expected impact of circum - nuclear starbursts on the observed properties of seyfert 2 s .
furtheremore , they offer alternative empirical diagnostics of the presence of circum - nuclear starbursts from a few easily measured quantities . |
ever since it has been realized that there are surprises on the way from one- to two - dimensional quantum magnets @xcite , spin ladders have attracted an enormous interest .
two - leg ladders , such as sr@xmath1ca@xmath2cu@xmath3o@xmath4 @xcite , have been under intense scrutiny mainly because of the potential interplay between their spin - gaped ground state and the occurrence of superconductivity @xcite .
n - leg ladders with @xmath5 are of particular relevance , not only because they allow for a generalized test of haldane s conjecture @xcite , but also because new , _ tube - like _ lattice structures can be realized if periodic transverse boundary conditions apply , such as in [ ( cucl@xmath6tachh)@xmath7cl]cl@xmath6 @xcite and cscrf@xmath8 @xcite with @xmath9 , cu@xmath6cl@xmath8@xmath10d@xmath11c@xmath8so@xmath6 @xcite with @xmath12 , and na@xmath6v@xmath7o@xmath13 @xcite with @xmath14 . for @xmath9 ,
magnetic frustration surfaces as an additional ingredient , already for tubes with only nearest neighbor exchange , leading to a rich variety of phenomena not present in two - leg ladders @xcite .
experimentally , the four - spin tube cu@xmath6cl@xmath8@xmath10d@xmath11c@xmath8so@xmath6 has been suggested to display frustrating antiferromagnetic next - nearest neighbor exchange @xcite .
theoretically , _ unfrustrated _ four - spin tubes have been considered in two seminal papers @xcite , both , in the weak and the strong rung - coupling limit . magnetic frustration ,
however , has not been considered in these studies .
therefore , in this work , we perform a first analysis of a frustrated four - spin tube ( ffst ) @xmath15 with a lattice structure and exchange couplings @xmath16 as shown in fig.[fig1 ] .
spin-@xmath0 moments are located on the solid circles and all couplings , @xmath17 , @xmath18 and @xmath19 are antiferromagnetic ( we set @xmath20 hereafter ) .
this ffst is simpler than the one proposed for cu@xmath6cl@xmath8@xmath10d@xmath11c@xmath8so@xmath6 , where only part of the surface squares experience diagonal exchange , and the leg - couplings seem to be in - equivalent @xcite .
apart its relation to existing materials , the ffst is of interest as a 1d variant of the anisotropic triangular lattice on the torus with four site circumference , i.e. cutting the tube longitudinally one obtains an anisotropic triangular lattice strip .
frustrated four - spin tube .
solid circles represent spin-@xmath0 moments .
plaquettes ( bold black lines ) are coupled by nearest ( @xmath18 ) and next nearest ( @xmath19 ) antiferromagnetic exchange , blue and red lines . on -
plaquette couplings ( @xmath17 ) are set to unity . ] for @xmath21 , the ffst resembles a chain of weakly coupled four - spin plaquettes each of which displays a singly degenerate singlet ground state , separated by a gap of @xmath20 from the first excited triplet .
therefore perturbation theory in @xmath22 applies .
motivated by this we investigate the ffst by series expansion ( se ) in @xmath22 .
moreover , we corroborate our approach and gauge our se results by employing density - matrix renormalization group ( dmrg ) calculations . the structure of the paper is as follows . in section [ ii ]
we clarify the region of applicability of the se . section [ iii ]
details our se method .
results are presented in section [ iv ] , including the ground state energy of the ffst in sub - section [ iva ] , the one - particle excitations in sub - section [ ivb ] , as well as two - particle states in [ ivc ] .
conclusions are presented in section [ v ] . for completeness
, technicalities of the two - particle se calculations are included in the appendix [ app1 ] .
proper application of se hinges on the adiabatic renormalization of the bare starting state in terms of the coupling constants . in case of intervening second order
quantum phase transitions , the se can be used directly to limit its range of applicability in terms of diverging susceptibilities or vanishing elementary excitation gaps . in case of a discontinuous or first order transitions ,
se based on a single bare ground state fails to signal any transition .
to put our se on firm grounds _ a - priori _ , we therefore search for potential first - order quantum phase - transitions of the ffst at small @xmath22 .
phase diagram of the ffst .
solid lines : transitions at the _ classical _ level ( all are discontinuous ) .
wave vectors a , b , c and d label pitch angle of classical phase .
circles : first - order _ quantum _ critical line from dmrg ( see text ) .
inset : ground state energy versus @xmath23 from dmrg for the quantum model at @xmath24 , showing first order transition at @xmath25 . ] to this end , it is instructive to first consider the classical phase diagram of the ffst .
we allow the spin structure to be a spiral , which , due to @xmath26 symmetry can be considered to be planar @xmath27 with @xmath28 , where @xmath29 , @xmath30 , and @xmath31 $ ] .
the transverse pitch vector @xmath32 has to be discretized according to @xmath33 .
the ground state energy is @xmath34 . from this , four classical phases result , shown in fig .
[ fig2 ] : a @xmath35 antiferromagnet for @xmath36 ( region a ) , a @xmath37 columnar antiferromagnet for @xmath38 and @xmath39 ( region c ) , a @xmath40 columnar antiferromagnet for @xmath41 ( region b ) . in the remaining region
d , the energy is minimized by two degenerate incommensurate spirals with pitch @xmath42 and @xmath43 , and @xmath44 . due to @xmath32 s discretization all of the classical transitions
are discontinuous .
while the quantum analog of this rather rich phase diagram clearly deserves future analysis , our se is confined to the region of @xmath45 .
therefore we focus only this region regarding potential first order quantum phase transitions .
first , the classical diagonal transition from the @xmath37 to the @xmath35 state has no quantum analog , since it is connected to the region of @xmath21 .
there , and instead of the classical @xmath35 or @xmath37 antiferromagnets , the quantum model shows a phase of weakly coupled plaquette - singlets and the quantum ground state is protected by the singlet - triplet gap of the bare plaquette . to assess the relevance to the quantum case of the classical transitions from the @xmath37 and @xmath35 states into the doubly - degenerate @xmath46 @xmath47 phase , we resort to a dmrg analysis of the ground state energy . for this
we use the alps package @xcite . since we are only concerned with first order transitions , we refrain from any detailed finite size scaling analysis .
we find that for @xmath48 there is negligible finite size dependence of the ground state energy for ffsts of lengths @xmath49 , i.e. @xmath50 spins , and that @xmath51 states kept in the density matrix lead to @xmath52 digits of precision , which is sufficient for our purpose .
the inset in fig .
[ fig2 ] shows a typical result for the ground state energy @xmath53 obtained from dmrg versus @xmath23 along the diagonal @xmath54 .
obviously the ground state energy displays a kink at @xmath25 , which we identify with a first order transition . the circles in fig .
[ fig2 ] summarize a scan of locations of this transition which we have performed .
the size of the circles is a rough measure of the numerical accuracy for the locations of the transition .
these locations are remarkably close to those of the classical model .
below these transitions se based on decoupled plaquettes is applicable . beyond @xmath55
the critical points are increasingly hard to detect accurately from the dmrg data .
we speculate that below the first - order transition lines at @xmath56 and @xmath57 the bare plaquette state is adiabatically connected to a luttinger liquid for @xmath58 and @xmath59 , respectively .
the main focus of this work is on se in terms of @xmath18 and @xmath19 , starting from the limit of isolated plaquettes . to this end
we decompose the hamiltonian of the ffst into@xmath60 where @xmath61 is the plaquette hamiltonian at site @xmath62@xmath63_{n}\label{eq:3}\end{aligned}\ ] ] and the perturbation @xmath64 is given by@xmath65 the eigenstates of an isolated plaquette consist of four equidistant energy levels @xmath66 , labeled by the quantum number @xmath67 , and can be classified according to the total and the @xmath68-component of the plaquette spin @xmath69 .
table [ tab1 ] lists that the ground state is a singlet , the first excited state at @xmath70 is a triplet , the @xmath71 sector is composed of a singlet and two triplets , and for @xmath72 one quintet remains .
.[tab1 ] spectrum of the single plaquette .
it consists of four equidistant energy levels @xmath66 labeled by the quantum number @xmath67 , the total spin @xmath73 , and its @xmath68-component @xmath74 .
the last column enumerates the states .
[ cols="<,^,^,^,^,^",options="header " , ] in turn , @xmath75 displays an equidistant spectrum , labeled by @xmath76 . at @xmath77
, the @xmath78 sector refers to the _ _ unperturbed _ _ singlet ground state @xmath79 composed of @xmath80 singlets on all plaquettes .
the @xmath81 sector comprises single @xmath82 triplets inserted into @xmath83 at site @xmath84 .
the @xmath85 sectors are of multiparticle nature .
the perturbation @xmath86 can be rewritten as @xmath87 , where @xmath88 represent raising ( @xmath89 ) or lowering ( @xmath90 ) operators within the spectrum of @xmath75 . for the ffst
we find @xmath12 .
it has been shown quite generally @xcite , that models with the preceding type of spectrum allow for se through a continuous unitary transformation ( cut ) using the flow equation method of wegner @xcite .
the basic idea is to transform @xmath91 onto an effective hamiltonian @xmath92 which is _ block - diagonal _ in the quantum number @xmath93 .
this transformation can be achieved exactly order - by - order in @xmath22 leading to@xmath94 where the @xmath95 are weighted products of the @xmath96 operators which conserve the @xmath93-number and have their weights determined by recursive differential equations , see @xcite for details . due to @xmath93-number conservation several observables can be accessed directly from @xmath92 in terms of a se in @xmath22 . for systems with coupled spin - plaquette cut se has been used for one @xcite , two @xcite and three @xcite dimensions .
in this section we present our findings from se up to @xmath97 , sectioning the discussion according to a. the ground - state energy , b. the one- , and c. the two - particle excitations . to assess the quality of the se we complement our analysis by dmrg calculations for selected cases .
first we consider the ground state energy @xmath98 .
@xmath93-conservation leads to @xmath99 where @xmath83 is the _ unperturbed _ ground state .
ground state energy per site @xmath100 versus @xmath18 and @xmath19 , showing a monotonously decreasing behavior in the parameter range shown .
along the line of maximum frustration @xmath101 , the energy gain is smallest . ] evaluating this matrix element on chains with periodic boundary conditions ( pbc ) of a length @xmath102 , sufficient not to allow for wrap - around of graphs with length @xmath103 , i.e. @xmath104 , one can obtain _ analytic _ ses which are valid to @xmath105 with respect to eqn .
( [ eq:5 ] ) in the thermodynamic limit . evaluating the ground state energy per spin @xmath106 up to @xmath107 we
get @xmath108 here , the first term corresponds to the bare energy per spin listed in table [ tab1 ] .
since the ffst can be mapped onto an identical ffst with @xmath109 by a @xmath110-twist of the plaquettes around the tube , one expects that @xmath111 , which is obviously fulfilled .
figure [ fig3 ] shows @xmath112 to be monotonously decreasing with @xmath22 . along the line @xmath101
, the energy gain is smallest .
speaking differently , along its maximally frustrated direction in parameter space the energy landscape is flattest .
( a ) comparison between @xmath100 from plain se ( red solid lines ) with dmrg for @xmath113 and @xmath114 ( blue crosses ) , along @xmath115 .
black dashed lines refer to three dlogpad approximants to eqn .
( [ eq:7 ] ) . for reference the ground state energy of the 1d hafm
is also depicted ( green dash - dotted ) .
( b ) comparison of plain se ( red solid lines ) with dmrg ( blue crosses ) along @xmath101 . up to the first - order transition discussed in section [ ii
] , the agreement is excellent . ] in fig .
[ fig4 ] we asses the accuracy of eqn .
( [ eq:7 ] ) in several ways .
first , in panel ( a ) we compare @xmath100 with dmrg calculations along @xmath115 where we expect no first - order transition .
the dmrg results have been obtained for @xmath113 and @xmath114 .
obviously the agreement between se and dmrg is very good for @xmath116 . in panel
( b ) we display a similar comparison along @xmath101 .
here the se agrees very well with dmrg to even larger values of @xmath22 , however only up to the first - order transition discussed in section [ ii ] .
third , in panel ( a ) we also include three dlogpad approximants to eqn .
( [ eq:7 ] ) , which , similar to the dmrg , start to depart from the se for @xmath117 and may be used to increase the interval of confidence of the se only slightly by @xmath118 as evident from the dmrg results . for reference panel ( a ) also depicts the ground state energy per site of the 1d hafm with exchange constant @xmath18 which is @xmath119 .
evidently , in the parameter range we consider , the ffst is far away from the decoupled chain limit .
now we turn to the dispersion of one - particle eigenstates . by the latter we mean
excited eigenstates of the effective hamiltonian which , apart from having fixed @xmath93 , lattice momentum @xmath120 , and total spin quantum numbers @xmath121 , are linear combinations of _ single _ entries of table [ tab1 ] only .
@xmath81 eigenstates are single particle states by construction . for @xmath122 ,
@xmath93-conservation does not protect single states in table [ tab1 ] from decay into _ two _ states by virtue of @xmath92 .
eg . a @xmath123 singlet with @xmath97 at site @xmath124 could decay into two @xmath125 triplets with total spin @xmath126 and @xmath127 at sites @xmath128 ( see also appendix [ app1 ] ) . for the ffst however , and for @xmath97 , @xmath129 and up to @xmath107
, we find that all matrix elements of @xmath92 inducing such decay _ vanish _ identically .
this feature can be traced back to the @xmath130 symmetry of the tube .
in fact , we obtain that changing e.g. the exchange couplings @xmath23 into @xmath131 on one of the legs , renders the one - particle @xmath97 states unstable against decay . in summary , for each momentum @xmath120 the @xmath97 spectrum contains three genuine one - particle levels with @xmath129 , all of which are degenerate in @xmath84 . for the remaining _
two_-particle states with @xmath97 we refer to the next section . for the rest of this section we label the one - particle states by @xmath132 , where @xmath133 refers to the plaquette s site and @xmath134 is the total spin , where the index @xmath62 is only due to the fact that for @xmath97 there are two @xmath135 states , say , @xmath136 .
spin - z quantum numbers @xmath84 are discarded because of @xmath26 invariance . due to @xmath93-conservation and
@xmath26 invariance the effect of @xmath137 on @xmath132 is limited to @xmath138 which implies a shift in real space , and potentially a mixing of states of equal @xmath73 with identical @xmath93 .
the hopping amplitudes @xmath139 do not depend on @xmath140 due to translation invariance .
therefore , by fourier transformation @xmath141 we get the dispersion from @xmath142 where , obviously @xmath143 for @xmath144 .
however for @xmath145 , i.e. for the two @xmath97 , @xmath135 states , we find @xmath146 . in turn , @xmath147 is @xmath148(x)$ ] for @xmath134-[off]diagonal transitions . to obtain hopping amplitudes valid to @xmath105 , in the thermodynamic limit , the @xmath139 and the ground state energy @xmath149 in eqn.([eq:9 ] ) have to be evaluated on clusters with open boundary conditions ( obc ) , large enough to incorporate @xmath103-th order graphs for hopping processes of distance @xmath133 @xcite . depending on @xmath133 ,
these clusters are of either of length @xmath103 or @xmath150 .
we have calculated _ analytic _ expressions @xcite for @xmath151 to @xmath107 in @xmath22 . for @xmath152 eqn .
( [ eq:9 ] ) is already diagonal in @xmath134 , @xmath153 with eigenvalues @xmath154 . only for @xmath155 eqn .
( [ eq:9 ] ) displays a @xmath156-matrix structure , referring to @xmath136 , with eigenvalues @xmath157 and @xmath158 . in fig .
[ fig5 ] we show the one - particle dispersions for selected values of @xmath22 .
this figure displays only bare se results and no pad extrapolations .
several comments are in order .
first , the figure does not only display @xmath159 for all @xmath73 and @xmath160 , but for curiosity also the quintet with @xmath161 , assuming that the latter does not decay into multi - particle states - which we have not checked .
genuine one - particle dispersions on the ffst for @xmath160 at various values of @xmath162 .
these comprise three triplets ( red circle , green triangle up , green triangle down ) and one singlet ( blue diamond ) .
the @xmath93=3-quintet ( magenta x ) has not been tested against multi - particle decay and is only shown for reference .
dmrg results at zone - boundaries are shown with large crosses . ] as for the ground state energy , since exchanging @xmath109 maps the ffst onto an equivalent one by a @xmath110-twist there are related symmetries of the one - particle dispersions . these respect
the additional fact , that the single - particle states on the bare plaquette are the eigenstates of a four - site spin-@xmath0 chain with pbc , which carry one out of four momenta @xmath163 with @xmath164 transverse to the ffst . in turn , exchanging @xmath109 maps the one - particle dispersions onto identical ones up to a shift of the brillouin zone by one out of @xmath165 , and may also exchange the dispersion branches for the degenerate bare @xmath166 states .
we have checked this to be fulfilled by all @xmath159 .
, for @xmath167 we have @xmath168 and for @xmath169)$ ] we have @xmath170}^{\textrm{1pt } } ( k , j_{1},j_{2 } ) = e_{q,1b[a]}^{\textrm{1pt } } ( k+[-]\pi/2,j_{2},j_{1})$ ] .
additionally @xmath170}^{\textrm{1pt } } ( k , j_{1},j_{2})$ ] are degenerate at @xmath115 . considering the elementary triplet dispersion @xmath171 at @xmath115 , we observe a zone - boundary gap in fig .
[ fig5 ] ( a ) which decreases as @xmath18 increases , see panel ( b ) . for four - leg ladders without frustration ,
i.e. @xmath115 , the analytic expression for the gap we get is @xmath172 which , in passing , improves earlier se to @xmath173 on _ unfrustrated _ four - leg ladder @xcite by three orders .
the dispersion @xmath171 shows rather strong oscillations in panel ( b ) .
these are not related to convergence issues of the se , but are robust features of hopping to more than only nearest - neighbors .
increasing @xmath19 from zero , at fixed @xmath18 , as from fig .
[ fig5 ] ( a ) to ( c ) the bandwidth of @xmath171 is reduced . on the line of maximum frustration , i.e. at @xmath101 in panel ( d ) , it is only very small , albeit _ not _ exactly zero , namely @xmath174 .
a similar tendency can be observed in all other @xmath175 sectors , yet less pronounced . to assess the quality of the se
, we have also calculated the elementary @xmath167 gap by dmrg .
the results are shown by the large crosses at the zone - boundaries in fig .
as is obvious , the agreement is very good in all four panels .
this is particularly noteworthy for panel ( b ) , corroborating the preceding statement , that the oscillations in @xmath171 for larger values of @xmath19 are robust features of the se and no convergence issue .
[ ] 2 & @xmath176 + 1 & @xmath177 + 0 & @xmath178 + in this section we focus on two - particle states generated from two one - particle triplets out of the @xmath167 sector .
apart from generating a continuum of states at fixed total momentum with respect to the relative momentum , interactions may lead to additional collective ( _ anti_)_bound states _ which split off from the continuum .
the existence of such collective states is know for simpler 1d quantum magnets , such as chains and two - leg ladders @xcite .
here we show that similar states exist also on the ffst . to this end
we evaluate the two - particle spectrum following ideas developed for dimer - se of two - leg ladders @xcite . for the sake of completeness technical details of this procedure
are revisited in appendix [ app1 ] . on the ffst
, the @xmath97 sector allows for one - particle _ and _ two - particle states .
this is different from dimer systems where only two - particle states occur for @xmath97 @xcite .
the real - space representations of all @xmath97 states with proper spin quantum numbers @xmath121 are listed in table [ tab2 ] . here
the genuine two - particle states are labeled by @xmath179 , where @xmath180 refers to sites on the 1d lattice and @xmath121 to total spin , and spin-@xmath68 quantum numbers .
as discussed in the preceding section the one- and two - particle excitations at @xmath97 do not mix .
therefore we focus on the genuine two - particle excitations hereafter . because of translational invariance
the two - particle states can be classified according to a center - of - mass momentum @xmath181 and a relative distance @xmath182 @xmath183 these states have an exchange parity @xmath184 for @xmath185 , i.e. @xmath186 . the effective hamiltonian , eqn .
( [ eq:5 ] ) , will mix states at different @xmath182 .
the two - particle spectrum is obtained from the sum of the one- and two - particle irreducible hamiltonians ( see@xcite and appendix [ app1 ] ) , labeled by @xmath187 and @xmath188 , with _ analytic _
matrix elements calculated by se from the states of eqn .
( [ eq:10 ] ) .
we have evaluated these matrix elements to @xmath189 . for each @xmath181
, this leads to an eigenvalue problem with a rank set by the number of lattice sites ( or relative momenta ) @xmath102 .
this can be diagonalized numerically on finite , but large lattices .
see appendix [ app1 ] for definitions and formal details .
two - particle density of states close to the zone boundary at @xmath190 on an @xmath191 site ffst for @xmath192 .
delta functions at each two - particle energy broadened by @xmath193 .
( a ) _ reducible _ spectrum for s=0 and s=1 .
( b ) difference between bare and _ reducible _ spectrum for s=0 and s=1 .
( c ) complete spectra for s=0,1,2 . ] in fig .
[ fig6 ] we show the two - particle spectrum @xmath194 at fixed total momentum @xmath181 and spin @xmath121 , summed over the relative momentum @xmath195 , where @xmath196 are the two - particle eigenenergies .
we choose @xmath162 = @xmath197 and @xmath198 as an example .
to analyze the effects of the interaction we consider @xmath196 at various levels of approximation .
first , we set the irreducible two - particle interactions @xmath188 to zero - i.e. no actual interactions occur .
the corresponding spectrum is shown in fig .
[ fig6 ] ( a ) for @xmath126 and @xmath135 .
these spectra are indistinguishable on the scale of this plot .
the van - hove singularities are related to the extrema of the one - particle dispersion in the @xmath167 sector as in fig .
even though the irreducible one - particle interaction @xmath187 will act only on one of the two particles , this does _ not _ imply , that the spectra in fig .
[ fig6 ] ( a ) are identical to the convoluted _ bare _ spectra of two _ single _
@xmath199-particles at fixed total momentum @xmath181 and finite @xmath102 .
this is due to the hard - core constraint which forbids double occupation of plaquettes by @xmath199-particles .
this constraint is encoded in the matrix elements of @xmath187 when acting on the two particle sector .
since there are @xmath102 double occupancies within @xmath200 two - particle states , their removal amounts to a @xmath201 effect on the integrated spectral weight at fixed total momentum @xmath181 . at fixed total and relative momentum
, this effect will be _ quantitatively _ identical for two - particle states of identical exchange parity @xmath202 .
for this reason figs .
[ fig6 ] ( a ) and ( b ) contain only @xmath126 and 1 sates .
figure [ fig6 ] ( b ) depicts the difference between the _ bare _ spectra and that obtained from @xmath187 for @xmath126 and @xmath135 for @xmath191 .
clearly the integrated spectral weight for both values of @xmath73 is finite but of @xmath203 at most .
differences are visible between the @xmath126 and @xmath135 spectrum - albeit very small .
figure [ fig6 ] ( a ) shows that _ no _ ( anti-)bound state arises in the two - particle sector solely due to the hard - core repulsion . diagonal element of the irreducible interaction @xmath204 versus relative distance of two particles for @xmath129 , and 2 , and for @xmath205 and @xmath206 , multiplied by @xmath207 . ( the matrix elements of @xmath188 at @xmath189 are zero for @xmath208 . ) ] next we consider the complete @xmath97 spectrum in fig .
[ fig6 ] ( c ) for @xmath209 and @xmath210 , including the irreducible two - particle interaction @xmath188 .
the latter is _ spin - dependent _ and _ longer - ranged_. this figure clearly demonstrates a main point of this section , i.e. the occurrence of three split - off ( _ anti_)-_bound _ states ( above ) below the continuum depending on their spin .
apart from that , and similar to the hard - core constraint @xmath188 leads to an @xmath203 redistribution within the two - particle continuum .
the total momentum @xmath181 in fig .
[ fig6 ] has been chosen , such as to evidence the ( anti)-bound states clearly .
depending on @xmath181 they can approach and also merge with the continuum
. a qualitative understanding of binding versus anti - binding of the collective two - particle states can be obtained from considering the diagonal matrix elements @xmath211 of two particles versus their relative distance @xmath182 at fixed total momentum @xmath181 .
this matrix element is @xmath212 and independent of @xmath84 .
it is shown in fig .
[ fig7 ] for @xmath22 identical to that of fig .
first , this figure demonstrates that the irreducible two - body interactions are very rapidly decaying as a function of the relative distance of the particles ( note the scaling of the matrix elements by @xmath207 on the y - axis ) .
second , and for both total momenta depicted , the interactions for @xmath135 are attractive , while those for @xmath213 are repulsive .
this is consistent with binding versus anti - binding of the collective states in fig .
yet , one should keep in mind the matrix elements @xmath214 which are off - diagonal in @xmath215 .
we have not tested if these invalidate the simplified argument given here . finally , we emphasize , that the ordering of anti - bound versus bound states on the ffst is different from that on plain two - leg ladders . in the latter case two - particle singlets also form _
bound _ states with a binding energy larger than that of the two - particle triplets .
momentum resolved spectrum of all states on the ffst with @xmath160 for @xmath192 and @xmath216 .
apart from four genuine one - particle states , i.e. three triplets ( black dashed , blue solid , green solid ) , and one singlet ( black solid ) , there are three two - particle continua with s=0,1,2 , and three collective two - particle states : a bound triplet ( magenta dots ) , an anti - bound singlet ( black dots ) and an anti - bound quintet ( red dots ) .
only @xmath217% of the scattering states between the lower and upper bounds of the continua are depicted to avoid graphic cluttering .
dmrg results for one and two - particle states at zone - boundaries(centers ) are depicted with large crosses . ] in fig .
[ fig8 ] we show the dispersion of all one and two - particle states from the @xmath81 and @xmath97 sector as a function of total momentum for two sets of @xmath22 , obtained by se , together with dmrg results for one and two - particle states at the zone - boundary(center ) .
this figure is another main result of our paper and summarizes several aspects .
first , the spectrum is very rich and consists of various discrete states and three superimposed two - particle continua .
this should be contrasted against plain two - leg ladders , which show a less involved low - energy spectrum @xcite .
some of the discrete states are genuine one - particle states and some are collective two - particle states . for the parameters depicted , the collective two - particle states are clearly visible only close to the zone boundary .
the potential existence of critical parameters @xmath218 or wave - vectors @xmath219 for which the collective states merge with the continua is unclear at present .
the @xmath97 one - particle states are contained almost completely within the spectral range of the continua and split off from the latter only in a range of @xmath181 similar to the collective states .
this situation is also very different from two - leg spin - ladders in which the states which split off from the continuum close to the zone boundary are collective ( anti)-bound states only . from an experimental point of view
this may pose a challenge on discriminating between such genuine one - particle and collective two - particle states .
finally , the effects of frustration are clearly visible in going from fig .
[ fig8 ] a ) to b ) : the complete spectrum shows a tendency to localize .
however , in contrast to the one - particle dispersions the two - particle interactions are _ not _ reduced approaching the line of maximum frustration , @xmath54 . in turn
the relative splitting of the collective states from the continuum is enhanced by frustration . to conclude this section we provide some measure of convergence of the se for the two - particle continuum in fig .
this figure shows the relative change in energy of all two - particle states at a fixed total momentum versus their energy when switching from an @xmath189 to an @xmath107 evaluation of the matrix elements of @xmath220 . for this
we confine ourselves to @xmath126 .
as is obvious , the changes are completely negligible and fully justify the use of @xmath189 se for the two - particle states .
we note in passing , that @xmath107 se will not only lead to longer - range irreducible one - particle hoppings , but also increase the range of the irreducible two - particle interactions in fig .
[ fig7 ] , retaining however the rapid decay with distance depicted there .
to summarize , we have studied the strong rung - coupling regime of the frustrated four - spin tube .
remarkably , this spin model displays several twists with respect to conventional two - leg spin ladders .
it allows for frustration induced quantum - phase transitions into phases other than the plaquette singlet phase .
the nature of these phases remains to be explored .
the structure of the excitations is both richer , and different as compared to two - leg ladders , including a reordering of two - particle ( anti)bound states and additional elementary excitation modes , some of which can be made to decay by reducing the symmetry of the ffst .
finally geometric frustration on the ffst introduces a control parameter allowing for almost localization of excitations and a flattening of the energy landscape .
as compared to potential material realizations of the ffst in cu@xmath6cl@xmath8@xmath10d@xmath11c@xmath8so@xmath6 , several extensions have to be considered in the future . these comprise e.g. reducing the @xmath130-symmetry and considering the weak rung - coupling limit .
we thank d. c. cabra , a. v. chubukov , and o. starykh for helpful discussions .
m. a. has been supported by conicet ( coop .
r.2049/09 and pip 1691 ) and by anpcyt ( pict 1426 ) .
w. b. thanks dfg for financial support through grant number 444 arg-113/10/0 - 1 .
part of this work has been performed at the kavli institute for theoretical physics and within the advanced study group at the max planck institute for the physics of complex systems , which we would like to thank for their hospitalities .
the research at kitp was supported in part by the national science foundation under grant no .
nsf phy05 - 51164 .
relative error in all two - particle energies for an ffst of 200 plaquettes at fixed total momentum @xmath181 for switching from @xmath189 to @xmath107 se .
( the apparent two - curve structure is due to an oscillation of the error . ) ]
in this appendix we list some of the details necessary to evaluate the two - particle matrix elements by se .
first , we refer to the parity of the states in eqn .
( [ eq:10 ] ) .
because of this we may confine ourselves to @xmath221 ( @xmath222 , i.e. @xmath223 is forbidden ) .
the matrix elements of the one(two)-particle irreducible hamiltonian @xmath224 are obtained by subtracting the one(two)-particle reducible contributions from the matrix elements of @xmath92 @xcite @xmath225 @xmath226 here @xmath227 , @xmath228 and @xmath229 refers to the cluster on which these matrix elements are evaluated . for @xmath230 and @xmath231 to be size consistent
, the cluster has to be the largest linked cluster for a given one- and two - particle state in real space at a given order @xmath103 of the se .
the @xmath232 in eqn .
( [ eq : a2 ] ) refers to @xmath230 evaluated on the _ same _ linked cluster @xmath229 as the @xmath233 corresponding to a given @xmath231 and @xmath234 is the ground state energy of that cluster .
the two - particle spectrum results from the eigensystem of@xmath235 which , due to translational and spin - rotational invariance is diagonal in @xmath181,@xmath73 , and @xmath84 and is independent of @xmath84 . for each @xmath236 the two matrices on the left of eqn .
( [ eq : a3 ] ) are hermitian with , matrix - indices @xmath237 .
thus , for the remainder we may consider only @xmath238 .
we begin with @xmath239 . because a linked cluster can have at most @xmath150 sites , @xmath240 for @xmath241 .
i.e. @xmath242 is an @xmath243 matrix .
the action of @xmath188 is @xmath244 here @xmath62 must be restricted to @xmath245 , because of the sites @xmath246 to reside within a linked cluster at order @xmath103 .
i.e. @xmath62 is confined to the interval @xmath247 , which has its midpoint at @xmath248 @xmath249\ } next , we apply @xmath187 to a two - particle state .
this will only shift one of the particles , i.e.@xmath253 for each @xmath181 , @xmath187 connects states @xmath182 and @xmath254 .
this is equivalent to a band - matrix of width @xmath255 . for @xmath256 ,
the non - zero content of the columns of this matrix is independent of @xmath182 , for @xmath257 this is not so .
the latter relates to the exclusion of on - site double - occupation .
linked clusters with size @xmath258 and with initial ( half circles ) and final ( wedges ) two - particle states for all addend in eqn .
( [ eq : a2 ] ) at order @xmath12 and for @xmath259 .
( i)r labels ( ir)reducible graphs . ] for reference we explicitly display one specific low - order matrix element @xmath260 from eqns .
( [ eq : a3 ] , [ eq : a5 ] , [ eq : a6 ] ) , namely @xmath261 , @xmath12 and @xmath126 .
figure [ fig10 ] relates to the two - particle irreducible matrix element from @xmath188 , and depicts all @xmath231 on their respective clusters including a label for reducing contributions to be evaluated according to eqns .
( [ eq : a1 ] , [ eq : a2 ] ) . these are @xmath262 for @xmath263 and @xmath264 for @xmath265 for this particular matrix element .
note that on finite clusters @xmath266 and @xmath267 are not necessarily identical .
the one - particle irreducible matrix element from @xmath187 is straightforward .
the complete matrix element reads @xmath268 finally , in the @xmath97 sector the single - particle states @xmath269 , and @xmath270 need to be considered .
any one - particle irreducible hamiltonian matrix element between the two- and one - particle states for @xmath97 is zero due to orthogonality .
two - particle irreducible contributions require matrix elements of the form@xmath271 where @xmath272 is a one - particle state of spin quantum numbers @xmath121 with @xmath97 .
this implies a _ decay _ of @xmath97 single - particle states into @xmath97 two - particle states . _ up to the order that we have performed the se , we have not observed such decay . _ in turn the hamiltonian matrix in the @xmath97 sector of the four - tube is already diagonal w.r.t . the particle number . | we study the magnetism of a frustrated four - leg spin-@xmath0 ladder with transverse periodic boundary conditions : the frustrated four - spin tube ( ffst ) . using a combination of series expansion ( se ) , based on the continuous unitary transformation method and density - matrix renormalization group ( dmrg ) we analyze the ground - state , the one- , and the two - particle excitations in the regime of strong rung - coupling .
we find several marked differences of the ffst with respect to standard two - leg ladders .
first we show that frustration destabilizes the spin - gap phase of the ffst which is adiabatically connected to the limit of decoupled rung singlets , leading to a first order quantum phase transition at finite inter - rung coupling .
second , we show that apart from the well - know triplon branch of spin - ladders , the ffst sustains additional elementary excitations , including a singlon , and additional triplons .
finally we find , that in the two - particle sector the ffst exhibits collective ( anti)bound states similar to two - leg ladders , however with a different ordering of the spin - quantum numbers .
we show that frustration has significant impact on the ffst leading to a flattening of the ground - state energy landscape , a mass - enhancement of the excitations , and to a relative enhancement of the ( anti)binding strength . where possible we use dmrg to benchmark the findings from our se calculations , showing excellent agreement . |
in a zinc blende semiconductor the hole wave function originates from atomic @xmath9 orbitals resulting in an angular momentum @xmath10 . in the long wavelength approximation ,
the effective luttinger hamiltonian for holes is quadratic in the hole momentum @xmath12 @xcite ( see also ref .
@xcite @xmath139 1 , 2 , 3 are the crystal axes of the cubic lattice , @xmath140 is the electron mass , @xmath141 denotes the anticommutator , and @xmath142 , @xmath143 and @xmath144 are luttinger parameters . in gaas @xmath145 , @xmath146 , @xmath147 @xcite .
the hamiltonian ( [ lut ] ) can be rewritten as @xmath148 where @xmath149 the irreducible 4@xmath150 rank tensor @xmath151 depends on the orientation of the cubic lattice , the tensor is proportional to @xmath152 . neglecting @xmath152 compared to @xmath143 , the luttinger hamiltonian can be approximated by the following rotationally invariant ( independent of the lattice orientation ) hamiltonian @xmath153 \ .\ ] ] due to the confining potential @xmath154 , motion perpendicular to the 2d plane of the heterostructure is quantized , leading to the formation of 2d subbands , where only the lowest subband occupied in the low - temperature experimental regime . assuming a square well confining potential of width @xmath155 we have @xmath156 . since @xmath157
, we may expand @xmath158 , with the leading term becoming diagonal in a basis of states with @xmath159 .
due to the sign of the interaction , states with @xmath160 ( heavy hole ) are lower in energy , and the splitting between these and states with @xmath161 ( light hole ) at @xmath162 becomes @xmath163 here we take @xmath164 nm .
the splitting between the lowest and the next heavy hole band at @xmath162 is @xmath165 numerical diagonalization of the full luttinger hamiltinian ( [ lut ] ) using the nextnano++ package @xcite gives the energy levels ( 2d dispersions ) plotted in fig .
[ hh12 ] .
the hh1-hh2 splitting at @xmath162 is pretty close to ( [ hsplit ] ) while the hh1-lh1 splitting in fig .
[ hh12 ] is somewhat smaller than ( [ ssplit ] ) because of the @xmath166 tensor corrections . at hole density corresponding to our experiment , @xmath167 , only the lowest band
is populated .
the lowest band dispersion @xmath168 enters eq .
( [ heff ] ) .
we describe this band by the effective spin @xmath17 , @xmath18 , @xmath19 .
the fermi momentum is @xmath169 and the fermi energy , @xmath170mev , is shown in fig .
[ hh12 ] by the red horizontal line .
the heavy hole effective mass , @xmath46 , follows from fig .
[ hh12 ] . at @xmath171
the mass is about @xmath172 and at @xmath173 the mass is @xmath174 .
obviously only the latter mass is relevant to our analysis .
the off - diagonal part of @xmath175 in the hamiltonian ( [ hl ] ) , @xmath176 , leads to heavy - light hole mixing .
@xmath177e^{i{\bm k}\cdot{\bm r}}\nonumber\\ & & |{\bm k},\downarrow\rangle=\left[|-\frac{3}{2}\rangle + a k_-^2|+\frac{1}{2}\rangle \right]e^{i{\bm k}\cdot{\bm r}}\nonumber\\ & & a=\frac{\sqrt{3}\gamma_2 } { 2m_e\delta_{hl}}= \frac{\sqrt{3}}{4\langle k_z^2\rangle}\ .\end{aligned}\ ] ] taking the square well width @xmath164 nm and the hole density @xmath178 , we arrive at the following estimate for the mixing probability , @xmath179 .
this very small mixing , of order @xmath180 in probability , is responsible for the soc considered here .
the zeeman interaction of a @xmath10 hole with magnetic field @xmath107 is @xcite , @xmath181 where @xmath182 . taking the matrix element of @xmath183 between states eq.([se ] ) we find the effective matrix of @xmath37 @xmath184 comparing this with @xmath37 in eq.([heff ] ) we determine the coefficient @xmath32 in this equation to be @xmath185 according to our fit of sdh data @xmath186 .
hence we find that @xmath187 and the probability of the heavy - light hole mixing is @xmath188 , which is remarkably consistent with the estimate presented after eq.([se ] ) .
it is worth noting that eq.([al ] ) is approximate , since one should expect a comparable contribution to @xmath32 which is not accounted for by the calculations presented .
so far , we have neglected the coupling to the vector potential created by @xmath189 , @xmath190 .
this coupling also gives a contribution to the coefficient @xmath32 , see refs .
this contribution is highly sensitive to the exact shape of the confining potential and therefore can not be reliably calculated @xcite .
the kinematic form of @xmath37 however remains unambiguous and we can fit the value of @xmath32 to the experimental data .
the @xmath191d hole system resides within a symmetrically doped @xmath192nm - wide gaas / al@xmath86ga@xmath87as quantum well , grown on the low symmetry plane @xmath193 by molecular beam epitaxy .
a heavily doped @xmath194 gaas layer located @xmath195 m below the quantum well , acts as an _ in situ _ back gate , allowing the 2d density to be tuned @xcite . at zero back - gate voltage , the density of the 2d hole system is @xmath196 with a corresponding mobility of @xmath197 .
transport measurements were performed in a kelvinox @xmath198 dilution refrigerator within the bore of a @xmath199 t magnet at a base temperature of @xmath200mk , using standard lock - in techniques , with a constant ac current of @xmath201na at a frequency of @xmath202hz . to perform tilted field measurements ,
the sample was mounted on a piezo - electric rotator which allowed for _ in situ _ rotation to be conducted with an accuracy of @xmath203 @xcite .
initially the 2d device was rotated to @xmath204 , so the magnetic field lies perpendicular to the sample plane , @xmath205 , @xmath206 , and the sample orientation confirmed by measuring the hall plateaus as a function of perpendicular field , shown in fig.[m1 ] ( blue ) .
the corresponding low - field oscillating longitudinal resistivity is shown in ( red ) , with spin - splitting appearing for @xmath207 .
for the purposes of our analysis we are only interested in low field data between @xmath208 t and @xmath209 t .
the coincidence method using tilted fields was first pioneered by fang and stiles in 1968 @xcite to study the land g - factor in 2d electron systems . here
we perform a similar set of tilted field transport measurements for a 2d hole system , taken along two crystal directions : the high symmetry @xmath210 $ ] and the low symmetry @xmath211 $ ] , as depicted in fig.[fig2]a . to achieve this ,
the device was first mounted on the rotator such that it tilts between the crystal axes @xmath212 $ ] and @xmath213 $ ] , where the 2d plane is fully perpendicular to the field at @xmath214 .
the sample was then rotated towards the @xmath210 $ ] direction till @xmath215 to introduce a parallel field component @xmath216 , and the total field @xmath107 swept , changing the sign of the in - plane field @xmath217 .
this procedure was repeated for a number of different @xmath218 with increasing in - plane field components .
the experiment was then repeated for equivalent @xmath219 and the results plotted in fig.[fig2]b . during a second cooldown , the sample was re - oriented to perform tilted measurements along the @xmath212 $ ] and low symmetry @xmath211 $ ] crystal axes .
the experiment was then repeated for both @xmath220 and the results shown in fig.[fig2]c .
the electric potential across the quantum well was tuned via the _ in situ _ back gate , to adjust the confining potential .
the presence of the rashba soc results in beatings of the sdh oscillations even without any tilting of the magnetic field @xcite .
the rashba interaction is sensitive to the back - gate voltage ( @xmath221 ) , so by varying the applied bias - voltage , we can tune the system to minimize the amount of beatings and hence to eliminate the rashba interaction .
fig.[m2 ] shows these beatings in detail , where the sdh oscillations at each back gate voltage are periodic in @xmath222 and the amplitudes of these oscillations normalized for clarity by multiplying the datasets by @xmath223 to remove the envelope .
the data are taken without any tilting , @xmath214 . from fig.[m2 ] we select @xmath224 as the final operating point with the least amount of beating in the sdh oscillations . we will show that the major part of the dresselhaus interaction does not influence dynamics at @xmath214 . hence , minimizing the beating we tune the rashba interaction to be close to zero .
this back - gate voltage is used as the operating point for the rest of the experiment . at this point
the carrier density is @xmath225 and the mobility is @xmath226 .
the comparison of the experimental result with possible theories is presented in fig.[fig3 ] . panels
a and b show the non - abelian theory and panels c and d show the abelian theory .
the non - abelian theory is consistent with experiment while the abelian theory is not consistent .
since the conclusions are based on our fits , a natural question which arises is ` how sensitive is the number of coincidences with respect to variation in our fitting parameters ? ' in fig.[s1 ] we show the response of the non - abelian prediction , eq.(7 ) , as the fitting parameters are varied .
the layout and colour scheme are similar to fig.3 : the top panels show the theoretical phase and the experimental coincidence points for orientations of the external field @xmath227 ( red ) and @xmath228 ( green ) .
the bottom panels show the phase and the experimental coincidence points for the field orientation @xmath121 ( blue ) .
solid lines in fig.[s1 ] are identical to that in fig.3a , b . in fig.[s1 ] panels a & b correspond to @xmath229 variation of @xmath230 , the c & d panels correspond to @xmath231 variation of @xmath88 , and the e & f panels correspond to @xmath229 variation of @xmath34 . from these plots ,
the presented deviations are larger than those accepted in fig.3 .
the curves corresponding to the lower boundaries of the parameters ( dotted lines ) are too far away from the experimental points . on the other hand the curves corresponding to the upper boundaries of the parameters ( dashed lines )
demonstrate an additional coincidence point ( panels b , d , f ) which is not observed experimentally .
this shows that the selected parameters , @xmath232 , provide the best fit to the data .
the curves in fig.[s1 ] do not account for the dresselhaus interaction .
there is no point to account for the interaction for purposes of the present analysis , since it hardly effects the red and green curves which are used to determine the fit parameters , and it does not change the number of coincidences .
dresselhaus only deforms the blue curves in panels b , d , and f exactly in the same way as in the panel b of fig .
3 . + a similar comparison for the abelian theory , eq.(8 ) , is presented in fig.[s2 ] , where once again the parameters @xmath233 are varied .
the top panels show the theoretical phase @xmath85 and the experimental coincidence angles for orientations of the external field @xmath234 ( red ) and @xmath228 ( green ) .
the bottom panels show the phase and the experimental coincidence points for the field orientation @xmath235 ( blue ) .
solid lines in fig.[s2 ] are identical to that in fig.3c and 3d .
[ h ! ] in fig.[s2 ] panels a and b correspond to @xmath231 variation of @xmath230 , panels c and d correspond to @xmath231 variation of @xmath88 , and panels e and f correspond to @xmath229 variation of @xmath34 . despite the significant amount of variation in these parameters , the abelian theory always predicts three concidence angles for @xmath236 ( blue ) , whilst experimentally only one angle is observed .
this discrepancy illustrates that the experimental data can not be reconciled with the paradigm of berry phases alone and renders our evidence for the non - abelian gauge field unambiguous .
dresselhaus spin - orbit interaction arises due to the lack of inversion symmetry in the bulk gaas crystal .
the interaction is cubic in the momentum @xmath35 and linear in the angular momentum @xmath237 , see ref .
@xcite . in the coordinate system defined in fig.[fig2]a
the leading term of the dresselhaus hamiltonian is @xmath238 where @xmath239ev@xmath240 , see ref.@xcite .
we neglect the subleading terms cubic in @xmath241 . the transformation ( [ b1 ] ) to the co - rotating frame does not change ( [ dh3 ] ) .
hence , in the co - rotating frame the dresselhaus interaction works as a weak periodic `` magnetic field '' superimposed on the constant `` magnetic field '' defined by eq.([po ] ) .
the projection of the periodic `` field '' on the direction perpendicular to the direction of the constant `` field '' , @xmath242 generates spin flips . accounting for the resonant part of the periodic perturbation one finds that @xmath66 given by eq.([prefactor ] ) is replaced by @xmath243 , @xmath244 where @xmath245 without tilting , @xmath214 , the dresselhaus perturbation ( [ dh2 ] ) is zero , @xmath246 .
hence , our back - gate tuning performed at @xmath214 does not compensate the dresselhaus interaction .
the interaction becomes important at intermediate values of @xmath98 .
the dashed curves in figs .
3a , b display eq.([prefactor1 ] ) calculated with @xmath247 .
we chose this value of @xmath248 to shift the coincidence angle in fig.3b from @xmath126 to @xmath249 .
this is an additional fitting parameter . because of weakness of the dresselhaus interaction the dashed and the solid curves in fig.3a are practically indistinguishable . on the other hand , because of the resonance
, the effect of the dresselhaus interaction in fig.3b is significant .
we stress again the point made in the main text , the correct value of @xmath248 necessarily leads to the extended range of @xmath98 over which the phase of the sdh oscillation inverts , clearly seen in the blue traces in fig 2c .
the value of @xmath248 can be also calculated using eq.([d0 ] ) . with @xmath250tesla and with parameters of the system discussed in the paper , eq.([d0 ] )
gives @xmath251 .
so , the dresselhaus interaction is weak and insignificant compared to the dominating magnetic field controlled spin orbit effects that drive the non - abelian berry phase .
the strength of the dresselhaus interaction measured in our experiment is slightly smaller but comparable with theoretical estimates presented in refs.@xcite .
99 c. n. yang , r. l. mills , _ phys .
rev . _ * 96 * , 191 ( 1954 ) .
s. l. glashow , _ nuclear phys .
_ * 22 * , 579 ( 1961 ) .
m. v. berry , _ proc .
london , ser . a _ * 392 * , 45 ( 1984 ) .
f. wilczek , a. zee , _ phys .
lett . _ * 52 * , 2111 ( 1984 ) .
a. tomita , r. y. chiao , _ phys .
lett . _ * 57 * , 937 ( 1986 ) .
yau , e. p. de poortere , m. shayegan , _ phys .
lett . _ * 88 * , 146801 ( 2002 ) .
b. grbi , r. leturcq , t. ihn , k. ensslin , d. reuter , and a.d .
wieck , _ phys .
lett . _ * 99 * , 176803 ( 2007 ) .
lin , _ et .
al . _ , synthetic magnetic fields for ultracold neutral atoms .
_ nature _ * 462 * , 628 ( 2009 ) .
j. w. zwanziger , m. koenig , a. pines , _ phys .
a _ * 42 * , 3107 ( 1990 ) . g. moore , n. read , _ nucl .
b _ * 360 * , 362 ( 1991 ) .
d. p. arovas , y. lyanda - geller , _ phys . rev .
b _ * 57 * , 12302 ( 1998 ) . s. murakami , n. nagaosa , s .- c . zhang , _ science _ * 301 * , 1348 ( 2003 ) .
p. a. lee , n. nagaosa , x .-
wen , _ rev .
phys . _ * 78 * , 17 ( 2006 ) .
j. dalibard , f. gerbier , g. juzeliunas , and p. ohberg , _ rev .
phys . _ * 83 * , 1523 ( 2011 ) .
f. gerbier , n. goldman , m. lewenstein , _
b _ * 46 * , 130201 ( 2013 ) .
g. engels , j. lange , th .
schpers , and h. luth , phys .
b * 55 * , 1958 ( 1997 ) .
b. grbc , _ et .
phys . rev .
b * 77 * , 125312 ( 2008 ) .
l. onsager , _ philos . mag .
_ * 43 * , 1006 ( 1952 ) .
k.s . novoselov _ et .
_ , _ nature _ * 438 * , 197 ( 2005 )
. j. g. analytis , _ et .
* 6 * , 960 ( 2010 ) .
j. luttinger , w. kohn , _ phys .
_ * 97 * , 869 ( 1955 ) .
i.m . lifshitz and l.m .
kosevich , _ sov . phys .
jetp - ussr _
* 6 * , 67 ( 1958 ) .
thomas , _ nature _ * 117 * , 514 ( 1926 ) .
j. c. budich , d. g. rothe , e. m. hankiewicz , and b. trauzettel , _ phys .
b _ * 85 * , 205425 ( 2012 ) . z.
k. keane _ et .
al . _ , nano lett . *
11 * , 3147 ( 2011 ) e. a. chekhovich _ et .
_ , _ nature materials _ * 12 * , 494 ( 2013 ) .
r. winkler , _ spin - orbit coupling effects in two - dimensional electron and hole systems _
( springer - verlag berlin heidelberg , 2003 ) , pp . 212 , 220 . s. birner , _ et .
al _ , ieee trans .
electron dev .
* 54 * , 2137 ( 2007 ) .
i. vurgaftman , j. r. meyer , and l. r. ram - mohan , _ j. appl
_ * 89 * , 5815 ( 2001 ) .
y. komijani _ et .
al _ , _ europhys . lett . _ * 102 * , 37002 ( 2013 ) .
m. y. simmons , _ et . al .
_ , _ appl .
* 70 * , 2750 ( 1997 ) .
l. a. yeoh , _ et .
_ , _ rev .
instrum . _
* 81 * , 113905 ( 2010 ) .
f. f. fang , p. j. stiles , _ phys . rev . _ * 174 * , 823 ( 1968 ) .
j. p. eisenstein , h. l. stormer , v. narayanamurti , a. c. gossard , and w. wiegmann , _ phys .
_ * 53 * , 2579 ( 1984 ) .
m. v. durnev , m. m. glazov , e. l. ivchenko , phys .
b * 89 * , 075430 ( 2014 ) . | gauge theories , while describing fundamental interactions in nature , also emerge in a wide variety of physical systems .
abelian gauge fields have been predicted and observed in a number of novel quantum many - body systems , topological insulators , ultracold atoms and many others .
however , the non - abelian gauge field , while playing the most fundamental role in particle physics , up to now has remained a purely theoretical construction in many - body physics . in the present paper
we report the first observation of a non - abelian gauge field in a spin - orbit coupled quantum system .
the gauge field manifests itself in quantum magnetic oscillations of a hole doped two - dimensional ( 2d ) gaas heterostructure .
transport measurements were performed in tilted magnetic fields , where the effect of the emergent non - abelian gauge field was controlled by the components of the magnetic field in the 2d plane .
gauge theories were originally conceived to describe elementary particles and their interactions @xcite .
the concept of the emergent gauge field is relevant to a wide class of quantum systems whose initial formulation has no apparent relationship to gauge fields .
such emergent gauge fields arise naturally in many geometrical contexts and the idea that physical systems can be classified according to their geometrical properties has become an overarching paradigm of modern physics .
one example of an abelian gauge theory in this context is the berry phase @xcite , which is associated with the adiabatic evolution of a nondegenerate quantum state .
the emergence of non - abelian gauge fields in degenerate quantum systems was first theoretically proposed by wilczek and zee @xcite shortly after the work of berry .
while abelian gauge fields have been observed in systems ranging from optical fibers @xcite and semiconductor rings @xcite to bose condensates of ultracold atoms @xcite , signs of non - abelian effects have so far only been observed in the nuclear quadrupole resonance of @xmath0cl in a single crystal of sodium chlorate @xcite .
non - abelian gauge fields have been theoretically predicted in a number of many - body systems including fractional quantum hall liquids @xcite , spin - orbit coupled systems @xcite , cuprate superconductors @xcite and ensembles of ultracold atoms @xcite . in spite of the theoretical excitement and great interest all previous attempts to observe these fields were unsuccessful .
this demonstrates the challenge involved in the experimental realization of emergent non - abelian gauge fields .
the idea of our experiment is partially based on previous theoretical work by arovas and lyanda - geller @xcite as well as murakami , nagaosa and zhang @xcite who proposed that effects relating to non - abelian gauge fields must be pronounced in hole - doped zinc blende semiconductors due to the strong spin - orbit coupling ( soc ) . in this context the gauge fields
are closely associated with spin dynamics along curved trajectories : ref .
@xcite proposed the use of mesoscopic rings to bend the trajectory , whilst ref .
@xcite suggested use of an external electric field for the same purpose . in this work
we use a 2d gaas hole - doped heterostructure in a relatively small ( fraction of a tesla ) magnetic field applied perpendicular to the 2d plane to curve the hole trajectories .
in addition , we apply an in - plane magnetic field ( @xmath1 several tesla ) , which allows us to control the magnitude of the spin - orbit coupling .
the combination of the soc and curved trajectories makes the non - abelian gauge field observable .
the perpendicular magnetic field gives rise to quantum magnetic oscillations which are influenced by non - abelian spin dynamics .
we measure the oscillations via the shubnikov - de haas ( sdh ) effect .
the sdh effect has been measured previously in numerous experiments with 2d systems with strong spin - orbit interaction , see e.g. refs .
@xcite however in all previous studies , effects related to the non - abelian berry phase are negligible , and what is measured is simply the densities of the spin - split subbands .
one needs very special conditions to distinguish between the abelian and the non - abelian berry phases , it is necessary to tune independently the spin precession , the orbital dynamics , and the spin - orbit interaction .
to do so in our experiment we use the following crucial points .
( i ) we can tune the spin - orbit coupling over a wide range using the in - plane field @xmath2 while keeping the orbital dynamics fixed . ( ii )
we use a low symmetry crystal with highly anisotropic coupling to @xmath2 , which allows us to control independently the larmor and the spin precession frequencies .
this is key to proving that the effects we observe can not be due to abelian physics , nor due to differences between datasets taken at different carrier densities , gate biases , or even from different samples .
( iii ) we use a device where we can minimise the undesirable rashba interaction , allowing a simple analytic theory to explain the data .
these three factors allow us to report the first observation of the non - abelian gauge phase which was elusive for 30 years since its theoretical prediction .
c c c & & + & & the spin dynamics of a particle moving around a circle in momentum space used in onsager quantization @xcite is illustrated in fig.[fig1 ] .
the three panels in this figure correspond to three qualitatively different situations : ( a ) spin dynamics being absent , ( b ) abelian spin dynamics and ( c ) non - abelian spin dynamics .
the spin ( red arrows ) is driven by a local effective magnetic field @xmath3 ( blue arrows ) , which is the sum of the external magnetic field @xmath4 and the momentum - dependent spin - orbit field @xmath5 . panel ( a ) depicts the trajectory of a non - relativistic electron in the absence of spin - orbit . in this case @xmath6 and the spin is simply aligned with the external field .
panel ( b ) illustrates the case of an ultrarelativistic dirac electron , _
e.g. _ an electron in graphene or on the surface of a topological insulator . in this case , although spin is precessing , it remains aligned with the driving field which itself is parallel to the momentum , @xmath7 .
the precession of spin around the orbit generates a geometric berry phase ( abelian gauge field ) which appears as the @xmath8-phase shift observed in magnetic oscillations @xcite .
the non - abelian case addressed in the present work is illustrated in panel ( c ) . here
, the driving field @xmath3 is not collinear with spin and the noncollinearity is proportional to the non - abelian gauge field . due to the non - abelian spin dynamics , the particle acquires a matrix - valued phase equal to the circulation of the gauge field around the trajectory in momentum space .
the phase manifests itself in quantum magnetic oscillations .
holes in gaas originate from atomic @xmath9 orbitals and hence posses an angular momentum @xmath10 . the electric quadrupole interaction leads to strong coupling between the angular momentum @xmath11 and the linear momentum @xmath12 , which is described by the luttinger hamiltonian @xcite .
the @xmath13-confinement in a 2d heterostructure enforces quantization of @xmath11 along the @xmath13-axis .
therefore , a hole quantum state with a given in - plane momentum @xmath14 splits into two doublets with @xmath15 ( heavy holes ) and @xmath16 ( light holes ) .
since light holes lie significantly higher in energy , we shall only consider heavy holes for the low energy dynamics .
the heavy - hole kramers doublet can be described by an effective spin @xmath17 , @xmath18 , @xmath19 .
the hamiltonian describing heavy holes consists of the kinetic energy , the zeeman interaction and the soc , @xmath20 , @xmath21 , \nonumber\end{aligned}\ ] ] where @xmath22 ; @xmath23 , @xmath24 , @xmath25 ; @xmath26 is the in - plane vector potential created by @xmath27 , @xmath28 is the elementary charge , @xmath29 are pauli matrices describing the spin , @xmath30 is bohr magneton , @xmath31 is the effective g - factor and @xmath32 is the soc strength . note that due to mixing between heavy holes states the dispersion @xmath33 can significantly differ from the simple quadratic form , see discussion in appendix [ apd ] .
note also that generally @xmath34 and @xmath32 depend on @xmath35 , and in combination with nonquadratic dispersion @xmath33 this dependence results in a very complex fan diagram of landau levels .
however , according to the landau theory of normal fermi liquids this complexity is irrelevant to the problem we address .
we do not need the full landau level fan diagram . according to normal fermi liquid theory
only the values of the parameters at the fermi energy are relevant .
this statement is very general , and even includes hole - hole coulomb interaction effects .
we will fit the experimental data to obtain the parameters @xmath34 and @xmath32 at @xmath36 .
a derivation of the spin - orbit interaction @xmath37 is presented in appendix [ apd ] , although we shall make two comments here on its origin : ( i ) the spin - orbit coupling arises from a small mixing between heavy and light holes , where the mixing probability is 1 - 2% ( see appendix [ apd ] ) .
( ii ) the kinematic structure of @xmath37 in eq.([heff ] ) is dictated by the fact that the pauli matrices @xmath38 correspond to @xmath39 .
if the perpendicular magnetic field is zero , @xmath40 , then the hole trajectories are straight lines and @xmath37 in eq.([heff ] ) simply splits the doubly degenerate band , @xmath41 , into a pair of chiral bands . in presence of @xmath27 the hole trajectory forms a circle , @xmath42 .
semiclassically , using the wave packet picture , the angle is @xmath43 ( the sign corresponds to @xmath44 ) , where @xmath45 is the cyclotron frequency and @xmath46 is the effective cyclotron mass at the fermi energy .
the spin - orbit field @xmath47 varies along the trajectory .
this variation can be removed by a local gauge transformation of the spinor wave function @xmath48 .
taking @xmath49 we gauge out the angle dependence of the soc , @xmath50g\ , \ \ \
\boldsymbol{\beta}^{\prime }
= \alpha k^2 \boldsymbol{b}_\parallel \ .\ ] ] since our choice of @xmath51 ensures that we perform a transformation to the co - rotating frame of the hole , it follows that @xmath52 does not vary along the trajectory .
the gauge transformation results in the covariant derivative @xmath53 , where @xmath54 is the non - abelian gauge field possessing a vortex structure in 2d momentum space @xmath55 the field tensor corresponding to this gauge field is zero , @xmath56=0 $ ] . however , the gauge field has a nonzero circulation along the hole trajectory @xmath57 and this circulation reveals itself in quantum magnetic oscillations . to understand quantum magnetic oscillations we need to consider the impact of spin - orbit coupling upon the landau level structure . for this analysis , we restrict ourselves to a semiclassical approximation , where the landau levels are determined by the onsager quantization condition . consider a hole traversing the circular trajectory , where the hole is initially prepared in a polarization state @xmath58 . under the combined action of @xmath37 and @xmath59
spin will precess along the trajectory , as shown in fig.[fig1]c . after a full cycle the spin wave function is @xmath60 , where @xmath61 is a unitary evolution matrix . in order to satisfy the semiclassical quantization condition ,
it is necessary for @xmath58 to be an eigenvector of @xmath62 , i.e. @xmath63 . here @xmath64 are the complex conjugate eigenvalues of @xmath62 .
hence , depending on the spin state , an additional phase @xmath65 appears in the onsager quantization condition due to spin dynamics .
sdh oscillations in the resistivity are given by the usual lifshitz - kosevich formula @xcite .
accounting for the additional phase @xmath66 we obtain , @xmath67 \mathcal{a}_1(b ) \cos \frac { 2\pi e_f}{\omega } + \dots \label{osc}\end{aligned}\ ] ] the amplitude factor depends on the hole scattering time @xmath68 , @xmath69 .
spin dynamics enters only via the spin evolution phase factor @xmath70 . for the semiclassical approximation approach
we assume large filling factors @xmath71 , hence only the lowest harmonic of magnetic oscillations is taken into consideration .
the matrix phase @xmath62 may be explicitly expressed as a path - ordered exponential which can be calculated using the gauge transformation from eq.([b1 ] ) @xmath72 d\theta \right\ } \nonumber\\ & & = \exp\left\{i\oint { \omega_{\boldsymbol{k } } \cdot \boldsymbol{dk } } - i\frac { 2\pi}{\omega_c } \left [ \boldsymbol{\beta}^{\prime } \cdot \boldsymbol{\sigma } + \frac { \delta}{2 } \sigma_z \right ] \right\ } \
.\end{aligned}\ ] ] hence , using eqs .
( [ flux ] ) and ( [ b1 ] ) we find the prefactor in eq.([osc ] ) for sdh oscillations , @xmath73 , @xmath74 here the @xmath75 term under the square root comes from the non - abelian gauge field .
it is worth noting that the effect of the gauge field is somewhat analogous to thomas precession in special relativity @xcite .
as previously mentioned , the gauge field can not be observed without the in - plane magnetic field .
this is evident from eq.([prefactor ] ) : if @xmath76 the gauge contribution is exactly @xmath77 and hence the phase shift is determined only by the zeeman splitting , @xmath78 .
the zeeman splitting with @xmath79 is @xmath80 .
a nave expectation for the spin accumulated phase would be @xmath81 , but eq.([prefactor ] ) is different from this .
a semi - nave expectation would take into account the abelian berry phase @xmath82 on top of the zeeman splitting .
the phase @xmath82 is given by the first term of the square root expansion in ( [ prefactor ] ) in powers of @xmath75 , yielding @xmath83 the subscript `` ab '' in @xmath66 stands for `` abelian berry '' .
the abelian berry phase approach provides a good description for magneto - oscillations in dirac fermion systems @xcite , and for quantum interference in mesoscopic rings with strong spin - orbit coupling @xcite .
however in our case , both the `` nave '' @xmath84 and the abelian berry phase @xmath85 approach are inconsistent with the data . in our experiments the 2d hole system is formed in a 20 nm wide symmetric gaas quantum well , grown in a ( 311)a gaas - al@xmath86ga@xmath87as heterostructure as indicated in fig.[fig2]a .
previous experiments on this system have shown that holes in a ( 311 ) oriented quantum well have a tensor g - factor with an unusual off - diagonal term @xmath88 @xcite .
although tilted field measurements revealed the presence of the @xmath88 term , no comparison of the shubnikov de haas oscillations with theory was possible , as there was no theory available for 2d hole systems in tilted magnetic fields .
we are now able to show that there is excellent qualitative agreement between the experimental data and the new theoretical model based on the non - abelian gauge field .
we use the coordinates @xmath89 , y\parallel[0 \bar{1}1 ] , z\parallel[311]$ ] shown in fig.[fig2]a .
the gyromagnetic tensor is not diagonal in the @xmath90,@xmath91 and @xmath13 axes , therefore the expression for @xmath92 presented in eq.([heff ] ) and used elsewhere is now replaced by @xmath93 note that the off - diagonal tensor component @xmath88 makes the magnetic response different for three orientations of @xmath2 : @xmath94 , @xmath95 , and @xmath96 @xcite .
this triples the amount of data we can get from the same sample .
details of our experimental setup / method are presented in appendixes [ apa ] , [ apc ] and [ apb ] .
so far we have only considered the effect of the external magnetic field , however spin dynamics can also be influenced by additional couplings , such as the rashba interaction ( stemming from the asymmetry of the interface ) and the dresselhaus interaction ( arising from the lack of inversion symmetry in the bulk gaas crystal ) .
we apply a voltage bias to the back - gate , to tune the symmetry of the gaas quantum well such that the rashba interaction is practically zero , see appendix [ apb ] .
the dresselhaus interaction is relatively weak , nevertheless it is important in some regimes .
moreover , as we discuss below , it brings an additional confirmation of the non - abelian dynamics .
the results of our measurements are presented in panels b and c of fig.[fig2 ] which plot resistivity versus @xmath97 where the in - plane field is altered by tilting the sample at an angle @xmath98 with respect to the applied field , such that @xmath99 , with @xmath44 .
panel b corresponds to tilting in the @xmath100-plane ( @xmath101 ) and panel c corresponds to tilting in the @xmath102-plane ( @xmath103 ) .
the data in panel b is symmetric with respect to @xmath104 , whilst the data in panel c exhibits asymmetry with respect to @xmath105 due to nonvanishing @xmath88 in eq.([gg ] ) . according to eqs .
( [ osc ] ) and ( [ prefactor ] ) , the normalized amplitude of resistivity oscillations , @xmath106 , is a function only of @xmath98 and is independent of the magnitude of the total magnetic field @xmath107 . at tilt angles corresponding to changes in the sign of @xmath108 , the first harmonic of the sdh oscillations invert (
i.e. maxima become minima , and vice versa ) . at these `` coincidence ''
angles the phase @xmath66 must coincide with a half - integer multiple of @xmath8 . in the data ,
these coincidences are observed at the tilt angle @xmath109 for the field applied along the @xmath100-plane in fig.[fig2]b . for the field applied in the @xmath102-plane ( fig.[fig2]c )
there are multiple coincidence angles at @xmath110 for @xmath111 and only a single coincidence at @xmath112 for @xmath113 .
the coincidence angles are plotted in fig . 3 , and are described by eqs . ( [ prefactor ] ) and ( [ gg ] ) .
there are three independent device - specific parameters in these equations , which are @xmath114 , @xmath115 , and @xmath116 .
we use the value @xmath117 derived in appendix [ apd ] as our reference point and hence we are left with unknowns @xmath34 , @xmath118 , and @xmath88 which we treat as free fitting parameters .
c c altogether we have three fitting parameters to describe five coincidence angles . to compare the experimental coincidence angles to those of theory
, we perform a least squares fit to @xmath66 using the observed four coincidences angles for the orientations @xmath119 and @xmath120 ( red and green symbols in fig .
3a ) , and use the values obtained to predict the coincidence angles for the orientation @xmath121 ( blue traces in fig .
3b ) .
the solid red and green curves in fig.[fig3]a show the calculated @xmath122 obtained from this fitting , with the following values of the fitting parameters : @xmath123 , @xmath124 , and @xmath125 . the solid blue line in fig .
3b shows @xmath122 for the @xmath121 orientation calculated using the fitting parameters from fig .
3a , which predicts that there will be only a single coincidence observed in the experimentally measured range of @xmath98 , in agreement with experiment ( blue square ) .
there is reasonable agreement between the predicted coincidence angle of @xmath126 and that observed in the experiment of @xmath127 , although we will shortly discuss the origins of this @xmath128 discrepancy .
to highlight the non - abelian dynamics we have attempted to fit the observed data using the abelian berry formula eq .
( [ preb ] ) instead of eq.([prefactor ] ) . using eqs .
( [ preb ] ) , ( [ gg ] ) we repeat the same procedure described above and fit @xmath85 to the observed four coincidence angles for the orientations ( @xmath119 ) and ( @xmath120 ) , as shown in fig .
3c . the fitting parameters obtained are @xmath129 , @xmath130 .
these parameters were then used to predict the coincidence angles occurring for the orientation @xmath121 , shown in fig .
3d .
the key point is that the abelian theory always predicts three coincidence angles in contrast to the single coincidence observed in experiment .
although the number of coincidence angles is not a topological invariant ( for example it depends on the range of tilt angles available in the experiment ) , it is robust within both the non - abelian and the abelian theories .
as shown in appendix [ apg ] , although the precise tilt angles at which the coincidences occurred are sensitive to the fitting parameters , the _ number _ of coincidences could not be changed even after significant variation of the parameters .
additional confirmation of the non - abelian dynamics comes from the dresselhaus interaction , neglected so far because of its smallness . in the co - rotating frame the spin - orbit coupling given by eq.([b1 ] ) results in an energy splitting @xmath131 between the `` up '' and `` down '' spin states , see eq.([prefactor ] ) .
the quantization axis for @xmath132 is tilted with respect to @xmath13 . the dresselhaus interaction in the co - rotating frame takes the form of a small periodic perturbation @xmath133 .
since the quantization axis is tilted , this perturbation drives transitions between the spin `` up '' and spin `` down '' . because of the smallness of the perturbation the transitions are significant only close to resonance , @xmath134 .
we use the amplitude of the dresselhaus interaction as an additional fitting parameter , and find that it is close to the value known from the literature , see appendix [ ape ] .
the effect of the dresselhaus perturbation is shown by the dashed curves in figs .
3a , b .
the tiny difference from the red and green solid curves in fig .
3a , which do not include the dresselhaus interaction , show that the effect of the interaction is very weak . on the other hand , for @xmath135 ( fig .
3b ) the resonance condition @xmath134 is satisfied and the dresselhaus term now becomes significant .
this causes a clear difference between the solid and dashed blue curves in fig .
3b , which completely removes the small disagreement between experimental and theoretical values of the coincidence angle .
of course , the inclusion of the dresselhaus interaction does not influence the number of coincidence angles , which is a very robust number .
moreover the inclusion of the dresselhaus term explains why the single coincidence for @xmath135 is not sharp , but occurs over a much wider range of angles than for @xmath136 or @xmath137 ( seen as the slow phase inversion and small amplitude of the sdh oscillations in the range @xmath138 for blue traces in fig .
2b ) .
this non - sharp transition for @xmath135 is explained by the inflection in the blue dashed curve in fig.3b , which is due to the dresselhaus interaction .
the `` inflection '' effect provides further confirmation of the non - abelian dynamics , since the small dresselhaus perturbation is always insignificant in the abelian theory .
finally we present in figs .
2b and c theoretical sdh curves calculated with modified lifshitz - kosevich formula ( [ osc ] ) .
the agreement between theory ( including dresselhaus interaction ) and experiment is very good .
overall , our data on the number of coincidences , supported by the slow phase flip of the sdh oscillations for @xmath135 , provide unambiguous evidence for the non - abelian gauge field .
the non - abelian gauge field features centrally in theoretical proposals to exploit hole systems for spintronics and quantum information purposes , including the realization of the dissipationless spin hall effect @xcite and non - abelian manipulation of hole qubits @xcite .
the capacity of hole systems in this context is further enhanced by the suppression of decoherence due to absence of the hyperfine interaction @xcite .
the observation of the non - abelian gauge field in a 2d hole system has positive implications for future studies of hole systems which rely on this concept .
* acknowledgements
* we acknowledge baruch horowitz , ulrich zuelicke , roland winkler , and dimitry miserev for important discussions . |
the strong and attractive nature of the @xmath5 interaction in the @xmath6wave @xcite manifests itself in the sharp enhancement of cross sections in @xmath3-producing reactions near threshold @xcite .
this enhancement was attributed to the possible existence of exotic @xmath7nucleus quasi - bound states @xcite first proposed by q. haider and l. c. liu .
the experimental findings motivated theoretical searches of metastable states of @xmath3-mesic nuclei ( see refs .
7 - 15 for some theoretical predictions and an extensive list of references of other theoretical searches ) , as well as experimental searches for the evidence of such states @xcite in the eta meson producing reactions . to investigate this possibility in the three nucleon system , in the recent past total as well as differential cross sections for the @xmath0 reaction
have been measured @xcite at the celsius storage ring of the svedberg laboratory , uppsala , using the wasa / promice experimental apparatus for excess energies ( @xmath8 ) @xmath9 14 mev . due to the strong energy dependence of the total cross section close to threshold , the data at low energies are however better suited to study the effect of the final state interaction ( fsi ) and explore the existence of quasi - bound eta - nuclear states . for @xmath10 14 mev ,
two data sets exist for the @xmath0 reaction .
one measurement was performed with the spesiii spectrometer at saturne for two excess energies , namely , @xmath11 1.1 mev and 3.3 mev @xcite .
the other data set was obtained using the cosy-11 detection system at the cosy - jlich accelerator for @xmath11 3.2 mev , 6.1 mev and 9.2 mev @xcite .
these data show enhancements of more than an order of magnitude in comparison to the theoretical plane wave total cross sections , suggesting thereby a strong fsi effect .
at low energies , ignoring the @xmath4 fsi , an averaged squared @xmath13 production amplitude , @xmath14 ( which is the ratio of the cross section for the production of the @xmath13 system to arbitrarily normalized phase space ) is traditionally extracted from data in terms of the scattering length @xmath15 by parametrizing @xcite it as , @xmath16 is considered a ( basically unknown ) scale factor and in principle depends upon the particular reaction studied . at the low energies considered here ,
this factor is practically energy independent .
@xmath17 is the relative momentum of the @xmath13 system . within this parametrization , the analysis of the wasa data ( for @xmath18 14 mev ) , using different values of @xmath15 taken from ref .
20 suggests a preference for @xmath19 fm ( as mentioned in ref .
this value of the @xmath13 scattering length is obtained in ref .
20 using an @xmath5 scattering length of , @xmath20 fm .
conclusions drawn about @xmath15 and subsequent inferences made about the possible existence of quasi - bound @xmath13 states , from such studies , may suffer from limitations and uncertainties because of the reasons listed in the following two subsections .
the scattering length enhancement factor description involves the factorization of the @xmath0 reaction amplitude into the @xmath3 production amplitude and an fsi factor .
this may not be correct .
1 exhibits this .
we plot here @xmath14 using , ( i ) the factorized enhancement factor prescription ( eq .
( [ eq1 ] ) ) and ( ii ) using the definition , @xmath21 where , @xmath22 , is evaluated in a theoretical model involving the fsi of the @xmath1 , @xmath2 and @xmath3 @xcite . here
the @xmath13 fsi is included by numerically solving the lippmann - schwinger ( ls ) equation ( as described in the next section ) .
the @xmath23-matrix for @xmath3-d elastic scattering , @xmath24 , entering the ls equation is constructed from @xmath15 given in ref .
20 . for the definition of @xmath22 and the phase space details , see . ) of this article for the @xmath23-matrix which has been used to calculate the cross sections . ]
we plot the squared amplitude obtained from eqs .
( 1 ) and ( 2 ) in fig .
1(a ) and 1(b ) respectively , as a function of the excess energy , @xmath25 , in the @xmath13 centre of mass system .
in order to highlight the slopes of these results for different @xmath26 , we choose values for the normalizing constant , @xmath27 , in case of the amplitude obtained from eq .
( 1 ) , such , that the @xmath14 has the same value for the maximum value of @xmath25 ( @xmath28 3.1 mev ) shown in fig .
to do the same for the results obtained from eq .
( 2 ) , we multiply the results by a normalizing constant , @xmath29 .
we observe that , as we approach the threshold , the curves corresponding to the full calculations for different values of @xmath15 ( shown in fig .
1(b ) ) open up much faster than the approximate calculations ( fig .
this means that the value of @xmath15 extracted by fitting the data by using eq .
( 1 ) could be larger than that obtained by using the full calculation .
the @xmath0 reaction , in the final state , in addition to the @xmath13 interaction , involves the @xmath4 interaction , which at low energies becomes important .
there are two possible @xmath6wave @xmath4 final states , the spin - doublet ( @xmath30 ) and the spin - quartet ( @xmath31 ) states .
@xmath30 bears connection to the @xmath4 distribution in the @xmath32 bound state .
furthermore , since both the proton and deuteron are positively charged , at low energies there exists a sizable coulomb repulsion between them too .
the @xmath4 fsi in the present work is included through the multiplication of an inverse jost function @xcite .
this function is known to give the modification of the non - interacting wave function of two particles at short distances due to the interaction between them ( see next section for the jost function expressions or ref .
21 for more details ) . in fig .
2 we show the inverse of the squared jost function , @xmath33^{-2}$ ] for the @xmath6wave @xmath4 interaction . in order to show the effect of the coulomb repulsion , we have plotted the jost function for the doublet and the quadruplet states with and without inclusion of the coulomb part separately .
the contribution of the jost function without coulomb part can be obtained by keeping @xmath34 = 1 in eq .
( [ quad ] ) .
we also show the doublet jost function multiplied by a factor @xmath35 to reflect the fact that the @xmath30 @xmath4 state has a bound state in ( @xmath36he ) with a separation energy @xmath37 of 5.48 mev .
we see that at small values of @xmath38 relative momentum the strong interaction in the @xmath4 system can enhance the wave function and hence the cross section , by a large amount .
the coulomb interaction , being repulsive reduces this effect . however , the overall effect is still an enhancement ( up to a factor of 3 - 6 ) .
the above discussion thus makes it clear that to learn reliably about the @xmath13 interaction near threshold the data on the @xmath0 reaction in this energy range should be analyzed , as far as possible , with the full theoretical framework available and should include all the interactions . in what follows ,
we have made an attempt to explain the new ( preliminary ) data available from ref .
19 by using a formalism @xcite which includes the interaction between all the pairs in the final state . in the following section we describe the formalism only briefly and
move on to discussing results and conclusions . for more details of the formalism
we refer the reader to our previous articles @xcite .
the cross sections for the @xmath39 reaction close to threshold are evaluated using the two step model for @xmath3 production , where the proton interacts with a proton ( or neutron ) in the deuteron in the first step .
this produces a deuteron and a @xmath40 ( or @xmath41 ) .
the pion in the second step interacts with another nucleon in the original deuteron and produces an @xmath3 .
the two steps thus are @xmath42 and @xmath43 .
the @xmath23-matrix for this model in the plane wave approximation is written as @xmath44 where @xmath45 is the four momentum of the intermediate pion .
the summation is over internal spin projections and the matrix element @xmath46 represents the deuteron wave function in momentum space for which we use the paris parametrization @xcite . the above kind of model for the production amplitude which involves a two step process ( and hence a three body mechanism for production ) has often been used for meson producing reactions at intermediate and low energies .
the importance of such a model for the @xmath47he @xmath3 reaction was first shown by laget @xcite and then tested with more recent data in ref .
22 and 23 .
it has also been used for other meson producing reactions in ref .
26 and 27 .
though one worries about unitarity bounds on cross sections involving two step processes at high energies ( @xmath48 scattering for example refs .
28 - 30 ) , such worries are not relevant for the present work .
the model calculations done here are very close to threshold and in addition , the cross section for @xmath3 production saturates at high energies .
the pion production in the first step being at high energy , is described by an on - shell @xmath49 @xmath50matrix , which is parametrized using the data on this reaction @xcite .
the @xmath51 in eq .
( [ pwtm ] ) has been related to the amplitudes @xmath52 given in ref .
31 as @xmath53 and we relate it to the cross section using @xmath54 where @xmath55 ( @xmath56 ) is the initial ( final ) momentum in the @xmath57 centre of mass system and @xmath58 is the total energy of the same .
the second step @xmath43 being near threshold , is described by an off - shell @xmath50matrix which is taken from a coupled channel calculation @xcite which reproduces the @xmath43 data well .
we refer the reader to ref .
32 for some recent interesting investigations of the @xmath43 data . to take the fsi into account we include the interactions between the @xmath4 and @xmath13 pairs .
the @xmath59 interaction is implicit in our description of the production vertex , which uses the @xmath50matrix for @xmath43 .
though , in principle , a finer treatment of the three body problem ( involving the 3-body @xmath1-@xmath2-@xmath3 scattering ) would be more appropriate , we expect the present approach to work reasonably well for the problem under consideration .
some remarks in connection with this issue are in order here .
the choice of any approximation is guided by the characteristics of the problem at hand . in our case , because of the large momentum transfer ( @xmath28 770 mev / c ) , the reaction takes place over a small volume of space . in such situations ( first introduced by watson and migdal and subsequently discussed in most of the scattering theory text - books @xcite ) , an essential feature of the cross sections is that the energy dependence of the cross section due to fsi , factors out from that of the primary production amplitude .
the fsi factor , which has the correct high energy behaviour is usually taken to be the inverse of the jost function . in the present paper we have used this prescription for the p - d interaction . having done that , we could include the eta - d interaction fully , hence we did not resort to a similar factorization for it .
we included the eta - d interaction effect fully by keeping it inside the production amplitude and using an exact scattering wave function .
the above factorization method has been used in literature to describe the low momentum behaviour of the meson @xcite and pion spectra @xcite measured in proton - proton collisions , and study the total cross sections in the @xmath60 reaction @xcite . the transition matrix for the reaction , @xmath61 , which includes the interaction between the @xmath3 meson and the deuteron
is written as , @xmath62 where @xmath63 and @xmath64 are the proton momenta in the initial and final states , respectively .
@xmath65 is the @xmath3 momentum in the @xmath13 centre - of - mass . @xmath66 and @xmath67 are the spin projections for the proton and the deuteron in the initial and final states .
the final state @xmath3-d wave function , @xmath68 consists of a plane wave and a scattered wave and satisfies the lippmann schwinger equation for @xmath13 elastic scattering : @xmath69 here , @xmath24 is the @xmath23-matrix for @xmath3-d elastic scattering . replacing the above equation in the transition matrix for the @xmath0 reaction ( eq .
( [ fulltmatrix ] ) ) , leads to @xmath70 where the first term represents the @xmath3 production without any interaction with the deuteron and the second one takes care of the rescattering of the @xmath3-d pair to all orders .
it is indeed the half - off - shell @xmath23-matrix , @xmath71 , which is reponsible for converting the off - shell eta mesons in the intermediate state to the on - shell ones in the final state . since the low energy @xmath72 interaction is dominated by the @xmath73 resonance n@xmath74(1535 ) ,
we perform a partial wave expansion of the @xmath3-d @xmath23-matrix , @xmath75 ( @xmath76 ) and retain only @xmath77-waves .
@xmath78 is half off - shell and we write it as a product of an on - shell @xmath50matrix and an off - shell ( which is actually half off - shell here ) form factor .
thus , we write @xmath78 using @xmath77-waves only as , @xmath79 where the @xmath80 is written in terms of the @xmath81 scattering length as @xmath82 we choose to define the off - shell form factor in terms of the deuteron form factor as in ref .
40 , @xmath83 where an average over the directions of the momenta has been performed in order to extract the @xmath77-wave contribution .
the deuteron wave function , @xmath84 , is written using the paris parametrization @xcite . the @xmath81 scattering length , used in eq .
( [ ton ] ) , is taken from ref .
20 , where it has been deduced by solving three - body faddeev type equations .
this calculation includes the scattering of the @xmath3 on the break - up continuum states as well as on the deuteron in the intermediate state .
we use the @xmath15 in this work calculated for a wide range of @xmath5 scattering lengths since the @xmath5 interaction is still not very well known . for the @xmath4 interaction ,
as mentioned above , we use the jost function , which is justified because the @xmath3 production involves high momentum transfer . near threshold
this value is around 700 mev / c .
this means that the reaction occurs in a small interaction volume .
the jost function is given as in ref .
21 , @xmath85^{-2}\,=\,[j_o\,(k_{pd})]^{-2}_q\,+\ , [ ( 1\,+\,\frac{|e_b|}{e})j_o\,(k_{pd})]^{-2}_d .
\label{pdj}\ ] ] the expressions for spin quadruplet ( q ) and doublet ( d ) @xmath86^{-2}$ ] are given by @xmath85^{-2}_x\,=\ , \frac{(k_{pd}^2\,+\,\alpha^2)^2\,(b^c_x)^2}{4}\ , \times \frac{1}{3\,c_o^2\,k_{pd}^2}\,\left[\frac{a_x } { 1\,+\,{\rm cot}^2\,\delta_x}\,\right ] \label{quad}\ ] ] where @xmath87 is either @xmath8 or @xmath88 , @xmath89 and @xmath90 .
the factor @xmath34 above originates from the coulomb interaction .
details regarding the parameters used in eq .
( [ quad ] ) can be found in ref .
the final amplitude is calculated by multiplying eq .
( [ pdj ] ) to the squared @xmath23-matrix obtained after including the @xmath13 interaction , i.e. , eq .
( [ tfull ] ) .
we present the total cross sections calculated , using our full amplitude described above , along with the data for excess energies up to 20 mev in fig .
3 . the cosy-11 data shown in fig .
3 , though preliminary , seems to be in reasonable agreement with the previous data near threshold . in this figure
we show separately the relative contribution of the different pieces in the fsi .
the @xmath13 fsi has been calculated using an @xmath13 scattering length corresponding to @xmath93 fm .
we find that ( i ) both , the @xmath94 and @xmath4 fsi affect the results equally strongly and in the whole region of the excess energies .
( ii ) comparison of results with and without coulomb repulsion show that the coulomb interaction reduces the cross section by a large amount .
in fact , it is because of this interaction that the calculated cross sections come in accord with the experiments .
this is one of the main result of our work .
we have also studied the sensitivity of the results to the @xmath5 scattering length which is still not so well known .
the results shown in fig .
4(a ) have been obtained using the paris parametrization for the deuteron form factor and four different values of the @xmath13 scattering lengths , @xmath15 .
these @xmath15 are calculated for different input values of @xmath95 , ranging from a weak to a strong @xmath5 interaction . for the results presented here , these @xmath95 values are @xmath96 , @xmath97 , @xmath98 , and @xmath99 fm .
the different @xmath15 results in the figure are identified by their corresponding @xmath95 values .
we see that ( i ) all the calculated cross sections show a strong fsi effect , and ( ii ) the results for @xmath100 and @xmath98 fm seem to pass through the maximum number of data points . for the sake of completeness , in fig .
4(b ) , we show the results obtained by using in eq .
( [ deutform ] ) , prescriptions of the deuteron wave function other than the paris form , which are also sometimes used in literature .
we have thus calculated the total cross sections using the wave functions obtained from the paris potential , hulthen potential , and the gaussian form .
the scattering length @xmath95 for the calculation in fig .
4(b ) has been kept fixed at @xmath97 fm . the maximum difference in the total cross sections , obtained due to the use of different wave functions ,
is found to be a factor of the order of 1.6 at higher excess energies .
we have presented a theoretical analysis of the data on the @xmath0 reaction near threshold using a model which was reasonably successful in reproducing the data at higher energies .
we start by comparing the model calculations with a form of the scattering amplitude often used for data analysis and investigate the limitations of this form .
the cross sections for the @xmath0 reaction at low energies ( up to @xmath28 20 mev above threshold ) are then evaluated within a two step model for @xmath3 production to investigate the importance of including the fsi between all particles at low energies .
the conclusions of this study can be summarized as follows : 1 .
quantitative reproduction of the experimentally observed enhancement in the cross section near threshold , requires the inclusion of the interaction between all three particles , namely , the proton , deuteron and the eta meson in the final state .
both the strong and coulomb @xmath4 fsi are found to be important at low energies .
2 . the best agreement with experimental data seems to come from the @xmath3-nucleon scattering length values of @xmath101 fm and @xmath98 fm .
calculations based on the exact alt - grassberger - sandhas equations in ref .
20 lead to @xmath102 scattering lengths of ( @xmath103 ) fm and ( @xmath104 ) fm for the two above @xmath95 s respectively .
3 . in the energy range discussed here , the scattering length description of the @xmath102 @xmath23-matrix , @xmath24 , which enters the lippmann - schwinger equation for the @xmath13 elastic scattering wave function , seems to be sufficient .
4 . the extracted value of the @xmath3-deuteron scattering length , @xmath15 , from experimental @xmath14 using the simplified relation in eq .
( 1 ) can be misleading .
one of the authors , k. p. k. , would like to thank the fundao para a cincia e a tecnologia of the ministrio da cincia , tecnologia e ensino superior of portugal for financial support under the contract sfrh / bpd/40309/2007 . | a model for the @xmath0 reaction published earlier , including the final state interaction ( fsi ) of all particles , is revisited to investigate the low energy data on this reaction .
the three body problem of @xmath1-@xmath2-@xmath3 scattering in the final state is approximated in terms of pairwise interactions between the three particles in the final state .
apart from a comparison with some preliminary data , two new findings relevant to the near threshold data analysis are reported .
the first one points toward the limitations of an fsi factor used conventionally to extract the eta - deuteron scattering length and infer subsequently on the existence of eta - mesic states .
the second result emphasizes the role of the @xmath4 fsi and the strong coulomb repulsion near threshold .
finally , a comparison of the above model calculation with low energy data , excludes very large eta - nucleon scattering lengths . |
the _ wide - field infrared survey explorer _ ( ) @xcite is an extremely capable and efficient black hole finder . as demonstrated in selected fields by ( , * ? ? ? * ; * ? ? ? * ) , the same material that obscures agn at uv , optical and soft x - ray energies is heated by the agn and emits strongly at mid - ir wavelengths .
the all - sky survey identifies millions of obscured and unobscured quasars across the full sky ( , * ? ? ?
* ; * ? ? ?
* ) , as well as very rare populations of extremely luminous , heavily obscured agn . in terms of the latter , the extragalactic team has been pursuing sources that are faint or undetected in @xmath3 ( 3.4 @xmath21 m ) and @xmath5 ( 4.6 @xmath21 m ) , but are bright in @xmath7 ( 12 @xmath21 m ) and @xmath9 ( 22 @xmath21 m ) .
we refer to this population as @xmath11-dropouts @xcite .
this is a very rare population ; selecting to a depth of 1 mjy at 12 @xmath21 m , there are only @xmath10 such sources across the extragalactic sky ( @xmath22 per @xmath23 ) .
these objects are undetected by _
rosat _ and tend to be optically faint ( @xmath24 ) , below the detection threshold of sdss .
we have obtained spectroscopic redshifts for @xmath25 @xmath11-dropouts thus far , consistently finding redshifts @xmath26 , with our current highest redshift source at @xmath27 @xcite
. approximately half of the sources show clear type-2 agn signatures in the optical spectra , with the other half typically showing only ly@xmath28 emission , sometimes extended , which could be due to star formation and/or agn activity @xcite .
the lack of a far - ir peak in their broad - band seds suggests the dominant energy input for this population comes from a heavily obscured agn and not extreme starbursts ( , * ? ? ?
* ; * ? ? ?
related high - luminosity sources selected from the satellite have also recently been reported by @xcite and @xcite , while several teams have identified less rare , less luminous sources from surveys with less extreme colors ( , * ? ? ?
* ; * ? ? ?
* ) . here
we report on the first targeted x - ray follow - up of @xmath11-dropouts .
we observed two sources with both the _ nuclear spectroscopic telescope array _
* ) and @xcite ; a third source was only observed by .
unless otherwise specified , we use vega magnitudes throughout and adopt the concordance cosmology , @xmath29 , @xmath30 and @xmath31 .
[ fig : spectra ] presents the optical spectra of the three @xmath11-dropouts targeted for x - ray follow - up : wisea j181417.29 + 341224.8 ( hereafter , ) , wisea j220743.82 + 193940.1 ( hereafter , ) and wisea j235710.82 + 032802.8 ( hereafter , ) .
the spectra were all obtained with the low resolution imaging spectrometer ( lris ; * ? ? ?
* ) on the keck i telescope , between 2010
july and 2010 november .
the sources were selected on the basis of having unusually red colors across the passbands in the initial all - sky data release : @xmath32 , @xmath33 , and @xmath34 ( for further details on the @xmath11-dropout selection , see * ? ? ?
the most likely interpretation of sources with these extreme colors is that they host an extremely luminous , heavily obscured agn which only becomes evident at observed wavelengths @xmath35 m .
the @xmath3 flux limit essentially constrains the sample to @xmath36 for the host galaxy not to be detected .
note the diversity of the optical spectra of the three sources targeted for x - ray follow - up ( fig .
[ fig : spectra ] ) .
this is representative of the diverse optical spectroscopic properties of the @xmath11-dropout population in general @xcite .
@xcite discusses in depth : briefly , the optical spectrum is indistinguishable from an @xmath37 lyman - break galaxy ( lbg ) at @xmath38 ( , * ? ? ?
* ) , with no obvious signature of a ( buried ) agn .
however , typical lbgs have @xmath8 m flux densities a factor of 1000 lower @xcite .
the redshift of is based on a single , asymmetric , high equivalent width emission line which is reliably identified as ly@xmath39 ( for a detailed discussion of one - line redshifts , see * ? ? ?
. shows a high equivalent width ( @xmath40 , observed ) , slightly broadened ( fwhm @xmath41 ) , self - absorbed line identified as @xmath42 , which is a common strong line in agn .
however , quite unusually , no ly@xmath28 emission is evident .
the only strong feature blue - ward of the emission line is a continuum break at observed @xmath43 , which is consistent with the ly@xmath28 forest break for the longer wavelength emission line being .
@xcite reports on a detailed investigation of a similar sdss quasar with broad emission , but lacking broad ly@xmath28 emission .
they argue that the unusual spectrum can not be solely due to dust extinction in the broad line region ( blr ) , and instead suggest that most , but not all , of the spectral properties can be explained by an ununusually high density gas in the blr ( @xmath44 ) with an incident power - law continuum extending to @xmath45 m .
clearly the unusual optical spectrum of is worthy of future study , but such analysis is beyond the scope of the current paper which focuses on the x - ray properties of the @xmath11-dropout population . in terms of their radio properties ,
has a counterpart offset by 61 in the nrao / vla sky survey ( nvss ; * ? ? ?
* ) with a flux density @xmath46 mjy .
@xcite report on follow - up radio observations of this source with the jansky very large array which resolves the nvss emission into two distinct sources , the fainter of which is associated with . based on both its rest - frame 1.4 ghz radio luminosity ( @xmath47 ) and its radio - to - optical ratio ( rest - frame @xmath48 ) , qualifies as radio - loud .
has a counterpart offset by 71 in the nvss with @xmath49 mjy , suggesting that it is also radio loud .
has no radio counterpart in either nvss or the vla faint images of the radio sky at twenty centimeters survey ( first ; * ? ? ?
the radio luminosities of the first two sources further indicate the presence of a powerful agn in this -selected population .
lccccccc & & + ( rl)3 - 5 ( rl)6 - 8 & @xmath50 & observation i d & ut date & exposure & observation i d & ut date & mos1/mos2/pn exposure + & & & & ( ks ) & & & ( ks ) + & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) + & 2.452 & 60001114002 & 2012 oct 30 & 21.3 & 0693750101 & 2012 oct 07 & 29.6 / 29.6 / 19.6 + & 2.021 & 60001115002 & 2012 oct 30 & 20.8 & 0693750201 & 2012 nov 22 & 18.4 / 18.7 / 11.5 + & 2.113 & & & & 0693750401 & 2013 jan 02 & 29.1 / 28.7 / 14.4 + ( 1 ) : target name ; full name and coordinates are in table 2 .
( 2 ) : redshift . ( 3 ) and ( 4 ) : observation i d and start date .
( 5 ) : net on - axis exposure time .
this value applies for both fpma and fpmb .
the target was on - axis for both of the observations .
( 6 ) and ( 7 ) : observation i d and start date .
( 8) : net on - axis exposure time , corrected for flaring and bad events , for the mos cameras and the pn camera , as indicated .
[ table : obslog ]
in 2012 october obtained @xmath51 ks observations of and ; details of the observations , including net exposure times , are provided in table [ table : obslog ] .
we processed the level 1 data using the data analysis software ( nustardas ) v.1.2.0 , and produced calibrated and cleaned event files ( level 2 data ) for both focal plane modules ( fpma and fpmb ) using nupipeline and the most current available version of the calibration database files ( caldb 20130509 ) . neither source was detected , though a serendipitous broad - lined agn at @xmath52 was identified in the field @xcite .
we measured gross source counts in 45 radius apertures centered on the positions and local background counts from an annulus of inner radius 90 and outer radius 150centered on the sources .
we performed photometry in the observed - frame 3 - 24 kev , 3 - 8 kev and 8 - 24 kev bands , as well as the rest - frame 10 - 40 kev band for both fpms and used binomial statistics to determine the likelihood of the sources being detected .
binomial statistics are more accurate than poisson statistics at these faint limits since it takes into account uncertainty in the measured background ( , it takes the total background counts into account , not just the scaled background counts ) .
we use binomial statistics to calculate the probability that the measured source counts are purely due to background fluctuations ( , false , or no - source probabilities ; for details , see * ? ? ?
* ) . for both sources
the probability of a false detection based on the binomial statistics is @xmath53 . as we take a no - source probability @xmath54 to indicate a detection , neither source was detected . since binomial statistics are not amenable to plotting simple tracks , fig .
[ fig : xraycounts ] shows the x - ray counts with poisson no - source probabilities .
the latter provide a good approximation of binomial probabilities for our sources given the reasonably high background count rates .
table [ table.phot ] reports the 90% confidence limit upper limits to the flux in the 3 - 24 kev energy band , calculated using the bayesian method of @xcite . to convert count rate to source flux , we used xspec v12.8.0k , taking into account the response matrix file ( rmf ) and ancillary response file ( arf ) for each fpm .
we assumed a power - law model with @xmath55 . assuming a harder intrinsic spectrum does not qualitatively change these results ; , adopting @xmath56 only changes the rest - frame 10 - 40 kev luminosities by @xmath57 .
we obtained @xmath58 ks observations of the three @xmath11-dropouts between 2012 october and 2013 january with the epic - mos @xcite and epic - pn @xcite cameras .
details of the observations are provided in table [ table : obslog ] .
we used data products from the pipeline processing system ( pps ) , analyzed with the _ science analysis software _
( sas v.12.0.1 ) .
we measured 0.5 - 10 kev source counts in 15 radius apertures centered on the positions .
for most observations , we measured local backgrounds in slightly offset source - free circular apertures with radii of @xmath59 selected to avoid serendipitous sources and chip gaps .
the exceptions were the epic - mos observations of which had no nearby serendipitous sources and thus allowed for a 30 to 70 radius annulus centered the source position .
as above with the observations , we calculated binomial false probabilities and plot the equivalent using poisson statistics in fig .
[ fig : xraycounts ] . only is reliably detected , with a binomial no - source probability of 0.04% with epic - pn and 0.1% with epic - mos1 ; was not reliably detected with epic - mos2 .
is weakly detected with epic - mos2 with a no - source probability of 0.2% , and is undetected with the other two cameras .
we used xspec to convert count rate ( or count rate limits ) to flux , assuming a power - law model with @xmath55 , and the rmfs and arfs for each epic camera . again , adopting a harder intrinsic x - ray spectrum would have a modest quantitative effect on the derived luminosities , but would not affect our broad conclusions . specifically , were we to instead assume @xmath60 , the rest - frame 2 - 10 kev luminosities would change by @xmath61 .
table [ table.phot ] presents basic source properties and multi - wavelength photometry for the three x - ray - targeted @xmath11-dropouts .
we list mid - ir data from the allwise data release and , where the latter comes from the _ warm spitzer _
observations reported by @xcite . comparing the agn luminosity and reddening of derived solely from the mid - ir photometry ( 4.1 ) to the values in @xcite derived from 16-band multi - wavelength photometry , we obtain consistent values within @xmath62 .
cccc & & & + r.a .
( j2000 ) & 18:14:17.29 & 22:07:43.82 & 23:57:10.82 + dec .
( j2000 ) & + 34:12:24.8 & + 19:39:40.1 & + 03:28:02.8 + @xmath50 & 2.452 & 2.021 & 2.113 + & & & + @xmath3 ( 3.4 @xmath21
m ) & @xmath63 & @xmath64 & @xmath65 + @xmath5 ( 4.6 @xmath21 m ) & @xmath66 & @xmath67 & @xmath68 + @xmath7 ( 12 @xmath21 m ) & @xmath69 & @xmath70 & @xmath71 + @xmath9 ( 22 @xmath21 m ) & @xmath72 & @xmath73 & @xmath74 + @xmath75 $ ] & @xmath76 & @xmath77 & @xmath78 + @xmath79 $ ] & @xmath80 & @xmath81 & @xmath82 + @xmath83 & @xmath84 & @xmath85 & @xmath86 + @xmath87 & @xmath88 & @xmath89 & @xmath90 + @xmath91 & @xmath92 & @xmath93 & @xmath94 + @xmath95 & @xmath96 & @xmath97 & + & & & + @xmath98 & @xmath99 & @xmath100 & @xmath101 + @xmath102 & @xmath103 & @xmath104 & @xmath105 + @xmath106 & 1.04 & @xmath107 & @xmath108 + @xmath109 & @xmath110 & @xmath111 & + astrometry and photometry are from the allwise release ; mid - ir photometry is all in vega magnitudes and limits report the 95% confidence lower limit to the apparent magnitude .
photometry , in brackets , is from griffith et al .
x - ray fluxes , all in the observed frame , are in units of @xmath112 .
@xmath98 is the rest - frame @xmath113 m luminosity ( @xmath114 ) of the agn in units of @xmath115 .
rest - frame x - ray luminosities are in units of @xmath116 .
we report 68% confidence limit ( cl ) uncertainties on the x - ray fluxes ; upper limits correspond to the 90% cl .
[ table.phot ]
we modelled the spectral energy distribution ( sed ) of each source using the @xcite 0.03 - 30 @xmath21 m empirical agn and galaxy templates .
each sed is modelled as a best - fit , non - negative combination of an elliptical , spiral , and irregular galaxy component , plus an agn component .
only the agn component is fit for dust reddening ( for further details on the sed models , see * ? ? ?
* ; * ? ? ?
the modeling outputs @xmath98 , the derived intrinsic luminosity ( @xmath114 ) of the agn component at rest - frame @xmath113 m , as well as the reddening of the agn component , @xmath117 ( see table [ table.phot ] ) . for
the typical gas to dust ratio observed by @xcite for luminous agn , the nuclear reddening values of @xmath118 imply gas columns of @xmath119 , which reach into the compton - thick regime ( @xmath120 ) .
the mid - ir properties of @xmath11-dropouts indicate the presence of extremely luminous , heavily obscured agn , with bolometric luminosities approaching , or even exceeding @xmath13 @xcite .
given the heavy obscuration implied by their mid - ir seds and optical spectra , it is perhaps unsurprising that the three sources we targeted for x - ray follow - up are either undetected or only faintly detected , despite observing at the penetrating energies above rest - frame 10 kev which are less affected by absorption .
but what are their absorption column densities ?
we can obtain indirect estimates of this from their mid - ir luminosities since unobscured agn tend to have a fairly tight relation between their mid - ir and x - ray luminosities .
[ fig : lxlir ] compares the rest - frame @xmath113 m and x - ray luminosities for our targeted sources .
we show local relations from @xcite , @xcite and @xcite , as well as non - beamed agn with @xmath121 from the serendipitous survey @xcite ; the serendipitous sources all lie within the scatter of the published relations .
we also show compton - thick quasars observed at soft energies from @xcite , three sdss type-2 quasars observed by @xcite , and an obscured quasar at @xmath122 detected by in the extended deep field south ( ecdfs ; * ? ? ?
the literature obscured agn and the -selected @xmath11-dropouts generally lie significantly below the published relations . assuming the suppression of their x - ray emission is due to absorption rather than intrinsic x - ray weakness , we can estimate their column densities from the dashed lines in fig . [
fig : lxlir ] , which apply columns of @xmath123 to the published relations .
these were calculated using the mytorus model @xcite with photon index @xmath124 and a torus inclination angle @xmath125 .
the implication is that all three targeted sources are obscured , possibly heavily obscured or compton - thick . for the rest - frame 2 - 10 kev panel , the one detected source ,
, has a column of @xmath126 for the @xcite relation , and yet higher columns for the other two relations .
the undetected sources require minimum columns of @xmath127 for all but the @xcite relation .
constraints from the nondetections are less strict , but are again consistent with heavy absorption columns .
furthermore , two of the three -selected targets are radio - loud .
since optically selected radio - loud quasars tend to have extra x - ray emission quasars as compared to radio - quiet quasars @xcite , this suggests yet larger minimum absorption columns for and .
in addition , the jet - linked x - ray emission is apparently subject to similarly high obscuration as the nuclear x - ray emission .
we designed our x - ray integration times to provide robust detections for ( i ) typical intrinsic agn seds , and ( ii ) gas column densities @xmath128 .
none of the three sources was strongly detected , implying that at least one of the assumptions in our experimental design does not hold .
has now observed a range of obscured agn , from famous , local sources such as mrk 231 @xcite , circinus @xcite , ngc 424 @xcite , and ngc 4945 @xcite , to higher redshift obscured quasars at @xmath129 from sdss @xcite and @xmath0 in the ecdfs @xcite .
a recurring theme of these observations is that several agn which are extremely luminous at certain wavelengths , such as in the mid - ir or [ ] @xmath130 , remain faint at x - ray energies .
for some objects , this is the case even for the more penetrating hard x - rays @xmath131 kev .
for some sources , such as optically bright ( @xmath132 ) broad - absorption line ( bal ) quasars from the palomar - green ( pg ) survey @xcite and mrk 231 , we consider , and sometimes even favor attributing the hard x - ray faintness to intrinsic x - ray weakness @xcite .
such intrinsic x - ray weakness seems the most plausible scenario when the agn appears unobscured at certain wavelengths , through strong uv continuum emission , broad emission lines , and/or weak x - ray spectra that are well described by a moderately absorbed power - law . for a more detailed description of the intrinsically x - ray weak scenario ,
see @xcite and @xcite .
for the three -selected agn discussed here , we instead favor interpreting the x - ray faintness as being due to a typical agn seen through extremely high absorbing columns , consistent with their mid - ir seds and optical spectra . indeed , the x - ray column constraints from fig .
[ fig : lxlir ] are broadly consistent with the mid - ir measurements given typical luminous agn gas - to - dust ratios from @xcite . what is more surprising is that these sources , amongst the bolometrically most luminous agn known , appear heavily obscured .
various observations have shown that more luminous agn are less likely to be obscured ( , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this is consistent with the `` receding torus model '' , first proposed by @xcite , in which the height of the torus is independent of luminosity while the the inner radius of the torus , corresponding to the distance at which dust reaches its sublimation temperature , increases with luminosity .
therefore , in this model , more luminous agn have more sightlines into the nucleus and thus have a lower likelihood of being obscured .
the @xmath11-dropout population is a rare population , with a surface density of just one source per @xmath133 in the extragalactic sky .
the mid - ir luminosities imply intrinsic x - ray luminosities of a few @xmath134 from the relations of @xcite and @xcite .
@xcite report on x - ray follow - up of the most luminous quasars in the sdss available at the time , finding 34 quasars across 4188 deg@xmath135 , or one source per @xmath136 .
the x - ray luminosities of this luminous quasar sample prove comparable to the expected intrinsic x - ray luminosities from the -selected sample , implying the surprising discovery of comparable numbers of obscured and unobscured quasars at the top of the luminosity function .
@xcite and @xcite present more detailed comparisons between the @xmath11-dropout and luminous unobscured quasar populations .
the discovery of a significant population of heavily obscured , extremely luminous agn does not conform to the simple receding torus model , suggesting an additional source of obscuration .
indeed , the models of @xcite predict that compton - thick agn should be more common at higher redshift because of the high fueling rates of quasars require significant gas reservoirs .
this is consistent with models showing that mergers may be more prominent in fueling agns at @xmath0 ( , * ? ? ?
* ; * ? ? ?
* ) . indeed
, these models also predict that the unified model of agn will break down for high - luminosity agns at @xmath137 because the obscuration is not confined to the nucleus ( see also * ? ? ?
further investigations into this interesting population are clearly warranted .
deeper x - ray observations , achieving robust detections rather than faint and non - detections , would be valuable . for example , assuming the obscuration is from a nuclear torus , deeper x - ray observations should detect narrow , reflected fe k@xmath28 fluorescent emission at rest - frame 6.4 kev ( , * ? ? ?
* ) ; a significant non - detection could point towards obscuration on larger scales than the torus .
such large - scale obscuration could also be probed by high - resolution imaging , such as far - ir observations with alma , and , eventually , mid - ir observations with _ jwst_. near - ir spectroscopy could also look for reddening in the agn narrow - line region ( , * ? ? ? * ) , which is expected to be significantly larger than the torus . a clearer understanding of the geometry of the obscuring region combined with an improved reckoning of the @xmath11-dropout space density will enable us to better place this population within the context of agns and galaxy evolution .
this work was supported under nasa contract no .
nng08fd60c , and made use of data from the _ nustar _ mission , a project led by the california institute of technology , managed by the jet propulsion laboratory , and funded by the national aeronautics and space administration .
we thank the _ nustar _ operations , software and calibration teams for support with the execution and analysis of these observations .
this research has made use of the _ nustar _ data analysis software ( nustardas ) jointly developed by the asi science data center ( asdc , italy ) and the california institute of technology ( usa ) .
this publication makes use of data products 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 .
we acknowledge financial support from the science and technology facilities council ( stfc ) grants st / k501979/1 ( gbl ) , st / i001573/1 ( dma and adm ) and st / j003697/1 ( pg ) . rja was supported by gemini - conicyt grant number 32120009 .
rch acknowledges support from nasa through adap award nnx12ae38 g and the national science foundation through grant number 1211096 . | we report on a and program that has observed a sample of three extremely luminous , heavily obscured -selected agn at @xmath0 in a broad x - ray band ( @xmath1 kev ) . the parent sample , selected to be faint or undetected in the @xmath2 m ( @xmath3 ) and @xmath4 m ( @xmath5 ) bands but bright at @xmath6 m ( @xmath7 ) and @xmath8 m ( @xmath9 ) , are extremely rare , with only @xmath10 so - called `` @xmath11-dropouts '' across the extragalactic sky .
optical spectroscopy reveals typical redshifts of @xmath0 for this population , implying rest - frame mid - ir luminosities of @xmath12 and bolometric luminosities that can exceed @xmath13 . the corresponding intrinsic , unobscured hard x - ray luminosities are @xmath14 for typical quasar templates .
these are amongst the most luminous agn known , though the optical spectra rarely show evidence of a broad - line region and the selection criteria imply heavy obscuration even at rest - frame @xmath15 .
we designed our x - ray observations to obtain robust detections for gas column densities @xmath16 .
in fact , the sources prove to be fainter than these predictions .
two of the sources were observed by both and , with neither being detected by ( @xmath17 ) , and one being faintly detected by ( @xmath18 ) .
a third source was observed only with , yielding a faint detection ( @xmath19 ) .
the x - ray data require gas column densities @xmath20 , implying the sources are extremely obscured , consistent with compton - thick , luminous quasars .
the discovery of a significant population of heavily obscured , extremely luminous agn does not conform to the standard paradigm of a receding torus , in which more luminous quasars are less likely to be obscured .
if a larger sample conforms with this finding , then this suggests an additional source of obscuration for these extreme sources . |
we use a two - dimensional explicit electrostatic kinetic pic simulation model @xcite .
the functionality of the code is similar to what we have used earlier to model the positively biased e - sail tether @xcite , but the code is vectorised with agnervector library @xcite and parallelised with the standard message passing interface ( mpi ) and openmp tools .
we run the code with real electron and ion masses .
we simulate a negatively charged tether in plasma flow which is caused by the satellite s orbital motion through the ionosphere .
we use a cartesian coordinate system where @xmath3 is antiparallel with the tether - perpendicular component of the flow , @xmath4 is parallel to the tether and @xmath5 completes a right - handed coordinate system . for convenience
we take the plasma flow to be perpendicular to the tether , which implies no loss in generality .
the simulation is two - dimensional in the tether - perpendicular plane .
a constant external magnetic field is employed in some runs . by default
the plasma is cold oxygen plasma , temperature 0.1 ev , and it flows at 7.5 km / s which corresponds to 4.7 ev bulk flow ion energy .
the tether bias voltage is typically a few hundred volts negative .
the parameters of all reported runs are given in table [ tab : simparams ] .
the simulations are initialised from vacuum and the plasma flow starts entering the box at @xmath6 .
above @xmath7 500 km where the plasma brake is relevant ( at lower altitude the neutral drag typically brings satellites down rapidly enough ) , ionospheric plasma consists mainly of o@xmath8 ( 16 amu ) , n@xmath8 ( 14 amu ) , he@xmath8 ( 4 amu ) and h@xmath8 ( 1 amu ) ions .
the main species are o@xmath8 and h@xmath8 ; the minor species n@xmath8 can be summed to o@xmath8 because the masses are approximately similar .
the abundance of o@xmath8 decreases with altitude . at all altitudes
, solar activity tends to increase the oxygen abundance . for the plasma brake ,
the most relevant environment is o@xmath8 plasma because proton plasma produces order of magnitude smaller thrust and helium is usually not a dominant species .
for this reason , the main emphasis in this paper is on o@xmath8 plasma .
runs reported in this paper used 16 nodes on a cray xc30 .
the xc30 consists of compute nodes , each node has 2 processor chips and each chip has 10 2.6 ghz execution cores . the runs reported here used 16 xc30 nodes and achieved @xmath9 particle propagations per second which corresponds to @xmath10 tflops single precision floating point performance .
we first run the code with oxygen plasma and without magnetic field for different values of the tether potential @xmath11 .
the runs are detailed in table [ tab : runlist ] .
the vacuum potential of the tether ( the voltage difference between the tether and the simulation box boundary if there is no plasma inside the box ) is an input parameter to the simulation . after the simulation , the true potential of the tether
is calculated by evaluating the plasma potential at the tether and adding it to the vacuum potential .
for this reason , the tether potential values @xmath11 are not round numbers .
we label the runs partly by their vacuum potential values .
figure [ fig : vdep ] shows the simulated thrust ( determined from particle momentum balance averaged over the last 1/3 of the run ) with open circles .
coloured marks in fig .
[ fig : vdep ] are results from some runs containing a magnetic field and they are described in later subsections below .
the solid line is the following analytical formula , inspired by our earlier work ( * ? ? ?
* equation 20 ) : @xmath12 where @xmath13 is the dynamic pressure , @xmath14 is the ion mass ( @xmath15 amu for oxygen plasma here ) , @xmath16 is the plasma flow speed relative to spacecraft ( assumed to be perpendicular to the tether or else @xmath16 denotes only the perpendicular component ) , @xmath17 @xmath18 is the tether s effective electric radius ( * ? ? ?
* appendix a ) , @xmath19 is the effective debye length and @xmath20 is the bulk ion flow energy in voltage units .
the effective electric radius is approximately given by @xmath21 where @xmath22 is the tether wire radius , typically 12.5 - 25 @xmath0 m , and @xmath23 is the tether width , typically 2 cm ( a rough value of @xmath23 is sufficient to know because @xmath24 enters into eq .
( [ eq : dfdz ] ) only logarithmically ) . in this paper
we use the value @xmath25 mm . the numerical coefficient ( 3.864 ) in front of the expression has been selected to give a good fit to the present simulation results .
this value is about 2.25 times larger than the value 1.72 used in our previous work @xcite .
we think that this difference may arise because in the earlier work we used typical solar wind parameters so that the ratio @xmath26 was about 10 while this ratio is about 100 in the present ionospheric case so that the works explored different regions of the parameter space . in the ionospheric case ,
many ions are deflected backward by the potential well around the tether which increases the thrust markedly in comparison to the solar wind case where the much larger bulk flow speed causes the ions to deflect more modestly .
figure [ fig : curvebaseline ] shows the time history of thrust during the v400 ( baseline ) run , computed by two complementary methods : the direct coulomb force acting on the tether ( by evaluating the numerical gradient of the plasma potential at the tether position ) and the total momentum @xmath3 component lost by particles between entering and leaving the simulation box .
figure [ fig : baselinepcolor ] shows the two - dimensional instantaneous electron and ion density at the final state of the baseline run . increasing the tether voltage makes the ion sheath larger .
the highest voltage runs ( marked with `` g '' in table [ tab : runlist ] ) were performed with extended box size of 768@xmath27768 to accommodate the larger sheath . as a consistency check ,
run v500 was performed with both grid sizes .
the difference in the determined thrust was minimal ( table [ tab : runlist ] , compare v500 and v500 g ) .
the quasi - monochromatic oscillation seen in fig .
[ fig : curvebaseline ] occurs in runs v300 and higher . in lower voltage
runs the oscillation is absent .
when present , the oscillation neither increases nor decreases with time .
we performed three more runs with the same parameters as v400 ( baseline ) , but with each of the magnetic field components in turn set to 30000 nt , a typical leo field strength .
the results are shown in figs .
[ fig : curvebx]-[fig : curvebz ] .
it is seen that in runs bx and bz , the oscillation which was already present in the baseline run is now increasing in amplitude .
in contrast , in the by run the oscillation is damped and is going to disappear .
thus , @xmath28 ( magnetic field perpendicular to tether and flow ) is stabilising while flow - directed or tether - directed field is destabilising at the baseline voltage . the final state of run bx is shown in fig .
[ fig : bxpcolor ] .
the sheath surrounding the tether is unstable and radiates plasma waves in all directions in the @xmath29 plane . in the tether - fixed coordinate frame ,
the waves move approximately at the same phase speed as the bulk flow ( 7.5 km / s ) .
the waves appear to move with the ion flow which is reflected by the tether s potential well . the incoming flow has a mach number of @xmath30 with respect to ion acoustic wave speed . because of energy conservation , reflected ions move radially outward with the same speed as the incoming flow .
the tether s potential well fluctuates , which modulates the flow of outward reflected ions .
the modulations propagate outward approximately at the same speed as the ions because the ion acoustic speed is less than the bulk velocity of the outward moving ion population . in fig .
[ fig : bxpcolor ] one can also see that the @xmath3-directed magnetic field tends to restrict electron motion in @xmath5 , which creates some horizontal stripes seen in the figure ( reddish stripes on positive @xmath3 ) , visible in both electron and ion density and emanating from the edges of the electron cavity .
figure [ fig : bzpcolor ] shows the final state of run bz .
again the state is unstable and
the tether s potential well radiates plasma waves .
additionally , the boundaries of the plasma wake behind the tether display some turbulent behaviour .
this probably occurs because now the magnetic field is along the tether so that electrons respond to @xmath3 and @xmath5 directed electric fields by @xmath31 drifting ( similar to kelvin - helmholtz instability in plane perpendicular to @xmath32 ) .
not surprisingly , the horizontal stripes which were visible in the bx run ( fig .
[ fig : bxpcolor ] ) are now absent .
figures [ fig : curvehelium ] and [ fig : curveproton ] show thrust histories for helium and proton plasma , respectively , and with other parameters in their baseline ( v400 ) values . in both light ion cases ,
an unstable oscillation is present .
the instability grows faster in the proton run than in the helium run and has time to evolve into intermittent , nonlinear regime .
figure [ fig : protonpcolor ] shows the final state of the proton run .
some ions have become trapped by the potential well which has decreased the thrust to some extent ( fig .
[ fig : curveproton ] ) .
the sheath emits plasma waves in all @xmath29 directions .
the flow mach number with respect to ion acoustic speed is low ( 1.2 ) which is probably the reason why the wake behind the tether is rather short ( the wake gets filled quickly from the boundaries because the ion thermal speed is almost as large as the flow speed ) . in fig .
[ fig : protonpcolor ] the behaviour of the densities near the inflow boundary ( right side ) shows some @xmath5-directed striping and does not look completely natural .
it is possible that the finite box size affects the result to some extent in this case . in the proton case ( fig .
[ fig : protonpcolor ] ) it is also noteworthy that the emitted ion waves have short enough wavelength that the electron density does not follow the ion density , in other words , that the emitted wavelength is not much larger than the electron debye length .
formula [ eq : dfdz ] predicts 12.2 nn / m in the helium case and 2.92 nn / m in the proton case .
the simulated values at 2.43 ms are larger ( 14.6 and 5.05 nn / m , respectively ) , even though in the proton run some ion trapping had already occurred which had reduced the thrust .
probably the smaller dynamic pressure of light ion bulk flow depresses the sheath less than in the oxygen case , and this effect is not included in eq .
( [ eq : dfdz ] ) where the thrust is assumed to be linearly proportional to the ion mass . in the helium run ( not shown ) the behaviour is intermediate between the oxygen and proton runs . in subsection [
subsection : longerruns ] below we will investigate longer time behaviour of proton plasma where also a magnetic field is included . in all runs thus far presented we have assumed that the electron temperature is 0.1 ev ( the same as ion temperature ) . in the 700 - 900 km altitude range , the electron temperature is actually often 0.2 - 0.3 ev .
figure [ fig : curvete ] shows thrust behaviour of a run ( named te ) which is otherwise identical to baseline ( v400 ) , but the electron temperature is set to 0.3 ev .
the thrust remains essentially unchanged and the oscillation is damped .
so run te suggests that a higher electron temperature improves stability .
figure [ fig : curvebxlong ] shows longer time behaviour ( @xmath33 timesteps , 23.4 ms ) of the thrust in the presence of @xmath3-directed magnetic field .
the result indicates that after about 7 ms , the thrust no longer decreases while intermittent unstable wave activity continues .
the asymptotic value of the thrust as determined from the run is 40.9 nn / m which is only 11% less than in v400 ( baseline ) when the difference in voltage between the runs ( 311 v versus 337 v ) is compensated for by a square root dependence .
hence the simulation suggests that the instability caused by an @xmath3-directed magnetic field decreases the plasma brake thrust only slightly at the baseline voltage .
figure [ fig : curvebzlong ] shows the corresponding result for @xmath4-directed magnetic field .
the behaviour is similar except that the fluctuation spectrum extends to somewhat lower frequencies .
the final state thrust 38.7 nn / m is 17% lower than voltage - corrected baseline .
figure [ fig : curvev1000bzlong ] shows the result with higher voltage ( 757 v which is 2.4 times larger than baseline ) and @xmath4-directed field .
the fluctuations are strong and the thrust 53.8 nn / m is 27% lower than voltage - corrected baseline .
thus the relative gap between the bz and @xmath34 cases increases with voltage .
the increase of the gap is slower than linear , however , since it increases from 17% to @xmath727% if the voltage is made 2.4 times stronger .
the thrust curve in fig .
[ fig : curvev1000bzlong ] might not yet be completely stabilised at the end of the run so the thrust estimate derived from the simulation in this case might somewhat overestimated . the @xmath3-directed magnetic field case with high 856 v voltage is shown in fig .
[ fig : curvev1000bx ] .
this run provides a positive surprise since initial transients now die away quickly after which the state is stable and the thrust is even somewhat larger than eq .
( [ eq : dfdz ] ) prediction .
thus , although @xmath3-directed field at lower voltage is unstable and the instability lowers the thrust to a small extent ( run bxlong ) , at higher voltage the instability is absent and instead of a small thrust reduction we have a minor thrust enhancement .
next we look at the asymptotic state in the strongly unstable proton case ( see subsection [ subsection : ionmass ] above ) where in addition a destabilising bz magnetic field exists .
the result is shown in fig .
[ fig : curveprotonbzlong ] .
the fluctuations are strong , but the average magnitude of the thrust gets stabilised during the run .
thus , even strongly unstable behaviour does not cause a collapse of the ion sheath which surrounds the tether and whose spatial extent determines the thrust .
interestingly , the determined thrust of 2.91 nn / m at @xmath35 v is exactly ( within three decimal places ) equal to eq .
( [ eq : dfdz ] ) prediction .
likely the @xmath34 value of the thrust in case of proton plasma would be slightly above eq .
( [ eq : dfdz ] ) prediction , while @xmath36 tends to slightly lower the thrust . that the effects cancel out exactly may be fortuitous .
we also made a run where the magnetic field components had equal values : @xmath37 nt@xmath38 .
the result is an intermediate case of the bx , by and bz runs : a very slowly growing oscillation appears .
it therefore seems that the behaviour is stable if the magnetic field is predominantly in the @xmath5 direction and unstable otherwise .
the result suggests that if the magnetic field has general orientation , the behaviour can be qualitatively interpolated from the x , y and z directed runs . when the plasma density @xmath39 is changed , all spatial scales in the electrostatic pic simulation model scale naturally by the electron debye length i.e. as @xmath40 .
the only thing that breaks this scale invariance is the fixed value of the tether s effective electric radius @xmath18 .
however , the effect of @xmath18 is minor because it enters in eq .
( [ eq : dfdz ] ) only logarithmically .
thus we can say that changing the plasma density @xmath39 does not affect the stable / unstable nature of the solution and the thrust scales nearly as proportional to @xmath41 . as a consistency check
, we made one run to verify this behaviour .
we also performed a high voltage run with @xmath5-directed field , v1000by ( table [ tab : runlist ] ) with short 2.43 ms duration because only a weak stable oscillation is present and no instability is seen . as in the corresponding lower voltage run by , the determined thrust 80.7 nn / m is in close agreement with eq .
[ eq : dfdz ] .
in eq . ( [ eq : dfdz ] ) the thrust is nearly linearly proportional to the ion mass and the simulations are in agreement with this at least qualitatively . at @xmath42
v , the runs bxlong and bzlong suggest that although the plasma sheath is unstable when an @xmath3 or @xmath4 dominant magnetic field is present , the thrust is reduced by the instability only modestly ( 11% and 17% , respectively ) .
when the voltage is increased to @xmath43 v , a @xmath4-directed field is again unstable ( v1000bzlong ) and the thrust reduction is larger , @xmath44% . in case of @xmath3-directed field , however , the instability is absent and there is actually a small thrust enhancement ( v1000bx ) .
a @xmath5-directed field is always stable and there is neither thrust reduction nor enhancement . in other words ,
( [ eq : dfdz ] ) is in most cases able to predict the thrust well .
the only exception is that if the magnetic field is aligned with the tether ( @xmath4-directed case ) , then there is a moderate thrust reduction which increases with voltage .
the relative reduction is 17% at 320 v and @xmath44% at 760 v for @xmath4-directed magnetic field . when @xmath34 , the flow is stable with oxygen plasma and unstable with proton plasma ( runs v400 and proton ) .
when an instability sets in , the ion cloud formed by ions passing near the tether starts to oscillate at ion plasma frequency . in the oxygen case
the ion plasma oscillation is 4 times slower than in the proton case . in the oxygen case
the flow has time to move by more than one ion sheath diameter during one ion plasma period , while in the proton case it moves only a fraction of the sheath diameter .
thus in the proton plasma case , if the sheath starts to oscillate , the oscillations have more opportunities to disturb the upstream flow and perhaps cause a positive nonlinear feedback .
maybe this is why lighter ion mass flow tends to be more unstable .
another way to arrive at a qualitatively similar conclusion is to note that the ion bulk flow kinetic energy @xmath45 is linearly proportional to the ion mass .
thus the ratio @xmath26 is 16 times larger in proton plasma flow than in o@xmath8 ion flow .
any voltage dependent instability should then occur at 16 times lower voltage in proton plasma than in o@xmath8 plasma .
let us consider a vertical gravity gradient stabilised plasma brake tether using the untilted dipole approximation for earth s magnetic field .
the untilted dipole approximation can be used in this case because runs where all field components had nonzero values seemed to interpolate smoothly from purely aligned runs ( section [ subsect : misc ] ) , i.e. exact orientation of the magnetic field does not seem to matter .
in this approximation , in equatorial orbit the magnetic field is @xmath5-directed all the time and the thrust is consequently predicted well by eq .
( [ eq : dfdz ] ) .
the same is true in polar orbit in low latitudes . only for the high latitude portion of a polar orbit
the thrust is somewhat smaller than what eq .
( [ eq : dfdz ] ) predicts .
let us look at a numeric example . at 1 kv voltage , @xmath46 m@xmath47 density and oxygen plasma , the predicted plasma brake thrust from eq .
[ eq : dfdz ] is 85 nn / m .
a 5 km long tether would produce 0.43 mn braking force which is equivalent to 13400 ns impulse per year . at 800 km altitude , reducing the orbital altitude by 100 km requires 52 m / s of delta - v , thus the exemplary 5 km tether ( mass 0.055 kg , @xcite ) could lower the orbital altitude of a 260 kg object by 100 km during one year .
the gathered oxygen ion current per tether length is given by the orbital motion limited ( oml ) theory expression as @xmath48 where @xmath49 is the total width of the four - wire tether ( 160 @xmath0 m ) .
hence a 5 km tether gathers 0.4 ma current and consumes 0.4 w power in this plasma . in reality
the current is expected to be somewhat larger because of production of some secondary electrons from the tether when hit by oxygen ions . for comparison ,
the maximal lorentz force produced by the tether @xcite ( using the length - averaged current of 0.2 ma ) is only 0.03 mn which is 14 times smaller than the plasma brake force in this case .
we did not test voltages higher than about 860 v in this paper .
going to higher voltages tends to increase the needed computing time because the sheath becomes larger so one needs more grid cells , particles and timesteps to model its asymptotic evolution accurately . in a practical device ,
if one increases the negative voltage , at some point electron field emission from the surface of the tether wires starts to become an issue .
field emission adds to the ionic current gathered by the tether and hence increases power consumption .
we think that the point where field emission starts to become an issue is larger than 1 kv but smaller than perhaps 3 - 5 kv .
the value also depends on the geometry and possible coating of the tether . before experimental knowledge about the plasma brake
is obtained , it is not necessarily well motivated to try simulations with higher voltages than those presented in this paper . the reference @xcite contains an error : the quantities which are in the present paper denoted by @xmath50 and @xmath11 were confused with each other .
therefore the thrust versus voltage relationship as given in @xcite was too optimistic .
on the other hand , @xcite used an earlier formula ( * ? ? ? * equation 20 ) whose numerical coefficient had been found from a test particle calculation using solar wind parameters . in the ionospheric plasma brake case ,
the ratio between the tether voltage @xmath11 and the bulk flow energy @xmath51 ( in voltage units ) is much larger than in the solar wind , and consequently ions are reflected backward more efficiently in the ionosphere than in the solar wind , as is evident from the simulation results of the present paper .
all in all , the results of @xcite are not too far from reality : the error resulting from confusing @xmath50 and @xmath11 and the inaccuracy resulting from using a formula suitable for solar wind roughly cancel out each other . according to a series of electrostatic pic simulations performed at different parameters , eq .
( [ eq : dfdz ] ) can be used to predict plasma brake thrust in ionospheric conditions . only if the dominant component of the magnetic field is along the tether , the thrust is reduced , and the relative reduction grows from 17% to @xmath44% when the voltage increases from 320 v to 760 v. the thrust reduction is due to an instability which has ionic character .
the predicted performance of the plasma brake seems promising concerning satellite deorbiting applications .
for example in o@xmath8 plasma with @xmath46 m@xmath47 density and using 1 kv voltage , a 5 km long plasma brake tether weighing 0.055 kg could produce 0.43 mn breaking force at altitude @xmath43 km , which is enough to reduce the orbital altitude of a 260 kg debris mass by 100 km during one year .
seppnen , h. , t. rauhala , s. kiprich , j. ukkonen , m. simonsson , r. kurppa , p. janhunen and e. haeggstrm , one kilometer ( 1 km ) electric solar wind sail tether produced automatically , rev .
instrum . , 84 , 095102 , 2013 .
lcr parameter & symbol & value + grid size & & @xmath56 + grid spacing & @xmath57 & @xmath58 mm + normalised spacing & @xmath59 & 0.54 + x grid domain & & -2.34 .. 1.4 m + y grid domain & & -1.87 .. 1.87 m + timestep & @xmath60 & @xmath61
ns + normalised timestep & @xmath62 & 0.057 + run duration & @xmath63 & @xmath64 ms + number of timesteps & & 417000 + electrons per cell & @xmath65 & 500 ( in plasma stream ) + number of particles & & @xmath66 m + plasma density & @xmath39 & @xmath46 m@xmath47 + ion mass & @xmath14 & 16 amu ( o@xmath8 ) + plasma drift & @xmath16 & 7.5 km / s + electron temp . &
@xmath67 & 0.1 ev + ion temp . &
@xmath68 & 0.1 ev + magnetic field & @xmath69,@xmath28,@xmath36 & 0 nt + tether voltage & @xmath11 & 337.436 v + tether electric radius & @xmath18 & 1 mm + llllrr run & parameters & thrust / nnm@xmath70 & eq . ( [ eq : dfdz])/nnm@xmath70 & rel .
thrust & nature + v50 & @xmath71 v & 7.06 & 10.6 & -33% & stable + v100 & @xmath72 v & 19.4 & 19.8 & -2% & stable + v150 & @xmath73 v & 27.3 & 26.7 & + 2% & stable + v200 & @xmath74 v & 32.8 & 32.3 & + 2% & stable + v300 & @xmath75 v & 40.5 & 41.2 & -2% & steady oscillation + v400 & table [ tab : simparams ] , @xmath76 v & 47.9 & 48.5 & -1% & steady oscillation + v500 & @xmath77 v & 54.5 & 54.9 & -1% & steady oscillation + v500 g & @xmath77 v & 54.9 & 54.9 & 0% & steady oscillation + v600 g & @xmath78 v & 60.9 & 60.6 & 0% & steady oscillation + v800 g & @xmath79 v & 71.4 & 70.7 & + 1% & steady oscillation + v1000 g & @xmath80 v & 80.1 & 79.5 & + 1% & steady oscillation + bx & @xmath81 t , @xmath82 v & 51.1 & 48.2 & + 6% & unstable + bxlong & @xmath81 t , @xmath83 v & 40.9 & 46.4 & -12% & unstable + by & @xmath84 t , @xmath85 v & 48.5 & 48.6 & 0% & dying oscillation + bz & @xmath86 t , @xmath87 v & 47.2 & 48.4 & -2% & unstable + bzlong & @xmath86 t , @xmath88 v & 38.7 & 46.9 & -17% & unstable + helium & @xmath89 amu , @xmath90 v & 14.6 & 13.2 & + 11% & unstable + proton & @xmath91 amu , @xmath92 v & 5.06 & 3.24 & + 56% & unstable + protonbzlong & @xmath91 amu , @xmath35 v & 2.91 & 2.91 & 0% & unstable + v1000by & @xmath84 t , @xmath93 v & 80.7 & 79.5 & + 2% & steady oscillation + v1000bzlong & @xmath86 t , @xmath94 v & 53.8 & 74.1 & -27% & unstable + v1000bx & @xmath81 t , @xmath95 v & 83.6 & 78.0 & + 7% & stable + te & @xmath55 ev , @xmath76 v & 48.7 & 48.5 & 0% & dying oscillation + | the plasma brake is a thin negatively biased tether which has been proposed as an efficient concept for deorbiting satellites and debris objects from low earth orbit .
we simulate the interaction with the ionospheric plasma ram flow with the plasma brake tether by a high performance electrostatic particle in cell code to evaluate the thrust .
the tether is assumed to be perpendicular to the flow .
we perform runs for different tether voltage , magnetic field orientation and plasma ion mass .
we show that a simple analytical thrust formula reproduces most of the simulation results well .
the interaction with the tether and the plasma flow is laminar ( i.e. , smooth and not turbulent ) when the magnetic field is perpendicular to the tether and the flow . if the magnetic field is parallel to the tether , the behaviour is unstable and thrust is reduced by a modest factor .
the case when the magnetic field is aligned with the flow can also be unstable , but does not result in notable thrust reduction .
we also fix an error in an earlier reference . according to the simulations ,
the predicted thrust of the plasma brake is large enough to make the method promising for low earth orbit ( leo ) satellite deorbiting . as a numerical example
we estimate that a 5 km long plasma brake tether weighing 0.055 kg could produce 0.43 mn breaking force which is enough to reduce the orbital altitude of a 260 kg object mass by 100 km during one year .
the plasma brake @xcite is an efficient propellantless concept for deorbiting low earth orbit ( leo ) satellites .
the plasma brake is a very thin negatively charged tether which , when charged , causes a braking force by creating enhanced coulomb drag with ambient ionospheric plasma ram flow .
the plasma brake tether is somewhat similar to the more well known electrodynamic tether @xcite , but is much thinner and uses electrostatic rather than magnetic forces .
because the plasma brake 4-wire ultrasonically bonded tether is thin , it is lightweight , 11 grams per kilometre @xcite and does not form an appreciable threat to other satellites even in case of tether breakage . a broken piece of plasma brake tether deorbits itself quickly because of electromagnetic forces and neutral drag , and even if the tether piece would collide with a satellite , the 25 - 50 @xmath0 m wires draw only @xmath1 mm deep scratches on its surface .
said electromagnetic forces are a passive coulomb drag effect and to a lesser extent the passive electrodynamic tether effect .
the broken tether piece experiences an orbital motion induced natural electric field , and whenever this field has a component along the tether , one end of the tether gets charged negatively and the other one positively with respect to the local plasma , causing electrostatic coulomb drag and electromagnetic lorentz drag .
although the passive electromagnetic drag effects are weaker than the active ones , the very low specific mass of the tether ( @xmath2 kg / km ) makes passive deorbiting of a broken tether piece fast , typically . to maintain a tether at positive voltage in the solar wind requires an electron gun @xcite that pumps out negatively charged particles from the system and thereby cancels the tether s gathered plasma current .
likewise , to maintain a tether at negative voltage can be accomplished with a positive ion gun . however , in leo the maintenance of the plasma brake tether s negative voltage in most cases does not require an ion gun since the satellite s conducting body can be used as the current balancing electron gathering surface @xcite . if the satellite s grounded and conducting surface area is insufficiently small for this purpose , a relatively short positively biased tether made of similar material than the main tether could be used for gathering the required balancing electron current .
the topic of the paper is to use a realistic electrostatic particle in cell ( pic ) simulation to model the interaction between the negatively biased plasma brake tether with the surrounding ionospheric plasma ram flow which moves at a typical satellite orbital speed of 7.5 km / s .
we run the code with different voltages , different magnetic field orientations and different ion species and determine the thrust per tether length in each case
. we will also present a simple analytical formula for plasma brake thrust which reproduces the simulation results . |
in the last decade nanoplasmonics led to the emergence of many promising applications @xcite .
one of them is a near - field generator of nanolocalized coherent optical fields spaser - based nanolaser or spaser ( surface plasmon amplification by stimulated emission of radiation ) , which was shown to be an optical counterpart of the mosfet ( metal - oxide - semiconductor field effect transistor ) @xcite . the device was proposed by d. bergman and m. stockman in the paper @xcite .
operation principles of the spaser - based nanolaser are similar to operation principles of the usual laser , but instead of photons we deal with surface plasmons ( sps ) .
the first experimental observations were made by m. noginov s group @xcite and x. zhang s group @xcite and they are dated back to 2009 year .
since the spaser - based nanolaser is a source of coherent light , it has a narrow spectral line width above the generation threshold .
there are two main possible mechanisms which lead to the spectrum narrowing . for high q - factor resonators ,
the narrowing is determined by the domination of stimulated emission of radiation and thus , by the large number of excited quanta in the resonator @xcite .
this situation is typical for usual lasers .
the opposite limit of low q - factor resonators and relatively small excited quanta corresponds to spaser operation @xcite . in the case
, the narrowing can be determined by large number of the excited atoms in the gain media and their coherence , which is achieved by the mutual interaction of atoms through the resonator mode @xcite .
here we show that the spaser - based nanolaser can produce narrow spectrum of generation even if the mean number of excited quanta in resonator is of the order or less than unity , and find the spectrum of the generation and its statistics .
the exact analytical treatment of lasing involves quantum fluctuations , which are responsible for finite width of the spectral line @xcite .
description of the laser with maxwell - bloch equations , see e.g. @xcite , corresponds to mean - field approximation both for the resonator mode and atoms .
we develop a theory which allows to account for quantum fluctuations in a low @xmath2-factor resonator with the arbitrary number of quanta both below and above the generation threshold , which interacts with ensemble of @xmath3 identical active atoms . to solve the problem completely analytically we impose some restrictions .
first , we assume that the cavity decay rate @xmath4 is the fastest rate in the system .
thus , the resonator mode can be adiabatically eliminated @xcite , and the state of the spaser - based nanolaser is fully characterized by the state of @xmath3 identical two - level active atoms .
second , we believe @xmath5 and thereby the fluctuations of the state of atoms can be considered is the small - noise limit ( * ? ? ?
* ch.5.1.3 ) .
note that due to adiabatic mode elimination we can only resolve times @xmath6 .
the smaller times were considered in the paper @xcite , for the model with single active atom , @xmath7 , but only below the generation threshold , when mean number of quanta in the resonator is well below unity . as a result
, we describe the behaviour of the spaser - based nanolaser below and above generation threshold , and demonstrate that the spectral line narrows considerably , when passing through the threshold .
we find the average number of plasmons in the cavity and show that this number near the generation threshold can be of the order of unity . in the case
the coherence is preserved in a state of active atoms , which relax slowly than @xmath8 .
this fact fundamentally distinguishes the behaviour of the bad - cavity nanolasers in comparison with the good - cavity lasers , where the coherence is preserved in a state of photons .
the evaluation for the number of plasmons is in accordance with the experimental observations @xcite .
moreover , we obtain second - order correlation function @xmath1 , @xmath6 , and find that above the generation threshold the amplitude fluctuations of polarization of active atoms lead to the damped oscillations in @xmath1 .
a similar dependence was observed in numerical simulations in the paper @xcite , and it is usual for bad - cavity lasers @xcite .
however , in the case of good - cavity lasers there is no oscillations in second - order correlation function @xcite .
we assume that the shape of the curve @xmath1 indicates the mechanism of the spectral line narrowing , and therefore we investigate at what relationship between cavity decay rate @xmath4 and homogeneous broadening of active atoms @xmath9 the oscillations occur .
we think that obtained results are important for understanding the fundamental principles of operation of spaser - based nanolasers .
we consider @xmath5 identical two - level active atoms with resonant frequency @xmath10 coupled to single strongly damped cavity , with a short plasmon lifetime @xmath11 centered at the same frequency .
the interaction between the atoms and the field is described by the tavis - cummings hamiltonian @xcite , @xmath12 , where @xmath13 is coupling constant , @xmath14 and @xmath15 are the creation and annihilation operators of plasmons in the cavity mode @xcite , and @xmath16 is collective atomic operators , where @xmath17 , @xmath18 pauly matrices , and @xmath19 . in a bad - cavity limit
@xcite the cavity mode can be adiabatically eliminated and the following master equation in the schrdinger picture for the atomic density operator @xmath20 : @xmath21 + \frac{\gamma_{\uparrow}}{2 } \left ( \sum_{j=1}^n 2 \sigma_{j+ } \rho \sigma_{j- } + \frac{1}{2 } j_z \rho + \frac{1}{2 } \rho j_z - n \rho \right ) + & \\ \nonumber & \displaystyle + \frac{\gamma_{\downarrow}}{2 } \left ( \sum_{j=1}^n 2 \sigma_{j- } \rho \sigma_{j+ } - \frac{1}{2 } j_z \rho - \frac{1}{2 } \rho j_z - n \rho \right ) + \frac{\gamma_{p}}{2 } \left ( \sum_{j=1}^n \sigma_{jz } \rho \sigma_{jz } -n \rho \right)+&\\ \label{eq:1 } & \displaystyle + \frac{g^2}{\kappa } \left(2 j_- \rho j_+ - j_+ j_- \rho - \rho j_+ j_- \right),&\end{aligned}\ ] ] where the trace is taken over the field variables . here
the active atoms are incoherently pumped with rate @xmath22 and we take into account the spontaneous emission with rate @xmath23 .
the dephasing processes , which are caused mostly by the interaction with phonons , has rate @xmath24 .
the last term describes interaction of active atoms through the cavity mode .
adiabatic mode elimination can be performed only if @xmath25 .
note , that normal - ordered field operator averages can be restored by formal substitutions @xmath26 and @xmath27 , ( * ? ? ?
* ( 13.60 ) ) . the same model , but with an arbitrary number @xmath3 of active atoms , was considered numerically by v. temnov and u. woggon in papers @xcite , where they showed that the last term in equation ( [ eq:1 ] ) leads to the cooperative effects .
our final interest is the state of the resonator , thus it is enough to describe the system of atoms in terms of collective atomic operators , despite the fact that some terms in ( [ eq:1 ] ) can not be rewritten in terms of @xmath28 .
first , we define characteristic function @xmath29 which determines all normal - ordered operator averages in usual way .
next , we introduce the glauber - sudarshan @xmath30-representation @xmath31 as the fourier transform of @xmath32 , which can be interpreted as distribution function and allows to calculate normal - ordered operator averages as in statistical mechanics ( * ? ? ?
* ( 6.118a ) ) .
evolution equation on @xmath33 can be obtained by differentiation ( [ eq:2 ] ) with respect to time and replacement @xmath34 with ( [ eq:1 ] ) .
the results is @xmath35 where @xmath36+&\\ \nonumber & \displaystyle + \frac{\gamma_{\uparrow}}{2 } \left ( 2e^{-2 \frac{\partial}{\partial m } } -1 + 2 \frac{\partial^2}{\partial v \partial v^ * } \right ) \left ( \frac{\partial}{\partial v } v + \frac{\partial}{\partial v^ * } v^ * \right)+&\\ \nonumber & \displaystyle+ \frac{\gamma_{\downarrow}}{2 } \left [ \left ( e^{2 \frac{\partial}{\partial m}}-1 \right ) ( n+m ) + \frac{\partial}{\partial v } v + \frac{\partial}{\partial v^ * } v^ * \right]+&\\ \nonumber & \displaystyle+ \gamma_{p } \left [ \frac{\partial}{\partial v } v + \frac{\partial}{\partial v^ * } v^ * + \frac{\partial^2}{\partial v \partial v^ * } e^{2 \frac{\partial}{\partial m } } ( n+m ) \right]+ i \omega \left [ \frac{\partial}{\partial v } v - \frac{\partial}{\partial v^ * } v^ * \right]+&\\ \nonumber & \displaystyle+ \frac{g^2}{\kappa } \left [ 2 \left ( 1 - e^{-2\frac{\partial}{\partial m } } \right ) v v^ * - \left ( \frac{\partial}{\partial v } vm + \frac{\partial}{\partial v^ * } v^*m \right ) + \frac{\partial^2}{\partial v^2 } v^2 + \frac{\partial^2}{\partial v^{*2 } } v^{*2}\right].&\end{aligned}\ ] ] the closed form of the equation confirms the possibility of describing the system in terms of collective atomic operators . the exact solution of the equation ( [ eq:3 ] ) is strictly singular due to the exponential factors in @xmath37 , which describe transitions in active atoms .
moreover , the solution can not be found in analytical form .
however , we can obtain an approximate nonsingular distribution , replacing ( [ eq:3 ] ) by a fokker - planck equation .
the key element to such replacement is a system size expansion procedure ( * ? ? ?
the large system size parameter in our case is the number @xmath5 of active atoms .
following the system size expansion method we will obtain an adequate treatment of quantum fluctuations in the first order in @xmath38 . to make a systematic expansion of the phase - space equation of motion in @xmath38
, we move into rotating frame and introduce dimensionless polarization per one atom @xmath39/n$ ] and inverse population @xmath40/n $ ] .
next , we separate the mean values and fluctuations in the phase - space variables @xmath41 and introduce a distribution function @xmath42 , which depends on variables , corresponding to fluctuations . using the equation ( [ eq:3 ] ) and neglecting terms @xmath43 , we obtain the equation for a scaled distribution function @xmath30 .
more accurately , we obtain two sets of equations : the first set describes dynamics of macroscopic variables , and one more equation characterizes fluctuations .
first , we analyze the system of equations describing the dynamics of macroscopic variables , which takes a form @xmath44 where we introduce population relaxation time @xmath45 , homogeneous broadening @xmath46 and equilibrium inverse population @xmath47 .
the system has two different stable steady - states solutions , depending on _ _
pump-__parameter @xmath48 , where @xmath49 . in the case @xmath50 , one
obtain the solution @xmath51 .
thus , there is no macroscopic polarization and this situation corresponds to the nanolaser operating below generation threshold . in the opposite case @xmath52
, the solution takes a form @xmath53 and it corresponds to the spaser - based nanolaser operating above generation threshold .
overall , the picture is completely analogous to the good - cavity laser ( * ? ? ?
8.1.2 ) .
note , that in the case @xmath54 both solutions are the same .
this point corresponds to the spaser generation threshold , which was obtained in earlier semiclassical papers , e.g. @xcite .
second , we analyze the equation , which provides a linearized treatment of fluctuations about solution to the system of macroscopic equations . in the case below generation threshold , i.e. @xmath50 , we obtain @xmath55 p + \frac{1}{t_1 } \frac{\partial}{\partial \mu } \mu p + \frac{2\gamma_{\uparrow } ( \gamma + 2 \gamma_{\downarrow})}{(\gamma_{\uparrow}+\gamma_{\downarrow } ) } \frac{\partial^2}{\partial \nu \partial \nu^*}p.\ ] ] the equation can be solved by separation of variables , and we calculate the steady - state correlation functions as in statistical mechanics @xmath56 the result for two - time correlation functions is analogous to the case of good - cavity laser , up to the replacement @xmath57 , since we adiabatically eliminate the cavity mode , rather than the polarization of active atoms ( * ? ? ?
note , that due to adiabatic mode elimination we can only resolve times @xmath6 .
the smaller times were resolved in paper @xcite , but only for the model with a single active atom . in the case @xmath58 the drift term in the fokker - planck equation ( [ eq:8 ] ) vanishes and there is no restoring force to prevent the fluctuations from growing without bound .
thus , the average number of plasmons in the cavity mode ( [ eq:9 ] ) diverges at the point @xmath58 .
thereby , the equation ( [ eq:8 ] ) can not correctly describe the behaviour of system at the generation threshold .
note , that the operation of a bad - cavity laser at the threshold was discussed in paper @xcite .
now , we turn out to the description of fluctuations above the generation threshold .
as follows from the steady - state solution in the case @xmath59 , the phase of polarization is undetermined .
thus , in place of the first equation in ( [ eq:4 ] ) , we write @xmath60 where now the variable @xmath61 represents real amplitude fluctuations , which can be both positive and negative , but must fall within the range @xmath62 , and the variable @xmath63 represents phase fluctuations .
the distribution function in scaled variables , normalized with respect to the integration measure @xmath64 , is defined by @xmath65 . in the case of small amplitude fluctuations above the generation threshold , @xmath66 , we can partially separate variables @xmath67 . moreover , in accordance with experimental papers @xcite , we believe @xmath68 . with these assumptions
the evolution of the distribution functions @xmath69 and @xmath70 are governed by equations : @xmath71a+ \frac{1}{t_1 } \frac{\partial}{\partial \mu } \mu a + \frac{\gamma_p}{4 } \left ( 1 + \frac{1}{\wp_0}\right ) \frac{\partial^2}{\partial \nu^2 } a,&\\ \label{eq:15 } & \displaystyle \frac{\partial \phi}{\partial t } = \gamma_p \frac{\gamma t_1 ( \wp_0 + 1 ) \wp_0}{(\wp-1 ) } \frac{\partial^2}{\partial \psi^2 } \phi.&\end{aligned}\ ] ] solving the equation ( [ eq:14 ] ) , we can calculate the average number of plasmons in the cavity mode above the generation threshold @xmath72.\ ] ] here the second term in the left part corresponds to the steady - state solution of macroscopic equations and the right part describes fluctuations . our theory is correct if fluctuations are small .
far enough away from the threshold , when @xmath73 , this leads to the restriction @xmath74 .
thus , our theory is self - consistent if @xmath75 .
, for the following parameters : @xmath76 .
the dashed line corresponds to the mean - field theory , the solid line takes into account quantum fluctuations.,title="fig:",width=264 ] + next , completely ignoring the small amplitude fluctuations in equation ( [ eq:12 ] ) and using equation ( [ eq:15 ] ) , we can obtain fist - order correlation function ( @xmath6 ) @xmath77 where @xmath78 defines the width of spectral line above the generation threshold , and we rewrite it in terms of the output power @xmath79 . qualitatively , the result is similar to the case of good - cavity lasers , compare with ( * ? ? ?
* ( 8.138 ) ) . however , the mechanism which leads to the narrowing of the spectral line is quite different
. we will discuss it in detail in the next section .
finally , we obtain the second - order correlation function , taking into account the amplitude fluctuations in the main order . in the region of interest ,
@xmath73 , one can find ( @xmath6 ) @xmath80 the damped oscillations in @xmath81 are usual for bad - cavity lasers @xcite .
the similar behaviour had been observed in numerical calculations in the paper @xcite , but there authors dealt with the regime of large fluctuations . as we discussed above , our consideration is reliable only in small - noise limit , i.e. @xmath82 .
note , that in the case of good - cavity lasers the amplitude fluctuations does not lead to the oscillations in @xmath81 , see ( * ? ? ?
* ( 8.139 ) ) .
in the next section we will explain the origin of damped oscillations in detail and establish when the good - cavity behaviour is replaced by the bad - cavity damped oscillations .
( b ) the vector field obtained from the right parts of the macroscopic equations with @xmath83 , @xmath84 and @xmath85 .
the spiral movement to the steady - state leads to the damped oscillations in @xmath86.,title="fig:",width=415 ] + in the fig .
[ pic:1 ] we plot the dependence of the average number of plasmons in the cavity on the pump - parameter @xmath87 below and above the generation threshold , where we propose the following sets of parameters : @xmath88 and we variate the parameter @xmath3 .
the dashed line corresponds to the semiclassical mean - field theory , see equations ( [ eq:5])([eq:7 ] ) , and the solid line takes into account quantum fluctuations , see equations ( [ eq:9 ] ) , ( [ eq:16 ] ) .
emphasize , that near the threshold @xmath89 the average number of plasmons in the cavity mode @xmath90 .
despite this fact , a linewidth of the order of @xmath9 well below threshold is replaced by a much narrower line of the order of @xmath78 above threshold .
when @xmath91 , we find @xmath92 . in the good - cavity lasers the spectral line width becomes narrower above the generation threshold because the stimulated emission starts to play more important role than the spontaneous emission . in this case the average number of photons in the cavity near the threshold is sufficiently greater than unity @xcite . in our case the average number of plasmons is less than unity .
thus , in order to spectral line width becomes narrower , the born plasmon should be coherent with the already dead plasmon , and this is possible since the active atoms preserve the coherence . originally arising plasmon interacts with active atoms and make them coherent to each other
. then the plasmon dies after a short time @xmath93 , but the coherence is still alive in active atoms , which relax slowly , @xmath94 .
the next plasmon generated by such atoms can be coherent to the previous one .
this mechanism of the spectral line narrowing was demonstrated in experiment with the laser , which deals with photons , in the paper @xcite .
however , we can not directly applied our theory to this experiment , since the assumption @xmath68 is not fulfilled .
another experimental realization of the lasing regime is @xcite , where mean number of plasmons is also less than unity , @xmath95 .
indeed , the total pumping energy absorbed per one nanolaser is @xmath96 , where @xmath97 is the duration of the pumping pulse .
the measured value @xmath98 near the generation threshold , @xmath99 , @xmath100 and @xmath101 .
note , that numerical parameters in the fig .
[ pic:1 ] are slightly different from those from the paper @xcite .
the reason is that for the experimental parameters the amplitude fluctuations are large , see ( [ eq:16 ] ) , and our theory is not applicable well .
however , we believe that the picture does not change qualitatively . next , in the fig . [ pic:2](a )
we plot the second - order correlation function above the generation threshold , according to the expression ( [ eq:18 ] ) .
the dependence is valid if @xmath102 , because it is obtained under the bad - cavity approximation . in order to explain the nature of the damped oscillations , we plot the vector field on the @xmath103-@xmath104-plane , see fig . [ pic:2](b ) , which corresponds to the right parts of the mean - field equations ( [ eq:5])-([eq:7 ] ) .
the red point represents the steady - state solution of these equations . fluctuations move the system from its equilibrium state and then it relaxes to the steady - state .
the spiral movement corresponds to the damped oscillations in polarization amplitude @xmath104 and inverse population @xmath103 , and as a consequence in the second - order correlation function . , for different pump - parameters @xmath87 .
the area below and to the right to the corresponding curves responds to the non - oscillating regime .
the painted area above the dotted line corresponds to the bad - cavity lasers.,title="fig:",width=264 ] + in the same way we can analyze a laser with a resonator of arbitrary @xmath2-factor , but we should go beyond the bad - cavity approximation . instead of macroscopic equations ( [ eq:5])-([eq:7 ] ) , we need to consider a full system of three maxwell - bloch equations @xcite .
the evolution of amplitude fluctuations around the steady - state solution is defined by three eigenvalues .
one eigenvalue is always real .
two others can be either real or complex conjugate , which corresponds to non - oscillating and oscillating character of the second - order correlation function respectively .
the eigenvalues are fully defined by three parameters : @xmath105 and @xmath87 . in the fig .
[ pic:3 ] we plot a `` phase diagram '' in logarithmic coordinates for different pump - parameters @xmath106 .
the area to the right and below to the corresponding curves responds to the non - oscillating regime .
the parameters from the painted area above the dotted line always correspond to the oscillations in @xmath86 .
this is an area of bad - cavity lasers , where the mechanism of spectral line narrowing is based on the coherence conservation in the state of active atoms .
the asymptotic behaviour ( @xmath107 ) of the dotted line was obtained numerically and it corresponds to the dependencies @xmath108 and @xmath109 . in the area below the dotted line both the oscillating and non - oscillating behaviour of @xmath86 is possible , depending on pump - parameter @xmath87 .
thus , the shape of the second - order correlation function provides insufficient information to obtain a mechanism of spectral line narrowing .
the answer on this question is contained in the `` phase diagram '' in the fig .
[ pic:3 ] .
to summarize , the quantum theory of a spaser - based nanolaser was presented .
we found that the average number of plasmons in the cavity mode near the generation threshold can be less than unity both in our theory and experiments @xcite . despite this fact , the spectral line width narrows sufficiently , when passing through the threshold .
we proposed that it is possible behaviour when the coherence preserved by the active atoms , which relax slowly than the damping of cavity mode occurs .
we also studied the amplitude fluctuations of polarization and concluded that they lead to the damped oscillations in the second - order correlation function @xmath110 above the generation threshold .
it is unusual behaviour for the good - cavity lasers , and we investigated in detail at what relationship between cavity decay rate @xmath4 and homogeneous broadening of active atoms @xmath9 the bad - cavity damped oscillations are replaced by non - oscillating regime . in the future , we plan to extend our consideration beyond to the small - noise limit .
we thank v.v . lebedev and v.p .
drachev for fruitful discussions .
the work was supported by grant xxx . | we present a quantum theory of a spaser - based nanolaser , under the bad - cavity approximation . we find first- and second - order correlation functions @xmath0 and @xmath1 below and above the generation threshold , and obtain the average number of plasmons in the cavity .
the latter is shown to be of the order of unity near the generation threshold , where the spectral line narrows considerably . in this case
the coherence is preserved in a state of active atoms in contradiction to the good - cavity lasers , where the coherence is preserved in a state of photons .
the damped oscillations in @xmath1 above the generation threshold indicate the unusual character of amplitude fluctuations of polarization and population , which become interconnected in this case .
obtained results allow to understand the fundamental principles of operation of nanolasers
. 99 mark i. stockman , `` nanoplasmonics : past , present , and glimpse into future , '' 2202922106 ( 2011 )
. mark i. stockman , `` the spaser as a nanoscale quantum generator and ultrafast amplifier , '' j. opt .
* 12 , * 024004113 ( 2010 ) .
d. bergman and m. stockman , `` surface plasmon amplification by stimulated emission of radiation : quantum generation of coherent surface plasmons in nanosystems , '' 02740214 ( 2003 ) .
m.a .
noginov , g. zhu , a.m. belgrave , r. bakker , v.m .
shalaev , e.e .
narimanov , s. stout , e. herz , t. suteewong , and u. wiesner , `` demonstration of a spaser - based nanolaser , '' nature * 460 , * 11101112 ( 2009 ) .
r.f .
oulton , v.j .
sorger , t. zentgraf , r .-
m .
ma , c. gladden , l. dai , g. bartal , and x. zhang,``plasmon lasers at deep subwavelength scale , '' nature * 461 , * 629632 ( 2009 ) .
m.o .
scully and m.sh .
zubairy , _ quantum optics _ ( cambridge university press , 1997 ) .
h.j .
carmichael , _ statistical methods in quantum optics 1 _ ( springer , new york , 2010 ) .
j.g .
bohnet , z. chen , j.m .
weiner , d. meiser , m.j .
holland and j.k .
thompson , `` a steady - state superradiant laser with less than one intracavity photon , '' nature * 484 , * 7881 ( 2012 ) .
v.m . parfenyev and s.s .
vergeles , `` intensity - dependent frequency shift in surface plasmon amplification by stimulated emission of radiation , '' 04382415 ( 2012 ) .
j.i .
cirac , `` interaction of a two - level atom with a cavity mode in the bad - cavity limit , '' 43544362 ( 1992 ) .
e.s .
andrianov , a.a .
pukhov , a.v .
dorofeenko , a.p . vinogradov and a.a .
lisyansky , `` spectrum of surface plasmons excited by spontaneous quantum dot transitions , '' jetp * 117 , * 205213 ( 2013 ) .
v. temnov and u. woggon , `` photon statistics in the cooperative spontaneous emission , '' 57745782 ( 2009 ) .
s. gnutzmann , `` photon statistics of a bad - cavity laser near threshold , '' epjd * 4 , * 109123 ( 1998 ) .
e.s .
andrianov , d.g .
baranov , a.a .
pukhov , a.v .
dorofeenko , a.p . vinogradov and a.a .
lisyansky , `` loss compensation by spasers in plasmonic systems , '' 1346713478 ( 2013 ) .
h.j .
carmichael , _ statistical methods in quantum optics 2 _ ( springer , new york , 2008 ) .
v. temnov and u. woggon , `` superradiance and subradiance in an inhomogeneously broadened ensemble of two - level systems coupled to a low - q cavity , '' 24360214 ( 2005 ) .
j. trieschmann , s. xiao , l.j .
prokopeva , v.p .
drachev , a.v .
kildishev , `` experimental retrieval of the kinetic parameters of a dye in a solid film , '' 1825318259 ( 2011 ) .
j. kim , v.p .
drachev , z. jacob , g.v .
naik , a. boltasseva , e.e .
narimanov , and v.m .
shalaev `` improving the radiative decay rate for dye molecules with hyperbolic metamaterials , '' 81008116 ( 2012 ) . |
the physical properties of mott ( or charge transfer ) insulators are dominated by large on - site coulomb interactions @xmath5 which suppress charge fluctuations . quite generally ,
the coulomb interactions lead then to strong electron correlations which frequently involve orbitally degenerate states , such as @xmath6 ( or @xmath7 ) states in transition metal compounds , and are responsible for quite complex behavior with often puzzling transport and magnetic properties.@xcite the theoretical understanding of this class of compounds , with the colossal magnetoresistance ( cmr ) manganites as a prominent example , @xcite has substantially advanced over the last decade , @xcite after it became clear that orbital degrees of freedom play a crucial role in these materials and have to be treated on equal footing with the electron spins , which has led to a rapidly developing field _ orbital physics_.@xcite due to the strong onsite coulomb repulsion , charge fluctuations in the undoped parent compounds are almost entirely suppressed , and an adequate description of these strongly correlated insulators appears possible in terms of superexchange.@xcite at orbital degeneracy the superexchange interactions have a rather rich structure , represented by the so - called spin - orbital models , discovered three decades ago,@xcite and extensively studied in recent years .
@xcite although this field is already quite mature , and the first textbooks have already appeared,@xcite it has been realized only recently that the _ magnetic _ and the _ optical _ properties of such correlated insulators with partly filled @xmath8 orbitals are intimately related to each other , being just different experimental manifestations of the same underlying spin - orbital physics . @xcite
while it is clear that the _ low - energy _ effective superexchange hamiltonian decides about the magnetic interactions , it is not immediately obvious that the _ high - energy _ optical excitations and their partial sum rules have the same roots and may be described by the superexchange as well .
in fact , this interrelation between the magnetic and the optical properties makes it necessary to reanalyze the spin - orbital superexchange models , and to extract from them important constraints imposed by the theory on the system parameters .
we will show that also the opposite holds some general rules apply for the magnetic interactions in the correlated insulators with degenerate ( or almost degenerate ) orbitals , and therefore the magnetic measurements impose constraints on any realistic theory . at the same time
, we shall argue that such experiments provide very useful information concerning the orbital order ( oo ) and the strength of quantum fluctuations in a given compound , which can next be employed to interpret other experiments , including the optical spectroscopy .
the phenomena discussed in the present paper go well beyond the more familiar situation of a mott insulator without orbital degeneracy , or when the orbital degeneracy is lifted by strong jahn - teller ( jt ) distortions as for example in the high @xmath9 cuprate superconductors . in a mott insulator
the optical conductivity is purely incoherent , and the optical response is found at energies which exceed the optical gap .
when orbital degrees of freedom are absent , the optical gap is determined by the intraorbital coulomb interaction element @xmath10 .
naively , one might expect that the high - energy charge excitations at energy @xmath11 , which contribute to the optical intensities , are unrelated to the low - energy magnetic phenomena . however , both energy scales are intimately related as the superexchange follows from the same charge excitations which are detected by the optical spectroscopy
. the prominent example of this behavior is the nondegenerate hubbard model , where the virtual high - energy excitations determine the superexchange @xcite energy @xmath1 it decides , together with spin correlations , about the spectral weight of the upper hubbard band at half - filling.@xcite when temperature increases to an energy scale @xmath12 , the spin correlations are modified and the total spectral weight in the optical spectroscopy follows these changes .
@xcite the superexchange models for transition metal perovskites with partly filled degenerate orbitals have a more complex structure than for nondegenerate orbitals and allow both for antiferromagnetic ( af ) and for ferromagnetic ( fm ) superexchange.@xcite these different contributions to the superexchange result from the multiplet structure of excited transition metal ions which depends on the hund s exchange @xmath13 and generates a competition between high - spin and low - spin excitations .
the exchange interactions are then intrinsically frustrated _ even on a cubic lattice _ , which enhances quantum effects both for @xmath3 , @xcite and for @xmath0 systems.@xcite this frustration
is partly removed in anisotropic af phases , which break the cubic symmetry and effectively may lead to dimensionality changes , such as in @xmath14-type af phase realized in lamno@xmath4 , or in @xmath15-type af phase in lavo@xmath4 .
while rather advanced many - body treatment of the quantum physics characteristic for spin - orbital models is required in general , we want to present here certain simple principles which help to understand the heart of the problem and give simple guidelines for interpreting experiments and finding relevant physical parameters of the spin - orbital models of _ undoped _ cubic insulators .
we will argue that such an approach based upon classical oo is well justified in many known cases , as quantum phenomena are often quenched by the jahn - teller ( jt ) coupling between orbitals and the lattice distortions , which are present below structural phase transitions and induce oo both in spin - disordered and in spin - ordered phases .
@xcite however , we will also discuss the prominent example of lavo@xmath4 , where assuming perfect oo or attempts to decouple spin and orbital fluctuations,@xcite fail in a spectacular way and give no more than a qualitative insight into certain limiting cases .
significant corrections due to quantum phenomena that go beyond such simplified approaches are then necessary for a more quantitative understanding . in the correlated insulators with partly occupied degenerate orbitals not only the structure of the superexchange
is complex , but also the optical spectra exhibit strong anisotropy and temperature dependence near the magnetic transitions , as found in lamno@xmath4 , @xcite the cubic vanadates lavo@xmath4 and yvo@xmath4 , @xcite and in the ruthenates.@xcite in all these systems several excitations contribute to the excitation spectra , so one may ask how the spectral weight redistributes between individual subbands originating from these excitations .
the spectral weight distribution is in general anisotropic already when oo sets in and breaks the cubic symmetry , but even more so when @xmath14-type or @xmath15-type
af spin order occurs below the nel temperature .
the effective spin - orbital models of transition metal oxides with partly filled degenerate orbitals depend in a characteristic way upon those aspects of the electronic structure which decide whether a given strongly correlated system can be classified as a mott insulator or as a charge transfer ( ct ) insulator .
as suggested in the original classification of zaanen , sawatzky and allen,@xcite the energy of the @xmath16 ct excitation @xmath17 has to be compared with the coulomb interaction @xmath10 if @xmath18 , the first excitation is at a transition metal ion and the system is a mott insulator , otherwise it is a ct insulator .
both are strongly correlated insulators , yet in one limit the dominant virtual excitations are of @xmath19 type , whereas in the other limit they are of @xmath20 type .
one may consider this issue more precisely by analyzing the full multiplet structure , and comparing the lowest excitation energy ( to a high - spin configuration ) _ at a transition metal ion _
, @xmath21 , with that of the lowest ct excitation ( of energy @xmath17 ) _ between a transition metal ion and a ligand ion_. @xcite thus we argue that one can regard a given perovskite as a _ charge transfer insulator _ if @xmath22 , and as a _ mott - hubbard insulator _ if @xmath23 . by analyzing these parameters
it has been suggested that the late transition metal oxides may be classified as ct insulators.@xcite in this case important new contributions to the superexchange arise , @xcite called below @xmath17 ( charge transfer ) terms , as we shall discuss for two @xmath3 systems : kcuf@xmath4 and lamno@xmath4 .
a central aim of this paper is to provide relatively simple expressions for the magnetic exchange constants and for the optical spectral weights that can be used by experimentalists to analyze and compare their spin wave data with optical data .
while the full spin - orbital models are rather complex , they are nevertheless controlled by only very few physical parameters : ( i ) the superexchange constant @xmath1 , ( ii ) the normalized hund s exchange @xmath2 , and ( iii ) the charge transfer parameter @xmath24 .
there are two distinct ways to determine these effective parameters : either ( i ) from the original multiband hubbard model , or ( ii ) from experimental spin wave and/or optical data . here
the second approach is of particular interest because the simultaneous analysis of magnetism and optics provides a subtle test of the underlying model .
the paper is organized as follows . in sec .
ii we introduce the generic structure of the low - energy effective hamiltonian in a correlated insulator with orbital degeneracy , and discuss its connection with the optical excitations at high - energy .
this general formulation provides the important subdivision of a given spin - orbital model which is necessary to obtain the partial spectral weights for individual multiplet transitions . in the remaining part of the paper we concentrate on some selected cubic perovskites and
demonstrate that this general formulation allows one to arrive at a consistent interpretation of the magnetic and optical experiments in these correlated insulators using the superexchange interactions ( secs .
iii - vi ) , and to deduce the parameters relevant for the theoretical model from the experimental data , whereever available .
we start in sec .
iii with the simplest spin - orbital model for @xmath3 holes in kcuf@xmath4 , and demonstrate that this system is in the ct regime which changes the commonly used picture of superexchange in this system in a qualitative way .
next we present and analyze the spin - orbital model with @xmath3 orbital degrees of freedom for the undoped manganite lamno@xmath4 in sec .
we show that in this case much smaller contributions arise from the ct processes , and the system is already in the mott - hubbard regime of parameters , which explains the earlier success of a simplified effective model based entirely on @xmath19 excitations and sufficient for a semiquantitative understanding .
this justifies our approach to the early transition metal perovskites with @xmath0 degrees of freedom , titanates in sec .
v and vanadates in sec .
vi , which we treat as mott - hubbard insulators .
for all these systems we analyze the magnetic exchange interactions and the optical spectral weights , depending on the nature of the spin correlations in the ground state .
the paper is concluded in sec .
vii , where we provide a coherent view on the magnetic and the optical phenomena and summarize the experimental constraints on the model parameters .
we consider here effective models with hopping elements between transition metal ions , @xmath25 here @xmath26 are orbital energies , and @xmath27 are effective hopping elements via ligand orbitals they depend on the type of considered orbitals as discussed in refs . and .
the energy scale for the hopping is set by the largest hopping element @xmath28 : the @xmath29 element in case of @xmath3 systems , and the @xmath30 element when only @xmath31 bonds are considered in systems with degenerate and partly filled @xmath0 orbitals . for noninteracting electrons the hamiltonian @xmath32 would lead to tight - binding bands , but in a mott insulator the large coulomb interaction @xmath10 suppresses charge excitations in the regime of @xmath33 , and the hopping elements
can only contribute via virtual excitations , leading to the superexchange .
the superexchange in the @xmath6 _ cubic _ systems with orbital degeneracy is described by spin - orbital models , where both degrees of freedom are coupled and the orbital state ( ordered or fluctuating ) determines the spin structure and excitations , and vice versa . the numerical and analytical structure of these models represents a fascinating challenge in the theory , as it is much more complex than that of pure spin models .
the spin - orbital models have been derived before in several cases , and we refer for these derivations to the original literature.@xcite they describe in low energy regime the consequences of virtual charge excitations between two neighboring transition metal ions , @xmath34 , which involve an increase of energy due to the coulomb interactions .
such transitions are mediated by the ligand orbitals between the two ions and have the same roots as the superexchange in a mott insulator with nondegenerate orbitals@xcite at @xmath33 thus the resulting superexchange interactions will be called _
@xmath10 terms_. the essential difference which makes it necessary to analyze the excitation energies in each case separately is caused by the existence of several different excitations .
their energies have to be determined first by analyzing the eigenstates of the local coulomb interactions , @xmath35 with @xmath36 , which in the general case depend on the three racah parameters @xmath14 , @xmath37 and @xmath15,@xcite which may be derived from somewhat screened atomic values . while the intraorbital coulomb element @xmath38 is identical for all @xmath6 orbitals , the interorbital coulomb and exchange elements , @xmath39 and @xmath40 , are anisotropic and depend on the involved pair of orbitals ; the values of @xmath40 are given in table [ tab : uij ] .
the coulomb and exchange elements are related to the intraorbital element @xmath10 by a relation which guarantees the invariance of interactions in the orbital space , @xmath41 .2 cm cccccc orbital & @xmath42 & @xmath43 & @xmath44 & @xmath45&@xmath46 @xmath42 & @xmath47 & @xmath48 & @xmath48 & @xmath15 & @xmath49 @xmath43 & @xmath48 & @xmath47 & @xmath48 & @xmath48 & @xmath50 @xmath44 & @xmath48 & @xmath48 & @xmath47 & @xmath48 & @xmath50 @xmath45 & @xmath15 & @xmath48 & @xmath48 & @xmath47 & @xmath49 @xmath46 & @xmath49 & @xmath50 & @xmath50 & @xmath49 & @xmath47 [ tab : uij ] in cases where only the orbitals of one type ( @xmath3 or @xmath0 ) are partly filled , however , as e.g. in the titanates , vanadates , or copper fluorides , all relevant exchange elements @xmath40 are the same ( see table [ tab : uij ] ) and one may use a simplified form of onsite interactions,@xcite @xmath51 with only two parameters : the coulomb element @xmath10 ( [ u ] ) and a hund s exchange element @xmath13 , being @xmath49 for @xmath3 and @xmath48 for @xmath0 systems , respectively .
we emphasize that in the general case when both types of orbitals are partly filled ( as in the manganites ) and both thus participate in charge excitations , the hamiltonian ( [ hee ] ) is only approximate , and the full excitation spectra of the transition metal ions@xcite which follow from eq .
( [ heegen ] ) have to be used instead
. a few examples of spectra for @xmath34 charge excitations at transition metal ions are shown in fig .
[ fig : levels ] . as a universal feature
, the high - spin excitation is found at energy @xmath52 in all cases , provided that @xmath13 is understood as hund s exchange for that partly filled manifold ( @xmath3 or @xmath0 ) of degenerate @xmath8 orbitals which participate in charge excitations .
the structure of the excited states depends on the partly occupied orbitals @xcite and on the actual valence @xmath53 the distance between the high - spin and low - spin excitations increases with the number of electrons for @xmath54 ( holes for @xmath55 ) . at orbital degeneracy
the superexchange which connects ions at sites @xmath56 and @xmath57 along the bond @xmath58 involves orbital operators which depend on the bond direction .
therefore , we introduce the index @xmath59 to label the three _ a priori _ equivalent directions in a cubic crystal . in order to analyze the consequences of each individual charge excitation @xmath60 that contributes to the superexchange in a given transition metal compound with degenerate @xmath8 orbitals
, we shall use below a general way of writing the effective low - energy hamiltonian as a superposition of such individual terms on each bond @xmath58 , @xmath61 with the energy unit [ absorbed in individual @xmath62 terms ] given by the superexchange constant , @xmath63 it follows from @xmath19 charge excitations with an effective hopping element @xmath28 between transition metal ions , and is the same as that obtained in a mott insulator with nondegenerate orbitals in the regime of @xmath33.@xcite while @xmath10 is the uniquely defined on - site _
intraorbital coulomb element _
( [ u ] ) , increasing upon going from ti to cu along the transition metal series,@xcite the definition of the hopping @xmath28 between two nearest neighbor transition metal ions depends on the system.@xcite if degenerate @xmath3 orbitals are involved , it is the effective @xmath29 hopping element for a @xmath64-bond which involves @xmath65 orbitals on the intervening ligand ion ( e.g. for the hopping between two directional @xmath66 states along the @xmath67 axis ) , while for the systems with degenerate @xmath0 orbitals it stands for the effective @xmath30 hopping element due to @xmath31 bonds which involve @xmath68 orbitals on the ligand ion .
charge excitations in selected cubic transition metal oxides , as obtained from eq .
( [ hee ] ) for : ( a ) @xmath3 excitations of cu@xmath69 ( @xmath70 ) and mn@xmath71 ( @xmath72 ) ions [ in the @xmath72 case the spectrum@xcite was obtained from eq .
( [ heegen ] ) ] ; ( b ) @xmath0 excitations of ti@xmath71 ( @xmath73 ) and v@xmath71 ( @xmath74 ) ions .
the splittings between different states are due to hund s exchange element @xmath13 which refers to a pair of @xmath3 electrons in ( a ) , and to a pair of @xmath0 electrons in ( b ) , respectively . , width=291 ] in the superexchange hamiltonian eq .
( [ hj ] ) the contributions which originate from all possible virtual excitations @xmath34 just add up to the total superexchange interaction , in which the individual terms can not be distinguished .
yet each of these excitations involves a different state in the multiplet structure of at least one of the transition metal ions , i.e. , either in the @xmath75 or in the @xmath76 configuration or in both , depending on the actual process and on the value of @xmath53 .
as pointed out elsewhere,@xcite the same charge excitations contribute to the optical conductivity , and here they appear at distinct energies , thus revealing the multiplet structure of the excited transition metal ions .
moreover , they convey a rich and characteristic _ temperature dependence _ to the optical spectrum , determined by the temperature variation of the spin - spin and orbital - orbital correlations .
we emphasize that it is therefore important to analyze the various multiplet excitations separately , as they depend on these correlations in a different way , and will also contribute to a _ quite different temperature dependence _
, as we show in this paper on several examples .
as we will see in more detail below , the generic structure of each such individual contribution is for a bond @xmath77 given by @xmath78 where the orbital dependence of the superexchange is described by means of orbital projection operators @xmath79 which are expressed in terms of components of orbital pseudospin @xmath80 operators at sites @xmath56 and @xmath57 .
the coefficients @xmath81 and @xmath82 , which measure the strength of the purely orbital part and of the spin - and - orbital part of the superexchange , respectively , follow from second - order perturbation theory involving the charge excitation @xmath60 . in the present case of perovskites , where the bond between two transition metal ions through the ligand ion ( f or o ) connecting them is close to linear ( 180@xmath83 ) , the coefficients @xmath81 and @xmath82 are of similar magnitude ( in contrast to the situation in layered compounds like linio@xmath84 with 90@xmath83 bonds where the purely orbital interaction is stronger by an order of magnitude than the spin - and - orbital interaction@xcite ) . here
we consider systems having cubic symmetry at high temperature . yet at low temperature this symmetry is frequently spontaneously broken usually driven by the joint effect of ( i ) the orbital part of the superexchange interaction , and ( ii ) the jt coupling of the same degenerate ( and therefore jt active ) @xmath6 orbitals to lattice modes .
the result is the simultaneous onset of a macroscopic lattice distortion and of oo , i.e. , a cooperative jt effect . at temperatures well below the transition temperature @xmath85 of this combined structural and orbital phase transition ,
the oo is effectively frozen .
the remaining superexchange interactions between the spins may then be obtained by replacing the orbital projection operators in eq .
( [ hj ] ) by their expectation values , @xmath86 obviously , this leads to anisotropic magnetic interactions , @xmath87 which will in general induce a further magnetic phase transition at lower temperature .
it is noteworthy that in this situation the spin degrees of freedom get decoupled from the orbital degrees of freedom , although the purely orbital ( @xmath81 ) and spin - and - orbital ( @xmath82 ) superexchange terms are of similar strength .
responsible for this behavior is the jt contribution to the structural phase transition , which enhances @xmath85 above the value it would have if the transition were driven by orbital superexchange alone .
starting from the microscopic spin - orbital superexchange models , we will analyze the effective spin models which arise after such a symmetry breaking at low temperature .
rewritten from eq .
( [ hspinse ] ) , they are of the generic form @xmath88 with two different effective magnetic exchange interactions : @xmath89 along the @xmath67 axis , and @xmath90 within the @xmath91 planes .
the latter @xmath90 interactions could in principle still take two different values in case of inequivalent lattice distortions ( caused , e.g. , by octahedra tilting or pressure effects ) making the @xmath92 and @xmath93 axes inequivalent , but we intend to limit the present analysis to idealized structures with these two axes being equivalent .
we shall show that the spin - spin correlations along the @xmath67 axis and within the @xmath91 planes , @xmath94 next to the orbital correlations , play an important role in the intensity distribution in optical spectroscopy .
the spectral weight in the optical spectroscopy is determined by the kinetic energy,@xcite and reflects the onset of magnetic order @xcite and/or orbital order.@xcite as shown by ahn and millis,@xcite in the weak coupling regime one can analyze the total spectral weight in optical absorption using the hartree - fock approximation for the relevant tight - binding hamiltonian . in a correlated insulator
the electrons are almost localized and the only kinetic energy which is left is associated with the same virtual charge excitations that contribute also to the superexchange .
therefore , we will discuss here the individual kinetic energy terms @xmath95 , which can be determined from the superexchange ( [ hj ] ) using the hellman - feynman theorem,@xcite @xmath96 for convenience , we define the @xmath95 as positive quantities .
making use of the generic form of the superexchange contribution @xmath62 given by eq .
( [ hngamma ] ) , and assuming as above that spin and orbital degrees of freedom are decoupled in the temperature range of interest , we obtain @xmath97 each term @xmath95 ( [ hefa ] ) originates from a given charge excitation @xmath60 for a bond @xmath58 along axis @xmath98 .
these terms are directly related to the _ partial optical sum rule _ for individual hubbard bands , which reads@xcite @xmath99 where @xmath100 is the contribution of band @xmath60 to the optical conductivity for polarization along the @xmath98 axis , @xmath101 is the distance between transition metal ions , and the tight - binding model with nearest neighbor hopping is implied .
comparison with eq .
( [ kngamma ] ) shows that the intensity of each band is indeed determined by the underlying oo together with the spin - spin correlation along the direction corresponding to the polarization .
one has to distinguish the above partial sum rule ( [ opsa ] ) from the sum rule for the total spectral weight in the optical spectroscopy for polarization along a cubic direction @xmath98 , involving @xmath102 which stands for the total intensity in the optical excitations ( due to @xmath19 excitations ) .
this quantity is of less interest here as it has a much weaker temperature dependence and does not allow for a direct insight into the nature of the electronic structure .
in addition , it might be also more difficult to resolve from experiment . when the low - energy excitations are of ct type , two holes could also be created within a @xmath103 orbital on a ligand ( oxygen or fluorine ) ion in between two transition metal ions , described by @xmath104 processes these ct contributions lead to additional superexchange contributions , called below _
terms_. while the latter terms can be safely neglected in mott - hubbard systems , they substantially modify the superexchange in ct insulators , and may even represent there the dominating contribution.@xcite below we will analyze them in two situations which involve @xmath3 degrees of freedom , viz . in the cubic copper fluoride kcuf@xmath4 ( sec .
iii ) , and in the cubic manganite lamno@xmath4 ( sec .
iv ) , and we will show that in kcuf@xmath4 they represent an essential part of the superexchange .
the simplest spin - orbital models are obtained when transition metal ions are occupied by either one electron ( @xmath105 ) , or by nine electrons ( @xmath106 ) ; in these cases the coulomb interactions ( [ hee ] ) contribute only in the excited state ( in the @xmath73 or the @xmath70 configuration ) after a charge excitation @xmath34 between two neighboring ions . here
we start with the case of a single hole in the @xmath8 shell , as realized for the cu@xmath71 ions in kcuf@xmath4 with the @xmath107 configuration ( @xmath106 ) . due to the splitting of the @xmath6 states in the octahedral field within the cuf@xmath108 octahedra
, the hole at each magnetic cu@xmath71 ion occupies one of the @xmath3 orbitals .
the superexchange coupling ( [ hj ] ) is usually analyzed in terms of @xmath3 holes in this case,@xcite and this has become a textbook example of spin - orbital physics by now.@xcite orbital order occurs in kcuf@xmath4 below the structural transition at @xmath109 k. at @xmath110 the structure is tetragonal , with longer cu@xmath111cu distances within the @xmath91 planes ( @xmath112 ) than along the @xmath67 axis ( @xmath113 ),@xcite which favors strong af interactions along the @xmath67 axis . below the magnetic transition at @xmath114 k , long - range magnetic order of @xmath14-type sets in , @xcite and
the ordered moment is @xmath115 .
@xcite the superexchange between the cu@xmath71 ions in kcuf@xmath4 , @xmath116 consists of two terms : the @xmath10 term @xmath117 ( [ hj ] ) , and the ct term @xmath118 .
first we introduce the @xmath10 term @xmath119 following the general approach of sec .
[ sec : general ] .
it originates from three different excitations , leading to an intermediate @xmath70 configuration at an excited cu@xmath69 ion . using the model hamiltonian ( [ hee ] ) to describe the coulomb interactions between the @xmath3 electrons
, one finds an equidistant excitation spectrum of @xmath120 , @xmath121 ( @xmath122 and @xmath123 ) and @xmath124 states , with energies:@xcite @xmath52 , @xmath125 and @xmath126 , as shown in fig .
[ fig : levels](a ) .
this excitation spectrum is exact , and the element @xmath13 for a pair of @xmath3 electrons is given by the racah parameters @xmath37 and @xmath15 ( see table i ) : @xmath127 this definition of @xmath13 will be used for two systems with @xmath3 orbital degrees of freedom : for the copper fluoride kcuf@xmath4 ( considered here ) , and for the manganite lamno@xmath4 ( in sec .
[ sec : d4 ] ) . in
what follows , we will parametrize the multiplet structure of the different transition metal ions by the ratio of the hund s element @xmath13 and the intraorbital coulomb element @xmath10 , @xmath128 using eqs .
( [ hj ] ) and ( [ eta ] ) , one finds for each bond @xmath58 along a @xmath98 axis ( @xmath59 ) four contributions:@xcite @xmath129 with coupled spin and orbital operators .
the coefficients , @xmath130 follow from the above multiplet structure of @xmath70 ions.@xcite as explained below , it is straightforward to understand the generic structure of the superexchange term @xmath117 , given by eqs .
( [ som91])([som94 ] ) . here
@xmath131 is a spin @xmath132 operator , and @xmath133 are the spin projection operators on triplet ( @xmath134 ) and singlet ( @xmath135 ) spin states for a bond @xmath58 , so one recognizes the high - spin term ( [ som91 ] ) and three low - spin terms ( [ som92])([som94 ] ) , respectively .
the spin operators in eqs .
( [ som91])([som94 ] ) are accompanied by orbital pseudospin operators @xmath136 , which select the type of orbitals _ occupied by holes _ at sites @xmath56 and @xmath57 , and simultaneously dictate the allowed excited states .
the orbital operators @xmath137 depend on the direction of a considered bond @xmath58 , and are given by @xmath138 where @xmath139 and @xmath140 are pauli matrices acting on the orbital pseudospins and the signs @xmath141 in @xmath142 correspond to @xmath92 and @xmath93 axis , respectively . with the help of @xmath143 one
defines next the projection operators in the orbital subspace , @xmath144 for a given cubic axis @xmath98 they project ( at site @xmath56 ) either on the planar orbital @xmath145 in the plane perpendicular to the @xmath98 axis , or on the orthogonal directional orbital @xmath146 along this axis .
for instance , in the case where @xmath98 is the @xmath67 axis , they project on the @xmath147 orbital in the @xmath91 plane , and on the directional @xmath66 orbital along the @xmath67 axis . using the projection operators ( [ orbpro ] ) , the orbital dependence in eqs . ( [ som91])-([som94 ] )
becomes transparent .
first of all , @xmath148 in eq .
( [ som91 ] ) accompanies the high - spin @xmath120 excitation as this state may occur only when a pair of orthogonal orbitals is occupied at sites @xmath56 and @xmath57 , described formally by a superposition of two possibilities , @xmath149 .
in contrast , the operator @xmath150 selects two orbitals oriented along the bond @xmath58 for the high - energy low - spin @xmath124 excitation , see eq .
( [ som94 ] ) .
finally , the second @xmath151 ( [ som92 ] ) and third @xmath152 ( [ som93 ] ) term correspond to the doubly degenerate low - spin @xmath121 state which consists of two singlet excitations : ( i ) an interorbital singlet with two different orbitals occupied ( @xmath123 ) , and ( ii ) a double occupancy within a directional orbital at either site ( @xmath122 ) these two excitations have thus quite different orbital dependences , identical with those of the @xmath120 and the @xmath124 excitation , respectively . the sum over all the terms @xmath153 , with @xmath154 ,
gives the simplest version of the spin - orbital model for the cubic copper fluoride kcuf@xmath4 with _ degenerate _
@xmath3 orbitals .
its derivation and more details on the classical phase diagram can be found in ref . .
by considering further the electronic structure of kcuf@xmath4 , one can elucidate the role played by the ct part @xmath118 in the superexchange hamiltonian ( [ som9 ] ) . by analogy with the cuo@xmath84 planes of the high - temperature superconductors , where the ct processes give the dominating contribution to the af superexchange interaction ,
@xcite one expects that they are also important for a cubic copper(ii ) fluoride and modify the superexchange in kcuf@xmath4 . the ct term , @xmath155 with the coefficient @xmath156 resulting from the two - hole charge excitation at a common neighboring @xmath157 orbital of a fluorine ion in between two copper ions , in the process @xmath158 . as a double hole excitation
is generated at a single bonding orbital @xmath157 within each cu@xmath111f@xmath111cu unit , this term is necessarily af .
two holes can move to fluorine from two neighboring cu ions only if both of them occupy a directional @xmath3 orbital @xmath146 , oriented along the considered bond ( e.g. , @xmath159 orbitals along the @xmath92 axis ) , being the simplest ct term discussed recently by mostovoy and khomskii.@xcite therefore this process leads to the same orbital dependence in eq .
( [ som9d ] ) as the low - spin @xmath122 and @xmath124 excitations which involve double occupancies of directional @xmath146 orbitals .
a characteristic feature of spin - orbital superexchange models with @xmath3 orbital degrees of freedom is the presence of the purely orbital interactions in eqs .
( [ som91])([som94 ] ) and ( [ som9d ] ) , which favor particular type of occupied orbitals .
lda+u calculations @xcite have indicated that such purely electronic interactions would already drive the instability towards the @xmath15-type oo ( @xmath15-oo ) phase , with _ alternating _ orbitals in the @xmath91 planes , and _ repeated _ orbitals along the @xmath67 axis , which induces fm spin exchange in the @xmath91 planes , and strong af exchange between the planes .
experimentally , this oo sets in below the structural transition at @xmath109 k,@xcite i.e. , at much higher temperature than the characteristic energy scale of the magnetic excitations,@xcite suggesting that the jt effect plays an important role in this instability .
this observation is consistent with the large difference between @xmath85 and the nel temperature @xmath160 k , @xcite the latter being controlled by the magnetic part of the superexchange , and thus the orbital correlations decouple from the spin - spin correlations .
this motivates one to analyze the dependence of the magnetic exchange interactions and of the optical spectral weights on the type of oo stabilized below the structural transition .
here we are interested in the low temperature range of @xmath161 k , so we assume perfect oo given by a classical ansatz for the ground state , @xmath162 with the orbital states , @xmath163 and @xmath164 , characterized by opposite angles ( @xmath165 ) and alternating between two sublattices @xmath14 and @xmath37 in the @xmath91 planes .
the orbital state at site @xmath56 : @xmath166 is here parametrized by an angle @xmath167 which defines the amplitudes of the orbital states @xmath168 ) in fig .
[ fig : eg ] .
the oo state specified in eq .
( [ oo ] ) is thus defined by : @xmath169 with @xmath170 and @xmath171 .
orbitals as obtained for different angle @xmath167 in eq .
( [ mixing ] ) .
pairs of orthogonal orbitals , forming a basis in the orbital space , differ by angle @xmath172 .
directional @xmath3 orbitals @xmath146 along @xmath92 and @xmath93 axes are obtained from the @xmath173 basis ( [ egbasis ] ) by the transformation ( [ mixing ] ) with angle @xmath174 . dashed lines for @xmath175
indicate the possible oo of @xmath176 orbitals occupied by holes in kcuf@xmath4 , suggested for @xmath177 in ref . .
, width=264 ] the magnetic superexchange constants @xmath90 and @xmath178 in the effective spin model ( [ hs ] ) are obtained by decoupling spin and orbital variables and next averaging the orbital operators in the spin - orbital model ( [ som9 ] ) over the classical state @xmath179 as given by eq .
( [ oo ] ) .
the relevant averages are given in table [ tab : eg ] , and they lead to compact expressions for the superexchange constants in eq .
( [ hs ] ) , @xmath180 which depend on three parameters , viz . @xmath1 ( [ j ] ) , @xmath2 ( [ eta ] ) and @xmath24 ( [ r ] ) , and on the oo ( [ ood9 ] ) specified by the orbital angle @xmath167 .
it is clear that the fm term @xmath181 competes with all the other af low - spin terms . nevertheless , in the @xmath91 planes , where the occupied @xmath3 orbitals alternate , the large fm contribution ( when @xmath182 ) still makes the magnetic superexchange @xmath90 weakly fm ( @xmath183 ) , while the stronger af superexchange along the @xmath67 axis ( @xmath184 ) favors quasi one - dimensional ( 1d ) spin fluctuations .
.3 cm ccc @xmath98 & @xmath91 & @xmath67 @xmath185 & @xmath186 & @xmath187 @xmath188 & @xmath189 & @xmath190 @xmath191 & @xmath192 & @xmath193 [ tab : eg ] by considering the superexchange model , one can derive as well the pure orbital interactions which stabilize the oo .
the superexchange interactions are anisotropic below the structural transition at @xmath85 . in contrast , at sufficiently high temperature @xmath194 , when also spin correlations may be neglected , one finds isotropic orbital interactions , @xmath195 which multiply @xmath196 for each bond , contributing to an orbital instability towards alternating @xmath197-type oo , while actually @xmath15-type oo is observed below @xmath85 .
it is thus clear from experiment that this instability can not be of purely electronic origin , and that , similarly to what is the case in lamno@xmath4 , @xcite it is supported by the lattice . in fact , although it has been argued that the oo is caused primarily by the superexchange , @xcite the electronic interactions ( [ kc9 ] ) predict that @xmath198 , and not @xmath199 , as observed . note also that as soon as the af spin correlations develop along the @xmath67 axis , one finds anisotropic orbital interactions , @xmath200 , which amplifies the ongoing symmetry breaking in the tetragonal phase .
the spectral weights of the optical subbands also follow from the superexchange processes , and are determined from the effective hamiltonian ( [ som9 ] ) by the general relations given by eqs .
( [ hefa ] ) and ( [ opsa ] ) . following the excitation spectrum of fig .
[ fig : levels ] , one finds optical absorption at three different energies ( the degeneracy of the @xmath121 state is not removed ) , so we label the respective kinetic energy contributions @xmath95 by @xmath201 .
they are determined at low temperature @xmath202 k by rigid @xmath15-type
oo ( [ ood9 ] ) , with the classical averages of the orbital operators given in table [ tab : eg ] .
so one finds for polarization along the @xmath67 axis @xmath203 and for polarization in the @xmath91 plane @xmath204 similar to the exchange constants @xmath90 and @xmath89 , the kinetic energies depend on the multiplet structure described by two parameters , viz .
@xmath1 ( [ j ] ) , @xmath2 ( [ eta ] ) , and on the oo ( [ ood9 ] ) specified by the angle @xmath167 .
note that they depend on these parameters also indirectly , since the spin - spin correlations are governed by @xmath90 and @xmath89 as well .
we analyze this dependence in sec .
[ d9:kcufopt ] . in order to apply the above classical theory to
kcuf@xmath4 we need to determine the microscopic parameters which decide about the superexchange constants , given by eqs .
( [ jc9 ] ) and ( [ jab9 ] ) . in principle
, if the optical data would also be available , with this experimental input one would be able to fix the values of the relevant parameters @xmath1 and @xmath2 , and the orbital angle @xmath167 .
having only magnetic measurements , we give here an example of another approach which starts from the microscopic parameters for the local coulomb interaction and hund s element suggested by the electronic structure calculations performed within the lda+u method:@xcite @xmath205 , @xmath206 ev they lead to @xmath207 . note that these parameters are somewhat smaller than the values @xmath208 and @xmath209 ev deduced for cu@xmath71 ions in the cuo@xmath84 planes of the high - temperature superconductors by grant and mcmahan using the fixed charge method,@xcite but we believe that they reflect better the partly screened interactions within cuf@xmath108 units .
we are not aware of any estimation of the remaining microscopic parameters until now , but taking into account the expected contraction of the @xmath103 wavefunctions by going from o@xmath210 to f@xmath211 ions , we argue that @xmath212 is reduced , while @xmath17 and @xmath213 could be similar to their respective values for cuo@xmath84 planes.@xcite therefore , it is reasonable to adopt : @xmath214 , @xmath215 , and @xmath216 ev .
note that although the values of @xmath214 , @xmath17 and @xmath213 could not be really estimated , in the present approach they are not independent parameters ; also only a linear combination of @xmath17 and @xmath213 enters eq .
( [ r ] ) , so a change in the value of @xmath213 could to some extent compensate a modified value of @xmath17 .
the present parameters lead to @xmath217 mev and @xmath218 .
): @xmath89 ( [ jc9 ] ) and @xmath90 ( [ jab9 ] ) for the copper fluoride model as functions of the orbital angle @xmath167 which describes the oo [ see eq .
( [ oo ] ) ] .
the dashed lines show the @xmath10 term alone , while the solid lines include the ct contributions as well .
the oo induced by the jt distortions ( @xmath219 ) is indicated by dotted line .
parameters : @xmath217 mev , @xmath207 , @xmath218 .
, width=302 ] consider now the oo of the occupied orbitals ( by holes ) in kcuf@xmath4 .
recent resonant x - ray scattering experiments suggest that both sublattices are equivalent , with @xmath220 in eq .
( [ oo]),@xcite but the precise shape of the occupied orbitals in kcuf@xmath4 remains unresolved .
there are different views concerning the type of orbitals that participate in the oo state . on the one hand
, it is believed that the orbital angle @xmath167 should be close to the angle @xmath221 ( @xmath222 ) , as given by @xmath223 , which follows from the local lattice distortions.@xcite on the other hand , the electronic interactions in the symmetry broken @xmath14-af phase below @xmath224 would favor instead alternating @xmath225 orbitals,@xcite with @xmath226 . in reality
, one expects rather a certain compromise between the electronic interactions for finite spin correlations and those induced by the lattice .
thus , in the present study of the magnetic exchange constants and optical spectral weights we shall consider a range of possible values of @xmath227 , focusing in particular on the above values favored by the above individual terms in the effective hamiltonian .
first , we demonstrate that the model eq .
( [ som9 ] ) is capable of describing the experimentally observed exchange constants @xmath228 mev and @xmath229 mev .
@xcite remarkably , the value of @xmath230 is smaller by more than one order of magnitude than @xmath231 , being some challenge for the theoretical model .
consider first the values of @xmath89 and @xmath90 for varying angle @xmath167 which tunes the oo ( fig .
[ fig : j9 ] ) .
when only the @xmath10 part of the superexchange @xmath117 is considered , one finds @xmath232 mev and @xmath233 mev .
the fm term @xmath181 gives the largest contribution for the @xmath91 planes , which follows from the alternation of hole orbitals in the @xmath91 planes , being close to the planar orbitals @xmath234 ( @xmath235 ) suggested early on by kugel and khomskii , @xcite but is partly compensated by the af terms .
while @xmath90 is only weakly depending on @xmath167 near this type of oo , @xmath89 decreases steadily with increasing @xmath167 ( fig .
[ fig : j9 ] ) , as the overlap between the orbitals occupied by holes along the @xmath67 axis decreases when the amplitude of the @xmath236 states is reduced .
one finds that using only the @xmath10 term in the superexchange , rather extreme parameters , such as @xmath237 ev and @xmath238 ev ( with the present value of @xmath28 ) , would have to be assumed to reproduce the experimental values of @xmath89 and @xmath90 . the ct term ( [ som9d ] ) with @xmath218 enhances the af interaction @xmath89 by a factor larger than two but hardly changes @xmath90 ( see fig . [
fig : j9 ] ) . only then @xmath89 comes close to the experimental value @xmath228 mev.@xcite as in the @xmath10 model ( at @xmath239 )
, the value of @xmath89 decreases with increasing @xmath167 , and there is no serious difficulty to fit the parameters in order to obtain a reasonable agreement with experiment , once the value of the orbital angle @xmath167 would be known . as an illustrative example
we show the results obtained with the present parameters in fig .
[ fig : j9 ] one finds @xmath240 mev for the oo which agrees with the lattice distortions ( @xmath221 ) , and @xmath241 mev for the kugel - khomskii @xmath234 orbitals ( @xmath226 ) .
these results demonstrate that the ct superexchange term plays an essential role in kcuf@xmath4 and so this system has to be classified as a _ charge transfer insulator_. remarkably , the value of @xmath90 is almost unaffected by the ct term ( fig .
[ fig : j9 ] ) .
this is due to the alternating oo in the @xmath91 planes , which makes the value of the orbital projection @xmath185 in eq .
( [ som9d ] ) very small indeed in the physical range of @xmath167 ( compare table [ tab : eg ] ) .
in fact , for the alternating planar @xmath234 orbitals one of the operators @xmath242 equals zero , and the ct contribution vanishes , so one can not reduce the value of @xmath243 by increasing the af ct term that follows from @xmath118 . in kcuf@xmath4 , given by eqs .
( [ jc9 ] ) and ( [ jab9 ] ) , for increasing charge transfer parameter @xmath24 , obtained for a few values of @xmath2 and for the oo induced by the jt effect ( @xmath219 , solid lines ) .
the dashed line shows @xmath244 obtained for @xmath207 and for alternating @xmath234 orbitals ( @xmath235 ) .
, width=291 ] the strong anisotropy of the magnetic exchange interactions in kcuf@xmath4 is well illustrated in fig .
[ fig : ratio ] by the ratio @xmath244 , being close to 0.07 for either the jt oo ( @xmath219 ) , or for the oo suggested by the orbital superexchange at @xmath177 ( @xmath235 ) . note that for @xmath245 the ratio @xmath244 does not depend significantly on @xmath167 in the interesting range between @xmath246 and @xmath235
. now we turn to the optical spectral weights ( [ hefa ] ) and determine the kinetic energies for the corresponding hubbard subbands . as discussed in sec .
[ sec : general ] , they originate from different multiplet excitations , and depend on the oo and on the spin - spin correlations ( [ spins ] ) . here we analyze in detail the spectral weight distribution for polarization along the @xmath67 axis , where strong exchange interaction @xmath89 controls the spin - spin correlations @xmath247 ( [ spins ] ) which remain finite in a broad temperature regime .
knowing that the interchain fm exchange coupling @xmath90 is so weak , we describe the temperature variation of the spin - spin correlations @xmath248 employing the jordan - wigner fermion representation@xcite for a 1d spin chain .
one finds for perfect
oo at temperature @xmath110 @xmath249 where @xmath250 here @xmath251 is the 1d dispersion of pseudofermions .
the exchange interaction @xmath89 is constant as long as the orbitals remain frozen , and sets the energy scale for the temperature variation of @xmath247 .
( [ kappad9 ] ) and ( [ epsid9 ] ) were solved self - consistently to obtain @xmath247 ( [ scd9 ] ) as a function of temperature . in the limit
@xmath252 one finds @xmath253 , and @xmath254 .
this value represents an excellent analytic approximation to the exact result , @xmath255 obtained for the 1d af heisenberg chain from the bethe ansatz .
@xcite ) , as obtained for the @xmath67 axis with the oo ( [ oo ] ) suggested by the jt distortions ( @xmath219 ) in kcuf@xmath4 ( in units of @xmath1 ) : high - spin @xmath256 ( solid line ) and low - spin @xmath257 ( dashed lines ) , for increasing temperature @xmath258 .
parameters as in fig .
[ fig : j9 ] .
, title="fig:",width=283 ] -2 mm .2 cm cccccccc & & & & @xmath259 & @xmath260 k & @xmath261 k & & @xmath259 & @xmath260 k & @xmath261 k @xmath256 & 3.8 & 3.9 & 6.1 & & 3.2 & 3.3 & 5.2 @xmath262 & 16.9 & 16.8 & 11.9 & & 19.0 & 18.9 & 13.4 @xmath263 & 8.9 & 8.8 & 6.2 & & 11.2 & 11.1 & 7.9 @xmath264 & 29.6 & 29.5 & 24.2 & & 33.4 & 33.3 & 26.5 @xmath265 & 21.3 & 16.0 & 16.0 & & 19.5 & 14.6 & 14.6 @xmath266 & 0.0 & 3.9 & 3.9 & & 0.0 & 3.6 & 3.6 @xmath267 & 0.0 & 0.1 & 0.1 & & 0.0 & 0.0 & 0.0 @xmath268 & 21.3 & 20.0 & 20.0 & & 19.5 & 18.2 & 18.2 [ tab : kcuf3 ] the general theory presented in sec .
[ sec : general ] makes a clear prediction concerning the temperature dependence of the spectral weights in optical absorption .
first of all , a large anisotropy between the polarization along the af @xmath67 axis and the polarization in the ( weakly fm ) @xmath91 plane is expected when the af ( fm ) spin - spin correlations along the @xmath67 axis ( within the @xmath91 planes ) develop . indeed , using the self - consistent solution of eqs .
( [ kappad9 ] ) and ( [ epsid9 ] ) , one finds that the kinetic energy @xmath256 ( which determines the spectral weight of the high - spin excitation at energy @xmath52 ) is rather low ( fig .
[ fig : sw9 ] ) .
in contrast , the low - spin excitations @xmath262 and @xmath263 contribute with large spectral weights in the low temperature regime , reflecting the af correlations along the @xmath67 axis . when the temperature increases , the spin - spin correlations @xmath247 gradually weaken and the kinetic energy redistributes both low - spin terms @xmath262 and @xmath263 decrease , while the high - spin term @xmath256 increases as more high - spin excitations are then allowed ( fig . [
fig : sw9 ] ) . for the considered oo given by @xmath269 , the changes of the contributions at the two lower energies which correspond to @xmath270 and @xmath271 are particularly large between zero and room temperature ( up to @xmath272 ) , with the increase ( decrease ) of @xmath256 [ @xmath262 ] by @xmath274 ( @xmath275 ) percent of the reference value at @xmath259 .
this leads to rather similar values of all three contributions at room temperature .
this predicted behavior could be verified by future experiments .
the temperature variation of the spectral weights for @xmath91 polarization is more difficult to predict as it involves weak @xmath276 spin correlations which develop in the temperature range @xmath277 , and grow with increasing order parameter , @xmath278 at @xmath252 , one finds that the kinetic energy would come entirely from the high - spin optical excitations @xmath265 , while the low - spin excitations would be fully suppressed ( table [ tab : kcuf3 ] ) .
above @xmath224 the spin system is controlled by the dominating af exchange constant @xmath89 , and @xmath279 .
even then the high - spin excitations at low energy dominate and have large spectral weight as a result of the persisting oo .
some decrease of @xmath265 accompanied by the increase of @xmath256 with increasing temperature makes the anisotropy between @xmath256 and @xmath265 considerably less pronounced , but this anisotropy of the spectra in the low energy range remains still close to 3:1 even at room temperature ( table [ tab : kcuf3 ] ) .
note also that the highest energy excitation for the @xmath91 polarization vanishes at @xmath226 , and gives a negligible contribution for the jt angle @xmath221 , due to the orbital correlations within the @xmath91 planes .
finally , we would like to emphasize again that knowing only the exchange constants @xmath90 and @xmath89 in kcuf@xmath4 , one is not able to determine all the microscopic parameters of the ct model .
we emphasize that a better understanding of the properties of kcuf@xmath4 can be achieved only by combining the results of magnetic and optical experiments , after the latter experiments have been performed .
although @xmath3 and @xmath0 electrons behave quite differently in lamno@xmath4 and are frequently treated as two subsystems , @xcite the neutron experiments@xcite which measure the spin waves in the @xmath14-af phase below @xmath224 leave no doubt that an adequate description of the magnetic properties requires a magnetic hamiltonian of the form given by eq .
( [ hs ] ) , describing superexchange between total @xmath280 spins of the mn@xmath69 ions
. the high - spin @xmath281 ground state at each mn@xmath69 ion is stabilized by the large hund s exchange @xmath13 .
the situation is here more complex than either in kcuf@xmath4 ( sec .
[ sec : d9 ] ) or in the @xmath0 systems discussed in the following sections , however , as the superexchange terms between mn@xmath69 ions originate from various charge excitations @xmath282 , made either by @xmath3 or by @xmath0 electrons , leading to different excited states in the intermediate @xmath72 configuration on a mn@xmath283 ion .
such processes determine the @xmath10 term @xmath284 defined by eq .
( [ hj ] ) , and were analyzed in detail in ref . , and
lead to the structure of @xmath284 given below . however , the ct processes , @xmath285 , contribute as well and the complete model for lamno@xmath4 reads @xmath286 the superexchange constant @xmath1 is here defined again by eq .
( [ j ] ) , using an average hopping element along an effective @xmath29 bond , @xmath287 , where @xmath288 is an average ct excitation energy , introduced below in eq . ( [ bard ] ) .
first we analyze the structure of the @xmath10 term for lamno@xmath4 , @xmath284 , due to excitations involving @xmath3 electrons .
the energies of the five possible excited states:@xcite ( i ) the high - spin @xmath289 state(@xmath290 ) , and ( ii - v ) the low - spin ( @xmath291 ) states : @xmath292 , @xmath293 ( @xmath294 , @xmath295 ) , and @xmath296 , will be parametrized again by the intraorbital coulomb element @xmath10 ( [ u ] ) , and by hund s exchange @xmath13 between a pair of @xmath3 electrons , defined in eq . ( [ jhe ] ) .
@xcite the energies of the excited states are given in terms of the racah parameters in ref . ;
in order to parametrize this spectrum by @xmath13 we apply an approximate relation @xmath297 which holds for the spectroscopic values of the racah parameters for a mn@xmath71 ( @xmath72 ) ion:@xcite @xmath298 ev and @xmath299 ev . here
we use these atomic values as an example of the theory using them and eq .
( [ jhe ] ) one finds the excitation spectrum : @xmath52 , @xmath300 , @xmath301 , @xmath301 , and @xmath302 [ fig .
[ fig : levels](a ) ] . unlike @xmath13
, the value of @xmath10 is known with less accuracy hence we shall use it here only as a parameter which can be deduced _ a posteriori _ from the superexchange @xmath1 which is able to explain the experimental values for two exchange constants responsible for the @xmath14-af phase observed in lamno@xmath4 well below the structural transition ( here again @xmath303 ) . using the spin algebra ( clebsch - gordon coefficients ) , and making a rotation of the terms derived for a bond @xmath304 to the other two cubic axes @xmath92 and @xmath93 , one finds five contributions to @xmath284 due to different @xmath305 excitations by @xmath3 electrons,@xcite @xmath306 where the coefficients @xmath307 follow from the above multiplet structure of mn@xmath71 ( @xmath72 ) ions , and @xmath2 ( [ eta ] ) stands for the hund s exchange .
the meaning of the various terms is straightforward : the first term @xmath308 describes the high - spin excitations to the @xmath289 state while the remaining ones , @xmath153 with @xmath309 , arise due to the low - spin excited states @xmath292 , @xmath295 , @xmath294 and @xmath296 , respectively .
the orbital dependence is given by the same operators ( [ orbpro ] ) as in sec .
[ sec : d9 ] .
similar to the case of the @xmath121 state for the copper fluoride [ see eqs .
( [ som92 ] ) and ( [ som93 ] ) ] , the doubly degenerate @xmath293 state contributes here with two terms characterized by a different orbital dependence in eqs .
( [ som43 ] ) and ( [ som44 ] ) .
note that this degeneracy would be removed by the cooperative jt effect , i.e. , the structural phase transition ( and associated oo ) driven by the local jt coupling in combination with the elastic lattice forces .
the resulting small level splitting we neglect here , and so we set @xmath310 .
the superexchange mediated by @xmath0 electrons results from @xmath311 excitations which involve @xmath312 and @xmath313 configurations at both mn ions : mn@xmath71 and mn@xmath314 .
they give low @xmath291 spins of mn@xmath71 ions , and this part of the superexchange is af . using the present units introduced in eqs .
( [ u ] ) and ( [ jhe ] ) , one finds the excitation energies ( not shown in fig .
[ fig : levels ] ) : @xmath315 , @xmath316 , @xmath317 , and @xmath318 , with the first ( second ) label standing for the configuration of the mn@xmath71 ( mn@xmath314 ) ion , respectively . in the actual derivation
each of the excited states , with one @xmath0 orbital being either doubly occupied or empty , has to be projected on the respective eigenstates and the spin algebra is next used to construct the interacting total @xmath280 spin states .
this leads to the final contribution to @xmath117 which , in a good approximation , is orbital independent,@xcite @xmath319 here @xmath320 follows from the difference between the effective @xmath19 hopping elements along the @xmath64 and @xmath31 bonds , and we adopt the slater - koster value @xmath321 .
the coefficient @xmath322 stands for a superposition of the above @xmath0 excitations involved in the @xmath0 superexchange , @xmath323 there is no need to distinguish between the different excitation energies ; all of them are significantly higher than the first low - spin excitation energy for the configuration @xmath292 , which occurs after an excitation by an @xmath3 electron .
while earlier studies of the superexchange interactions in manganites were limited to model hamiltonians containing only the @xmath10 term , @xcite the importance of the ct processes was emphasized only recently.@xcite for our purposes we derived the ct term @xmath285 by considering again excitations by either @xmath64 or @xmath31 electrons on the bond @xmath58 , leading to two - hole excited states at an intermediate oxygen , @xmath324 . unlike in kcuf@xmath4 with a unique ct excitation , however , in the present case a number of different excited states occurs with the excitation energies depending on the electronic configuration of the two intermediate mn@xmath71 ions at sites @xmath56 and @xmath57 .
one finds that these various excitations can be parametrized by a single parameter @xmath24 given by eq .
( [ r ] ) , and the excited states on two neighboring transition metal ions contribute , as for the @xmath10 term , both fm and af terms , @xmath325 \big(\frac{1}{4}-\tau_i^{(\gamma)}\tau_j^{(\gamma)}\big ) \nonumber \\ & + & \frac{8}{5}\big[-c_4\big(\vec{s}_i\cdot\vec{s}_j+6\big ) + c_5\big(\vec{s}_i\cdot\vec{s}_j-4\big)\big ] \big(\frac{1}{2}+\tau_i^{(\gamma)}\big ) \big(\frac{1}{2}+\tau_j^{(\gamma)}\big)\big\ } \nonumber \\ & + & \frac{1}{8}jr\beta c_t\sum_{\langle ij\rangle } \big(\vec{s}_i\cdot\vec{s}_j-4\big),\end{aligned}\ ] ] where the coefficients @xmath326 , with @xmath327 , and @xmath328 are all determined by @xmath2 and @xmath24 via @xmath329 : @xmath330 the coefficients @xmath326 follow from ct excitations by @xmath3 electrons . as in the copper fluoride case ( sec .
[ sec : d9 ] ) , the lowest ( high - spin ) excitation energy will be labeled by @xmath17 , @xmath331 so the other possible individual ( low - spin ) excitations at each transition metal ion have the energies @xmath332 these excitation energies are used here to introduce an average ct energy @xmath288 , @xmath333 which serves to define the effective hopping element @xmath287 , and the superexchange energy @xmath1 ( [ j ] ) in a microscopic approach .
we emphasize , however , that such microscopic parameters as @xmath334 will not be needed here , and only the values of the effective parameters @xmath335 will decide about the exchange constants and the optical spectral weights .
each coefficient @xmath326 ( [ cn ] ) , with @xmath327 , stands for an individual process which contributes with a particular orbital dependence due to an intermediate state arising in the excitation process , and accompanies either a fm or an af spin factor , depending on whether high - spin or low - spin states are involved . as in the @xmath10 term ( [ som41])([som46 ] ) , a pair of directional @xmath146 orbitals accompanies low - spin excitations , while either high - spin or low - spin excited states are allowed when two different orbitals , one directional and one orthogonal to it ( planar orbital ) , are occupied at the two mn@xmath69 sites .
in contrast to the @xmath10 term , also configurations with two planar orbitals @xmath336 occupied at sites @xmath56 and @xmath57 contribute to @xmath285 in eq .
( [ hct4e ] ) .
these terms are accompanied by the projection operator @xmath337 in eq .
( [ hct4e ] ) .
note that in the case of the @xmath10 term such configurations did not contribute as the @xmath3 electrons were then blocked and could not generate any superexchange terms .
as the electrons from an oxygen @xmath157 orbital are excited instead to directional @xmath146 orbitals at two neighboring mn@xmath69 ions , again both high - spin ( @xmath290 ) and low - spin ( @xmath291 ) excitations are here possible , giving a still richer structure of @xmath285 .
the oo in lamno@xmath4 is robust and sets in below the structural transition at @xmath338 k.@xcite the orbital interactions present in the superexchange hamiltonian ( [ som4 ] ) would induce a @xmath197-type oo.@xcite the observed classical ground state , which can again be described by eq .
( [ ood9 ] ) , corresponds instead to @xmath15-type oo , as deduced from the lattice distortions .
note that in contrast to kcuf@xmath4 , the occupied orbitals refer now to electrons and thus the values of the expected orbital angle @xmath167 are @xmath339 ( which corresponds to @xmath340 ) and so are distinctly different from the copper fluoride case . the averages of the orbital operators in the orbital ordered state are given in table [ tab : eg ] , including the terms @xmath341 which contribute now to the ct part of superexchange .
the dependence on the orbital angle @xmath167 suggests that , similar to kcuf@xmath4 , these new terms are more significant along the @xmath67 axis for the oo expected in the manganites . for a better understanding of the effective exchange interactions it is convenient to introduce first the @xmath0 superexchange constant @xmath342 which stands for the interaction induced by the charge excitations of @xmath0 electrons . when the ct terms are included , @xmath342 consists of the two contributions given in eqs .
( [ som46 ] ) and ( [ hct4e ] ) , @xmath343 this interaction is frequently called the superexchange between the core spins .
we emphasize that this term is orbital independent and thus isotropic .
the coupling constant @xmath342 has a similar origin as the @xmath3 part of the superexchange @xmath344 , which however is orbital dependent and anisotropic .
we emphasize that both @xmath342 and @xmath344 are relatively small fractions of @xmath1 .
the @xmath3 contributions to the effective exchange constants ( [ hs ] ) in lamno@xmath4 depend on the orbital state characterized again by eqs .
( [ ood9 ] ) via the averages of the orbital operators , @xmath345 as the structural transition occurs at relatively high temperature @xmath346 k , at room temperature ( and below it ) the oo may be considered to be frozen and specified by an angle @xmath167 [ see eqs .
( [ ood9 ] ) ] .
the orbital fluctuations are then quenched by the combined effect of the orbital part of the @xmath3 superexchange in eqs .
( [ som41])-([som45 ] ) and the orbital interactions induced by the jt effect,@xcite and the spins effectively decouple from the orbitals , leading to the effective spin model ( [ hs ] ) .
the magnetic transition then takes place within this oo state , and is driven by the magnetic part of the superexchange interactions , which follow from @xmath284 and @xmath285 . for the @xmath15-type oo , as observed in lamno@xmath4,@xcite one finds the effective exchange constants in eq .
( [ hs ] ) as a superposition of @xmath342 and @xmath344 after inserting the averages of the orbital operators ( see table [ tab : eg ] ) in eq .
( [ je4 ] ) : @xmath347 considering @xmath321 to be fixed by the slater - koster parametrization , a complete set of parameters which determines @xmath89 and @xmath90 comprises : @xmath1 ( [ j ] ) , @xmath2 ( [ eta ] ) , @xmath24 ( [ r ] ) , and the orbital angle @xmath167 which defines the phase with oo by eqs .
( [ ood9 ] ) , referring now to the orbitals occupied by electrons .
equations ( [ jab4 ] ) and ( [ jc4 ] ) may be further analyzed in two ways : either ( i ) using an effective model which includes only the @xmath10 superexchange term due to @xmath19 transitions , as presented in ref . and discussed in appendix [ d4:eff ] ( i.e. , taking @xmath239 which implies @xmath348 ) , or ( ii ) by considering the complete @xmath349 model as given by eq .
( [ som4 ] ) , which includes also the ct contributions to the superexchange ( with @xmath350 ) . by a similar analysis in sec .
[ sec : d9 ] we have established that the ct terms are of essential importance in kcuf@xmath4 and should not be neglected , as otherwise the strong anisotropy of the exchange constants would remain unexplained .
here the situation is qualitatively different as we show in appendix [ d4:eff ] , using somewhat modified parameters @xmath1 and @xmath2 one may still reproduce the experimental values of the exchange constants , deduced from neutron experiments for lamno@xmath4,@xcite within the effective model at @xmath239 , and even interpret successfully the optical spectra.@xcite it is important to realize that the high - spin @xmath3 electron excitations play a crucial role in stabilizing the observed @xmath14-af spin order , as only these processes are able to generate fm terms in the superexchange .
they compete with the remaining af terms , and the @xmath14-af phase is stable only when @xmath351 and @xmath352 .
we have verified that the terms which contribute to @xmath90 and @xmath89 in eqs .
( [ jab4 ] ) and ( [ jc4 ] ) are all of the same order of magnitude as all the coefficients @xmath353 are of order one .
hence , the superexchange energy @xmath1 ( [ j ] ) is much higher than the actual exchange constants in lamno@xmath4 , i.e. , @xmath354 and @xmath355 .
next we consider the kinetic energies associated with the various optical excitations which determine the optical spectral weights by eq .
( [ hefa ] ) . again , as in the previously considered case of kcuf@xmath4 , the high - spin subband at low energy is unique and is accompanied by low - spin subbands at higher energies . while the energetic separation between the high - spin and low - spin parts of the spectrum @xmath356 is large , one may expect that the low - spin optical excitations might be difficult to distinguish experimentally from each other . as we will see below by analyzing the actual parameters of lamno@xmath4 , the low - spin excitations overlap with the @xmath16 ct excitations , and so such a separation is indeed impossible
thus we analyze here only the global high - energy spectral weight due to the optical excitations on the transition metal ions , expressed by the total kinetic energy @xmath357 for all ( @xmath3 and @xmath0 ) low - spin terms , and compare it with the high - spin one , given by @xmath358 . using the manganite model ( [ som4 ] )
one finds for polarization in the @xmath91 plane @xmath359(4-s_{ab}),\end{aligned}\ ] ] and for polarization along the @xmath67 axis @xmath360(4-s_c).\end{aligned}\ ] ] the optical spectral weights given by eqs .
( [ wab41])([wc42 ] ) are determined by @xmath1 , @xmath2 , and the orbital angle @xmath167 .
note that the leading term in the low - spin part comes from the @xmath3 optical excitations , while the @xmath0 excitations contribute only with a relatively small weight @xmath361 .
it is a challenge for the present theoretical model to describe both the magnetic exchange constants@xcite and the anisotropic optical spectral weights@xcite using only a few effective parameters @xmath335 and the orbital angle @xmath167 .
we shall proceed now somewhat differently than in sec .
[ sec : d9 ] , and analyze the experimental data using primarily these effective parameters , while we will discuss afterwards how they relate to the expectations based on the values of the microscopic parameters found in the literature .
the experimental values of the exchange constants,@xcite @xmath362 mev and @xmath363 mev , impose rather severe constraints on the microscopic parameters and on the possible oo in lamno@xmath4 .
the af interaction @xmath89 is quite sensitive to the type of occupied orbitals ( [ ood9 ] ) , and increases with increasing amplitude of @xmath236 orbitals in the ground state , i.e. , with decreasing orbital angle @xmath167 .
simultaneously , the fm interaction @xmath364 is enhanced . already with the effective model ( at @xmath239 )
it is not straightforward to determine the parameters @xmath1 and @xmath2 , as we discuss in appendix [ d4:eff ] .
this model is in fact quite successful , and a reasonable agreement with experiment could be obtained both for the magnetic exchange constants and for the optical spectral weights , taking the experimental excitation spectrum.@xcite here we will investigate to what extent this effective model gives robust results and whether the ct processes could play an important role in lamno@xmath4 . by analyzing the ct terms in the superexchange
[ compare eqs .
( [ jab4 ] ) , ( [ jc4 ] ) ] one concludes that these contributions are predominantly af .
therefore , it might be expected that a higher value of @xmath2 than 0.16 used in appendix [ d4:eff ] would rather be consistent with the magnetic experiments .
increasing @xmath2 gives an increased coefficient @xmath365 , so not only the fm term in @xmath366 is then enhanced , but also the optical spectral weight @xmath367 which corresponds to the high - spin transition .
simultaneously , the angle @xmath167 is somewhat increased , but the dependence of the spectral weight @xmath367 on the angle @xmath167 is so weak that the higher value of @xmath2 dominates and a somewhat lower value of @xmath1 than 170 mev used in appendix [ d4:eff ] has to be chosen .
altogether , these considerations led us to selecting @xmath368 mev and @xmath369 as representative parameters for which we show below that a consistent explanation of both magnetic and optical data is possible . : ( a ) @xmath89 along the @xmath67 axis , and ( b ) @xmath90 in the @xmath91 plane , as functions of orbital angle @xmath167 which defines the @xmath15-type oo , see eqs .
( [ ood9 ] ) . dashed lines for increasing @xmath239 , 0.15 and 0.30 from bottom to top ; solid lines for @xmath370 .
note that the dependence of @xmath90 on @xmath24 is nonmonotonic .
horizontal dotted lines indicate the experimental values of ref . for lamno@xmath4 ; diamond and circle show the experimental values of @xmath89 and @xmath90 for @xmath371 .
parameters : @xmath368 mev , @xmath369 .
, width=283 ] after these two parameter values have been fixed , it is of interest to investigate the dependence of the effective exchange interactions on the ct parameter @xmath24 and on the orbital angle @xmath167 . as in the kcuf@xmath4 case , one finds a stronger increase of @xmath89 with increasing @xmath24 , while these terms are weaker and lead to nonmonotonic changes for @xmath90 ( fig .
[ fig : j4r ] ) .
first of all , with the present value of @xmath369 , at @xmath239 the af interaction @xmath89 is close to zero for angles @xmath372 , while the fm interaction @xmath373 mev is somewhat too strong . with increasing
@xmath24 one finds that @xmath89 increases , while the fm interaction @xmath90 initially becomes weaker when @xmath24 increases up to @xmath374 and the term @xmath375 dominates the ct contribution to @xmath90 [ see eq .
( [ jab4 ] ) ] . at higher values of @xmath24
, however , the fm contributions due to @xmath376 and @xmath377 start to dominate , and @xmath90 decreases with increasing @xmath24 , particularly for small values of @xmath378 .
one finds that the experimental values of both exchange constants are well reproduced for @xmath370 and at the orbital angle @xmath371 .
although this fit is not unique , one has to realize that the experimental constraints imposed on the parameters are indeed rather severe as we show below the present parameters give a very reasonable and consistent interpretation of the experimental results for lamno@xmath4 . for the above parameters the fm and af terms to @xmath89 almost compensate each other in the @xmath10 term , and a considerable af interaction along the @xmath67 axis arises mainly due to the ct contributions ( fig .
[ fig : j4 ] ) .
this qualitatively agrees with the situation found in kcuf@xmath4 , where the ct term was of crucial importance and helped to explain the observed large anisotropy between the values of exchange constants .
also the ct term which contributes to @xmath90 is af and increases with increasing angle @xmath167 , while the @xmath10 term is fm but weakens at the same time .
this results in a quite fast dependence of the fm interaction @xmath90 on @xmath167 ( fig .
[ fig : j4 ] ) .
( [ jab4 ] ) and @xmath89 ( [ jc4 ] ) as functions of the orbital angle @xmath167 , obtained within the superexchange model ( [ som4 ] ) for lamno@xmath4 ( solid lines ) .
contributions due to the @xmath10 term and due to the ct term to both exchange constants ( @xmath364 and @xmath89 ) are shown by dashed and dashed - dotted lines , respectively .
isotropic and orbital independent @xmath0 terms superexchange terms @xmath379 and @xmath380 ( [ jt4 ] ) are : 1.45 and 0.83 mev .
experimental values @xcite of @xmath364 and @xmath89 ( indicated by diamond and circle ) are well reproduced at angle @xmath371 .
parameters @xmath368 mev , @xmath369 , @xmath370 .
, width=302 ] one thus recognizes a similar dependence of the exchange constants @xmath364 and @xmath89 on the orbital angle @xmath167 ( fig .
[ fig : j4 ] ) as that found before with the @xmath10 term alone ( see fig .
[ fig : j4u ] in appendix [ d4:eff ] ) .
the ct terms have mainly two consequences : ( i ) a large af contribution to the interaction along the @xmath67 axis @xmath89 , and ( ii ) an increase of the orbital angle well above @xmath381 .
these trends are robust in the realistic parameter range .
therefore , one expects that the occupied @xmath3 orbitals approach the frequently assumed alternating directional orbitals in the @xmath91 planes @xmath382 with @xmath383 , but can not quite reach them .
indeed , we have verified that orbitals with @xmath384 are excluded by the present calculations , as then the fm interaction within the @xmath91 planes changes sign and becomes weakly af .
thus , the mechanism for the observed @xmath14-af phase is lost , and one has to conclude that angles @xmath385 are excluded . indeed , we have found that the orbital angle @xmath386 reproduces well the experimental data for both exchange constants ( fig .
[ fig : j4 ] ) , and is thus somewhat smaller than the angle @xmath387 deduced from the lattice distortions in lamno@xmath4.@xcite this can be seen as a compromise between the orbital interactions involving the lattice and the purely electronic superexchange orbital interactions , so it is reasonable to expect that @xmath388 . finally , we emphasize that the @xmath3 electron excitations , contributing both to the @xmath10 and to the ct processes in the superexchange , are fm _ for all cubic directions_,@xcite and only due to a substantial @xmath0 term which follows from low - spin ( af ) excitations , @xmath389 mev , the exchange interaction along the @xmath67 axis changes its sign and becomes af .
alltogether , the present analysis shows that the @xmath0 superexchange plays a decisive role in stabilizing the observed @xmath14-af spin order without it already the undoped lamno@xmath4 would be a ferromagnet .
this peculiar situation follows from the large splitting between the high - spin and low - spin excitations @xmath390 ev in lamno@xmath4 , which is larger than in any other transition metal compound considered in this paper , due to the fact that the @xmath8 shell is half - filled in the mn@xmath71 excited states.@xcite this leads to relatively large fm contributions , even when the orbitals are not so close to the ideal alternation of directional and planar states ( as found along the @xmath67 axis ) , which would maximize the averages of the orbital operators , @xmath188 , that control the weight of the high - spin terms , see table [ tab : eg ] .
this result is remarkable but again qualitatively the same as found in the effective model of appendix [ d4:eff ] .
however , quantitatively the @xmath0 term is here somewhat stronger , as @xmath342 is now increased by @xmath391% over its value @xmath392 mev deduced from the effective @xmath19 model with @xmath10 terms only . as a common feature one finds that @xmath393 mev , so we emphasize that the superexchange promoted by @xmath0 electrons is quite weak and is characterized by a small value of @xmath394 ( or @xmath395 for the superexchange between @xmath291 core spins @xcite ) which provides an important constraint for the realistic models of manganites .
consider now the temperature dependence of the optical spectral weights ( [ opsa ] ) . as the orbital dynamics is quenched up to room temperature @xmath396 k
, it suffices to consider the temperature dependence of the intersite spin - spin correlations ( [ spins ] ) .
we derived these by employing the so - called oguchi method@xcite ( see appendix [ app : spincomn ] ) , which is expected to give rather realistic values of spin correlations functions for the large spins @xmath280 in lamno@xmath4 in the entire temperature range .
thereby we solved exactly the spin - spin correlations @xmath397 on a single bond @xmath58 coupled to neighboring spins by mean - field ( mf ) terms , proportional to the order parameter @xmath398 .
a realistic estimate of the magnetic transition can be obtained by reducing the mf result obtained for @xmath280 spins by an empirical factor @xmath399.@xcite using the exchange interactions obtained with the present parameters , one finds @xmath400 k which reasonably agrees with the experimental value of @xmath401 k.@xcite the calculations of spin - spin correlations are straightforward , and we summarize them in appendix [ app : spincomn ] . both correlation functions @xmath276 and @xmath247 change fast close to @xmath224 , reflecting the temperature dependence of the brillouin function which deterimes @xmath398 , and remain finite at @xmath402 .
the large splitting between the high - spin and low - spin excitations makes it possible to separate the high - spin excitations from the remaining ones in the optical spectra , and to observe the temperature dependence of its intensity for both polarizations . as in the effective model,@xcite
the present theory predicts that the low - energy optical intensity exhibits a rather strong anisotropy between the @xmath91 and @xmath67 directions .
it is particularly pronounced and close to 10:1 at low temperatures when the spin correlations @xmath403 and @xmath247 are maximal ( fig .
[ fig : anit ] ) .
unfortunately , this anisotropy at @xmath252 in only weakly dependent on the orbital angle @xmath167 , so it can not help to establish the type of oo realized in the ground state of lamno@xmath4 , and its possible changes with increasing temperature .
in fact , when the parameters are fixed and only the orbital angle @xmath167 is changed , a different value of the nel temperature follows from the modified exchange constants , being the main reason behind the somewhat different temperature dependence of the ratio @xmath404 ( fig .
[ fig : anit ] ) . , given by the ratio @xmath404 , for increasing temperature @xmath405 , as obtained for : @xmath371 ( solid line ) , @xmath406 ( dashed line ) , @xmath407 ( dashed - dotted line ) .
the arrows indicate the nel temperature @xmath224 derived from the exchange constants in each case ( see text ) .
parameters : @xmath368 mev , @xmath369 . ,
width=283 ] one finds a very satisfactory agreement between the present theory and the experimental results of ref .
, as shown in fig .
[ fig : anis ] .
we emphasize , however , that the temperature variation of the optical spectral weights could also be reproduced within the effective model at @xmath239,@xcite showing that the ct terms lead only to minor quantitative modifications .
note also that at this stage no fit is made , i.e. , the kinetic energies ( [ hefa ] ) which stand for the optical sum rule are calculated using the same parameters as found above to reproduce the exchange constants in fig . [
fig : j4 ] .
therefore , such a good agreement with experiment demonstrates that indeed the superexchange interactions describe the spectral intensities in the optical transitions .
we note , however , that the anisotropy in the range of @xmath408 is somewhat larger in experiment which might be due to either some inaccuracy of the oguchi method or due to the experimental resolution .
, obtained for lamno@xmath4 for the high - spin ( solid lines , @xmath270 ) and for the low - spin ( dashed lines , @xmath409 ) excitations at the mn ions for polarization : ( a ) in the @xmath91 planes , @xmath410 , and ( b ) along the @xmath67 axis , @xmath411 . the total kinetic energies @xmath412
are given by the long dashed - dotted lines , while the vertical dotted lines indicate the value of @xmath224 derived within the present model ( see text ) .
filled circles show the experimental intensities @xcite at low energy ( @xmath270 ) .
parameters : @xmath368 mev , @xmath413 , @xmath370 , and @xmath386 . , width=283 ] the distribution of the optical intensities @xmath95 and their changes between the low ( @xmath414 ) and the high temperature ( @xmath415 k ) regime are summarized in table [ tab : par ] . at temperature @xmath259
one finds that for the @xmath91 polarization the entire spectral intensity originates from high - spin excitations .
this result follows from the classical value of the spin - spin correlation function @xmath416 predicted by the oguchi method . as quantum fluctuations in the @xmath14-af phase
are small,@xcite the present result is nearly accurate .
when the temperature increases , one finds considerable transfer of optical spectral weights between low and high energies , discussed also in ref . , with almost constant total intensities @xmath412 for both @xmath417 and @xmath67 ( see also fig . [
fig : anis ] ) .
the optical weights obtained for the jt angle @xmath387 are similar to those for @xmath371 , showing again that the optical spectroscopy is almost insensitive to small changes of the orbital angle .
in contrast , for @xmath418 the exchange constants ( and hence also the estimated value of @xmath224 ) are too low .
.2 cm ccccccc & & & & & @xmath405 ( k ) & @xmath47 & @xmath261 & @xmath47 & @xmath261 & & @xmath90 & & & & @xmath89 & & & & @xmath265 & 140.1 & 94.2 & 135.8 & 88.1 & & @xmath419 & 0.0 & 38.1 & 0.0 & 42.1 & & @xmath420 & 140.1 & 132.3 & 135.8 & 130.2 & & @xmath256 & 13.7 & 42.9 & 12.6 & 41.2 & & @xmath421 & 80.0 & 44.0 & 74.9 & 40.1 & & @xmath422 & 93.7 & 86.9 & 87.5 & 81.3 & [ tab : par ] while the effective parameters @xmath335 used in this section give a very satisfactory description of both magnetic and optical properties of lamno@xmath4 , the values of the microscopic parameters , such as the coulomb interaction on the mn ions @xmath10 , and on the oxygen ions @xmath213 , the hund s exchange @xmath13 , the charge transfer parameter @xmath17 , and the @xmath16 hopping element @xmath212 , are not uniquely determined .
one could attempt to fix these parameters using the atomic value of hund s exchange @xmath206 ev .
with @xmath369 it leads to @xmath423 ev , quite close to other estimates,@xcite while the value of @xmath370 suggests then that , taking the commonly accepted@xcite value of the ct energy @xmath424 ev , the oxygen coulomb element is large , @xmath425 ev .
this value of @xmath213 is larger than usually obtained @xmath426 ev for oxygen ions , as for instance by analyzing the parameters of the three - band model for cuo@xmath84 planes.@xcite we note , however , that the optical data@xcite suggest a somewhat reduced value of hund s exchange @xmath427 ev ( using the present units ) , so it could be that the local exchange interactions are somewhat screened in reality .
fortunately , when the optical data are available , also the position of the low - energy excitation is known , and this may serve as an additional experimental constraint for the parameters.@xcite this excitation is found at about 2.0 ev,@xcite indicating that @xmath428 ev . with this constraint one finds , using again @xmath369 , that @xmath429 ev and @xmath430 ev .
these parameters give the low - spin @xmath292 and @xmath293 excitations close to 4.5 ev , in agreement with experiment.@xcite so the above values of the microscopic parameters @xmath431 appear to be consistent both with the present value of @xmath2 and with the spectra .
furthermore , for these empirical parameters one finds @xmath370 either with @xmath424 and @xmath432 ev , or with @xmath433 and @xmath434 ev .
these values , particularly the second set , are perfectly acceptable and in the usually considereed range.@xcite taking the above value of @xmath429 and @xmath368 mev , one finds @xmath435 ev , a somewhat lower value than that which follows from the effective model of appendix [ d4:eff ] .
altogether , these results indicate , contrary to what is frequently assumed , @xcite that the local exchange interactions are somewhat screened in reality by covalency effects , and that at the same time the screening of the intraorbital coulomb interaction @xmath10 is stronger than estimated before.@xcite
perovskite titanates , latio@xmath4 and ytio@xmath4 , are intriguing examples of mott insulators with orbital degrees of freedom due to @xmath0 electrons : in the ground state the ti@xmath69 ions are in the @xmath436 configuration . in an ideal perovskite structure
the @xmath0 orbitals are degenerate , but lattice distortions may contribute to the magnetic ground state@xcite and to the mott transition @xcite here we do not intend to discuss this controversial issue . in an ideal cubic system each @xmath0 orbital is perpendicular to a single cubic axis , for instance the @xmath437 orbital lies in the @xmath91 plane and is perpendicular to the @xmath67 axis .
it is therefore convenient to introduce the following short hand notation for the orbital degree of freedom:@xcite @xmath438 the labels @xmath59 thus refer to the cubic axes perpendicular to the planes of the respective orbitals . the superexchange spin - orbital model ( [ hngamma ] ) in cubic titanates couples @xmath132 spins and the orbital @xmath0 degrees of freedom at nearest neighbor ti@xmath69 ions . due to large @xmath10
the electron densities satisfy thereby the local constraint at each site @xmath56 , @xmath439 in titanates there is no need to consider ct processes , as these systems are mott - hubbard insulators@xcite and no qualitatively new effects apart from some negligible renormalization of the effective parameters @xmath440 could arise from ct excitations .
this simplifies our considerations , so we analyze the superexchange in the leading order of perturbation theory , given by the contributions which result from virtual excitations between the neighboring ti@xmath69 ions , @xmath441 .
these charge excitations involve the coulomb interactions in the @xmath73 configuration of a ti@xmath71 ion , parametrized as before by the intraorbital coulomb element @xmath10 , and by hund s exchange element @xmath13 for a pair of @xmath0 electrons , defined as follows ( see table [ tab : uij]),@xcite @xmath442 the charge excitations lead to one of four different excited states @xcite shown in fig .
[ fig : levels](b ) : the high - spin @xmath443 state at energy @xmath52 , and three low - spin states degenerate @xmath444 and @xmath121 states at energy @xmath125 , and an @xmath124 state at energy @xmath445 .
as before , the excitation energies are parameterized by @xmath2 , defined as in eq .
( [ eta ] ) , and we introduce the coefficients @xmath446 one finds the following compact expressions for the superexchange @xmath447 as given by eq .
( [ hj]):@xcite @xmath448 where @xmath449 as in secs .
iii and iv the orbital ( pseudospin ) operators @xmath450 depend on the direction of the @xmath58 bond .
their form follows from the simple observation@xcite that only two @xmath0 orbitals ( flavors ) are active along each cubic axis , e.g. if @xmath451 , the active orbitals are @xmath92 and @xmath93 [ see eq .
( [ t2 g ] ) ] , and they give two components of the pseudospin @xmath80 operator @xmath452 . here the operators @xmath453 describe the interactions between the active orbitals along the particular bond , which include the quantum effects due to their fluctuations , and take either the form of a scalar product @xmath454 , or lead to a similar expression , @xmath455 which involves double excitations due to @xmath456 and @xmath457 terms ( as in the @xmath458 hubbard model ) .
the interactions along axis @xmath98 are tuned by the number of electrons occupying active orbitals , @xmath459 , which is fixed by the number of electrons in the inactive orbital @xmath460 , because of the constraint ( [ cond1 ] ) .
the exchange constant for a bond @xmath58 along axis @xmath98 is obtained from eqs .
( [ som11])([som13 ] ) by averaging over the orbital states at both sites @xmath56 and @xmath57 , @xmath461 the cubic titanates are known to have particularly pronounced quantum spin - orbital fluctuations , and their proper treatment requires a rather sophisticated approach.@xcite here we shall ignore this complex quantum problem , and shall illustrate the general theory by extracting the magnetic exchange constants from eq .
( [ jd1 ] ) , and the optical spectral weights ( [ opsa ] ) , using an ansatz for the oo in the ground state , in analogy to the approach employed in secs .
[ sec : d9 ] and [ sec : d4 ] for the more classical @xmath3 systems .
in general a classical orbital state in the titanates with gdfeo@xmath4-type lattice structure can be parametrized as follows , @xmath462 ) wavefunctions at each site .
the occupied orbitals @xmath463 refer to four sublattices ( @xmath464 ) , with ti@xmath465 , ti@xmath466 ,
ti@xmath467 , ti@xmath468 positions.@xcite the minus signs in @xmath469 and @xmath470 reflect a mirror symmetry present in the gdfeo@xmath4 structure . note that this state resembles @xmath197-type oo , and is thus different from the @xmath15-type oo encountered for @xmath3 orbitals [ due to the change of sign for the @xmath471 orbitals in eqs .
( [ and1 ] ) along the @xmath67 axis ] . using the ansatz ( [ and1 ] )
one finds after a straightforward calculation the exchange constants , @xmath472 , \\
\label{jd1ab } j_{ab}&=&\frac{1}{4}j\big[2(r_1+r_2r_3)(\gamma^2+\alpha\beta)^2 \nonumber \\ & & \hskip .5 cm -(r_1-r_2)(1+\gamma^2)\big].\end{aligned}\ ] ] they are determined once again by : ( i ) the superexchange parameter @xmath1 given by eq .
( [ j ] ) , with @xmath28 standing now for the effective @xmath30 hopping element , ( ii ) hund s exchange element @xmath2
( [ eta]),@xcite and ( iii ) the orbital state ( [ and1 ] ) , specified in the present case by the coefficients @xmath473 . following the general theory of sec . [ sec : general ] , the optical excitations corresponding to the high - spin hubbard band at energy @xmath52 @xmath474 , and to the two low - spin hubbard bands at @xmath125 @xmath475 and @xmath445 @xmath476 of fig . [ fig : levels](b ) , have total intensities given by the respective kinetic energies @xmath95 , see eq .
( [ opsa ] ) . using the classical wavefunctions given by eqs .
( [ and1 ] ) , one finds the following optical spectral weights : for polarization along the @xmath67 axis , @xmath477 for polarization in the @xmath91 plane , @xmath478\big(\frac{3}{4}+\!s_{ab}\big ) , \\
\label{wab12 } \hskip -.5 cm k_2^{(ab)}\!\!&=&\!\!jr_2\big[\frac{1}{2}(1\!+\!\gamma^2 ) + \frac{1}{3}(\gamma^2\!+\!\alpha\beta)^2\big ] \big(\frac{1}{4}-\!s_{ab}\big ) , \\
\label{wab13 } \hskip -.5 cm k_3^{(ab)}\!\!&=&\!\!\frac{2}{3}jr_3(\gamma^2+\alpha\beta)^2 \big(\frac{1}{4}-s_{ab}\big).\end{aligned}\ ] ] the kinetic energies @xmath95 depend on the same parameters @xmath440 as the exchange constants ( [ jd1c ] ) and ( [ jd1ab ] ) , on the orbital state ( [ and1 ] ) via the coefficients @xmath473 , and on the spin - spin correlations ( [ spins ] ) .
it is now straightforward to investigate the dependence of the magnetic and the optical properties in the four - sublattice classical state ( [ and1 ] ) on the effective parameters @xmath440 and on the type of oo given by the coefficients @xmath473 . _ a priori _
, the average electron density in the @xmath471 orbitals , @xmath479 , is different from the densities in the other two orbitals ( @xmath480 and @xmath481 ) , and the cubic symmetry of the expectation values @xmath366 and @xmath482 is explicitly broken by the oo given by eq .
( [ and1 ] ) , unless all the orbital amplitudes are equal , @xmath483 as argued in refs . .
first we consider this isotropic state and evaluate the exchange constants @xmath366 for increasing hund s parameter @xmath2 .
one finds then af superexchange @xmath484 which decreases with increasing @xmath2 [ fig .
[ fig : jd1eta](a ) ] , but the interactions remain af in the physically relevant range of @xmath485 . note that this classical consideration seriously overestimates the actual exchange interaction as one expects instead @xmath486 when quantum effects are included .
@xcite nevertheless , having no information about the optical excitations , we give here an example of estimating the exchange constants and optical spectral weights using the spectroscopic values @xcite for hund s exchange @xmath487 , and for coulomb intraorbital interaction of @xmath488 ev , which give @xmath489 .
assuming now a hopping parameter @xmath490 ev , one finds @xmath491 mev which gives @xmath492 mev . while one might expect that the effective @xmath10 is somewhat reduced as in the case of lamno@xmath4 , and the value of @xmath2 would then be larger , the present crude estimate is accidentally quite close to the experimental value of @xmath493 mev.@xcite , as obtained for the classical wavefunctions ( [ and1 ] ) with isotropic orbital amplitudes ( [ tiso ] ) , for increasing hund s exchange @xmath494 : ( a ) exchange interactions @xmath495 , and ( b ) kinetic energy terms @xmath95 at @xmath259 : high - spin ( @xmath270 , solid line ) , and low - spin intermediate - energy ( @xmath271 , dashed line ) and high - energy excitation ( @xmath496 , long - dashed line ) .
the total kinetic energy @xmath412 is shown by the dashed - dotted line . , width=283 ] in the af state , realized in latio@xmath4 , all three hubbard subbands contribute in the optical spectroscopy , and taking the nel state with @xmath497 one finds the highest spectral weight at @xmath259 for @xmath271 [ fig .
[ fig : jd1eta](b ) ] . for the realistic value @xmath498 the spectral weight of the lowest - energy ( high - spin )
hubbard subband @xmath499 is relatively weak , and is similar to that of the @xmath496 low - spin excitation at the highest energy [ fig .
[ fig : jd1eta](b ) ] .
note , however , that in early optical experiments the intensity was found to be practically independent of energy @xmath500,@xcite so different excitations might be difficult to separate from each other .
let us verify now whether the wavefunctions as given by eqs .
( [ and1 ] ) could also lead to isotropic af states for other choices of orbital amplitudes than in eq .
( [ tiso ] ) .
thus , we considered equal amplitudes @xmath501 and @xmath98 in the states parametrized by @xmath502 , @xmath503 with @xmath504 .
the normalization condition gives @xmath505 which vanishes at @xmath506 .
as expected , the exchange interactions are anisotropic for small @xmath507 ( fig . [
fig : jd1 ] ) .
the increasing occupancy of the @xmath67 orbitals with increasing @xmath502 results in a somewhat enhanced exchange interaction @xmath90 near the isotropic state ( for @xmath508 ) , while the interaction @xmath89 decreases almost linearly.@xcite at larger orbital anisotropy the interaction @xmath90 decreases as well .
it is remarkable that the af interactions become again isotropic close to @xmath509 which gives a large charge density @xmath510 in the @xmath471 orbitals , but this feature is not generic and will likely be modified by quantum effects .
note also that the state at @xmath511 in not a fm state , contrary to some earlier suggestions.@xcite and @xmath90 , eqs .
( [ jd1 ] ) ( solid and dashed line ) for the titanite model for different types of orbital order given by eqs .
( [ and1 ] ) .
the coefficients in the wavefunctions ( [ and1 ] ) vary between the isotropic state @xmath512 and the state with linear combinations of @xmath513 and @xmath471 orbitals , with the coefficients given by : @xmath514 and @xmath505 .
parameters : @xmath489 .
, width=283 ] finally we remark that the temperature dependence of the optical spectral weights can again be studied in a similar way as for lamno@xmath4 , by solving the spin - spin correlation functions at increasing temperature using the method of appendix [ app : spincomn ] for @xmath132 spins .
having no experimental data neither for latio@xmath4 nor for ytio@xmath4 , we shall limit ourselves to making predictions concerning the overall change of the spectral weights between low temperature @xmath515 where the spin - spin correlations are maximal , and room temperature @xmath402 ( or @xmath516 ) , where the spin - spin correlations may be neglected .
we take rigid classical wave functions with oo ( [ and1 ] ) and discuss first af states .
apart from the isotropic af phase ( [ tiso ] ) , called state @xmath14 , we consider in table [ tab : d1 ] also somewhat modified orbital amplitudes , as proposed by itoh _
et al._:@xcite @xmath517 called here state @xmath37 .
one finds that the exchange interactions @xmath89 and @xmath90 are then rather anisotropic ( see table [ tab : d1 ] ) , so we believe that this state can be experimentally excluded on the basis of the neutron data.@xcite .2 cm cccccccc & & & & & & & & & @xmath259 & @xmath402 & @xmath259 & @xmath402 & @xmath259 & @xmath516 & @xmath89 & & & & @xmath90 & & & & @xmath256 & 0.176 & 0.264 & 0.172 & 0.258 & 1.057 & 0.792 & @xmath262 & 0.465 & 0.232 & 0.476 & 0.238 & 0.000 & 0.190 & @xmath263 & 0.118 & 0.059 & 0.124 & 0.062 & 0.000 & 0.000 & @xmath264 & 0.759 & 0.555 & 0.772 & 0.558 & 1.057 & 0.982 & @xmath265 & 0.176 & 0.264 & 0.207 & 0.311 & 1.057 & 0.792 & @xmath266 & 0.465 & 0.232 & 0.451 & 0.226 & 0.000 & 0.190 & @xmath267 & 0.118 & 0.059 & 0.107 & 0.053 & 0.000 & 0.000 & @xmath268 & 0.759 & 0.555 & 0.765 & 0.590 & 1.057 & 0.982 [ tab : d1 ] due to the af order in the ground state for the wavefunctions @xmath14 and @xmath37 , the lowest energy high - spin excitations are suppressed at low temperature , but we predict that their weight should increase by about @xmath518% above the magnetic transition ( table [ tab : d1 ] ) . an even stronger temperature dependence is predicted for the low - spin part of the spectra with its intensity decreasing by a factor 2 when the magnetic order is lost .
thus , the spectral weight is shifted from the high - energy to the low - energy part of the optical spectrum with increasing temperature , and the total spectral weight decreases , both features being typical of af bonds .
an experimental confirmation of this trend would also allow one to identify the energy splitting between the high - spin and low - spin parts of the upper hubbard band , and to determine the value of hund s exchange @xmath13 from experiment .
although the anisotropy in spectral weights found in the state ( [ itoh ] ) proposed by itoh _
et al._@xcite might be too weak to be seen experimentally , the predicted anisotropy in the exchange interactions is certainly detectable .
it is quite remarkable that an isotropic fm ground state of ytio@xmath4 can also be described classically by the oo state with four sublattices , but the phase factors in the wavefunctions have to be then selected differently:@xcite @xmath519 we call this state the orbital @xmath520 state . neglecting again orbital ( quantum ) fluctuations , one finds @xmath521 in this state , while @xmath522 . at finite @xmath2
this leads indeed to an isotropic fm state with exchange constants , @xmath523 in lowest order @xmath524,@xcite so this type of oo leads to a markedly different exchange constant from that given by eq .
( [ jd1c ] ) . assuming again the same parameters as above one finds @xmath525 mev , which is somewhat higher than the experimental value @xmath526 mev.@xcite however , in view of the simplicity of the present considerations this agreement can be regarded as satisfactory .
consider now the spectral weights in the optical spectroscopy .
one finds rather simple expressions for the ( isotropic ) optical spectral weights for the @xmath520 state : @xmath527 thus , the optical spectrum of ytio@xmath4 would be quite different from that of latio@xmath4 ( see table [ tab : d1 ] ) . for the fm state at @xmath259
( with @xmath528 ) only the high - spin spectral weight @xmath529 contributes , and for the present parameters @xmath530 .
thus , future optical experiments on ytio@xmath4 combined with magnetic experiments could help to determine the superexchange constant @xmath1 .
furthermore , for the fm @xmath520 orbital state ( [ fmd1 ] ) one finds at high temperature a moderate reduction of the low - energy spectral weight by @xmath531% , and a corresponding increase for the higher energy excitation @xmath271 [ the third excitation does not contribute as long as the oo remains the same as in eqs .
( [ fmd1 ] ) ] .
the energy difference between these excitations is @xmath532 ( fig .
[ fig : levels ] ) , so they might serve to determine this parameter from the optical spectroscopy .
the above results show once again that the magnetic interactions and the optical spectral weights depend in a crucial way on the underlying orbital state .
once the oo in @xmath520 state ( [ fmd1 ] ) has been fixed , the exchange constants @xmath533 ( [ jf1 ] ) were completely determined by the superexchange @xmath1 ( [ j ] ) , and by hund s exchange parameter @xmath2 . for the optical spectral weights @xmath95 ( [ wf11])([wf13 ] ) one needs in addition the spin - spin correlation function @xmath534 .
most importantly , a strong temperature dependence of the optical spectra follows here from a classical orbital picture , and this should be a crucial experimental test of the validity of the present approach .
the last example of spin - orbital physics we want to give here are the cubic vanadates , in which the orbital degrees of freedom originate from the @xmath535 configuration of the v@xmath69 ions present in the ground state . here our goal will be to answer the question to what extent these systems could be described by decoupling spin and orbital degrees of freedom and assuming classical states with oo .
the oo in the cubic vanadates is complementary to the spin order the @xmath15-type af phase is found for @xmath197-type oo , and the @xmath197-type af phase is accompanied by @xmath15-type oo.@xcite here we shall consider in more detail @xmath197-type oo , relevant for the case of lavo@xmath4 .
we will show in particular that the superexchange spin - orbital model@xcite allows one to understand the microscopic reasons behind the @xmath15-af phase observed in lavo@xmath4 , and predicts that @xmath536 , as actually observed.@xcite the comparable size of fm and af exchange constants @xmath89 and @xmath90 , respectively , is unexpected when considering the goodenough - kanamori - anderson rules,@xcite which would suggest that @xmath537 is by a factor @xmath538 smaller . in the case of the vanadates
the superexchange between the @xmath134 spins of v@xmath69 ions in a perovskite lattice results from virtual charge excitations @xmath539 .
these charge excitations involve the coulomb interactions in the @xmath74 configuration of a v@xmath71 ion , parametrized for a pair of @xmath0 electrons , as for the titanates , by the intraorbital coulomb element @xmath10 , and by hund s exchange element @xmath13 [ see eq .
( [ jht ] ) and table [ tab : uij]].@xcite the excitation spectrum which leads to the superexchange model includes three states:@xcite ( i ) a high - spin state @xmath296 at energy @xmath52 , ( ii ) two degenerate low - spin states @xmath540 and @xmath541 at energy @xmath10 , and ( iii ) a @xmath542 low - spin state at energy @xmath445 ( fig .
[ fig : levels ] ) .
we parametrize it by two coefficients : @xmath543 and @xmath544 .
a general hamiltonian was already given in ref .
; here we shall analyze it assuming that the @xmath42 ( @xmath471 ) orbitals are singly occupied at each v@xmath69 ion , as concluded from experiment@xcite and from electronic structure calculations.@xcite therefore , the electron densities in the remaining two orbitals satisfy at each site @xmath56 the local constraint , @xmath545 one finds then that the superexchange @xmath546 for a bond @xmath547 along the @xmath67 axis consists of @xmath548 and for a bond in the @xmath91 plane , @xmath549 where the operators @xmath550 describe orbital pseudospins @xmath80 defined ( for each direction @xmath98 ) by the orbital doublet @xmath551 . at each site @xmath56 there is precisely one electron in these two orbitals , and both of them are active along the @xmath67 axis .
it has already been shown before@xcite that the quantum fluctuations play a decisive role in lavo@xmath4 , and the observed @xcite anisotropy and temperature dependence of the high - spin excitations in optical spectroscopy is reproduced when the theory includes them .
in the next sec . [ sec : d2oo ] we will present now the limitations of a simplified approach which is widely accepted @xcite and assumes that the oo in lavo@xmath4 is quite rigid already at the magnetic transition .
quantum fluctuations lead to important corrections which go beyond this picture and modify the temperature variation of the optical spectral weights , as we discuss in sec .
[ sec : d2mot ] . after decoupling of spin and orbital variables , the effective spin exchange constants @xmath89 and @xmath90 in eq .
( [ hs ] ) can be obtained from eqs .
( [ h1cd2])([h3ad2 ] ) by averaging over orbital correlations , as derived in ref . , @xmath552 , \\ \label{jabd2 } j_{ab}&=&\frac{1}{4}j\big[1-\eta r_1-\eta r_3 \nonumber \\ & & + ( r_1-\eta r_1-\eta r_3)\big(\frac{1}{4 } + \langle\tau_i^z\tau_j^z\rangle\big)\big].\end{aligned}\ ] ] they are determined by the orbital correlations , which result not only from the superexchange @xmath546 but also from the interactions with the lattice ( from the jt term),@xcite as discussed in detail in sec .
[ sec : d2mot ] .
here we shall consider first classical states with alternating @xmath553 and @xmath554 orbitals in the @xmath91 planes : @xmath197-type oo when these orbitals alternate also along the @xmath67 axis , and @xmath15-type oo for repeated either @xmath553 or @xmath554 orbitals along the @xmath67 axis .
these classical states are frequently assumed as a consequence of a strong jt term and observed gdfeo@xmath4 distortions .
@xcite such classical order would naturally follow from a strong jt effect , but it is still controversial whether the jt interaction is actually that strong in the vanadates . fortunately , there are already experimental results which can help to resolve this controversy , and we address this issue in more detail in sec .
[ sec : d2mot ] .
first we consider the limit of strong jt interaction with rigid oo .
this implies @xmath555 along each bond for @xmath197-type oo , so one finds from eqs .
( [ jcd2 ] ) and ( [ jabd2 ] ) fixed values of the exchange constants : @xmath556.\end{aligned}\ ] ] the fm interaction @xmath557 increases in lowest order linearly with hund s exchange @xmath2 , and the above values of the exchange constants give @xmath15-af spin order .
whether or not such a classical oo is realized in the ground state , can be investigated by analyzing the consequences of the present theory for the distribution of spectral weight in the optical spectroscopy .
the optical spectral weights follow by averaging the individual contributions to the superexchange , see eqs .
( [ hefa ] ) and ( [ opsa ] ) .
one finds for @xmath197-type oo the optical spectral weights : for polarization along the @xmath67 axis , @xmath558 for polarization in the @xmath91 plane , @xmath559 as in all other cases , they depend on two model parameters , @xmath1 and @xmath2 , and on the spin - spin correlations @xmath560 ( the oo is already fixed ) . again , as in the case of lamno@xmath4 ( sec .
iv ) , the analysis of the optical spectra suggests that the effective parameters are somewhat different from the atomic values , primarily due to the screening of both coulomb @xmath10 and hund s exchange @xmath13 .
we use here the parameters deduced from the neutron experiments@xcite and from the optical spectra@xcite one finds@xcite @xmath561 mev and @xmath562 .
these values imply that @xmath563 mev and @xmath564 mev , which lead to a mf estimate of the transition temperature @xmath565k .
this value has to be still reduced by an empirical factor@xcite ( close to 68% for @xmath134 ) to estimate the effect of thermal fluctuations , so one finds @xmath566 k in reasonable agreement with the experimental value of 140 k. @xcite we note that the above values of the exchange constants are in good agreement with the neutron experiments.@xcite it is instructive to test this classical approach by analyzing its predictions for the optical spectral weights .
we evaluated the spin correlation functions for bonds in the @xmath91 plane and along the @xmath67 axis using the oguchi method , and used an order parameter @xmath398 at neighboring sites , which acts on the considered bond by mf terms and vanishes at @xmath567 , as explained in ref . .
the correlation @xmath247 for fm bonds along the @xmath67 axis can be obtained analytically,@xcite while @xmath276 for the af bonds was determined by a numerical approach described in appendix [ app : spincov ] . as expected , the spectral weight due to the high - spin excitations dominates for @xmath67 polarization . however , when the oo is perfect , the anisotropy between @xmath256 and @xmath265 at @xmath259 , being close to 8:1 ( fig .
[ fig : kd2 ] ) , is now less pronounced than in the case of joint spin and orbital dynamics.@xcite at low temperature the low - spin excitations dominate the optical spectral weight for @xmath91 polarization .
one finds that the low - energy spectral weight along the @xmath67 axis @xmath256 decreases with increasing temperature .
simultaneously , the low - energy spectral weight in the @xmath91 planes @xmath265 increases , and the anisotropy goes down to @xmath568 at @xmath569 ( i.e. , at @xmath570 k for @xmath571 k ) .
it is quite remarkable that the present classical approach gives qualitatively a very similar distribution of the spectral weights and their temperature dependence for the fm and af bonds in lavo@xmath4 and in lamno@xmath4 ( cf . figs .
[ fig : anis ] and [ fig : kd2 ] ) . ) per bond ( [ hefa ] ) , as obtained for the @xmath15-af phase of the cubic vanadates : high - spin @xmath358 ( solid lines ) and low - spin @xmath572 ( dashed lines ) and @xmath573 ( long - dashed lines ) , for increasing temperature @xmath574 : ( a ) along the fm @xmath67 axis ( @xmath451 ) ; ( b ) within the af @xmath91 plane ( @xmath417 ) . in part
( a ) the experimental points from ref . for the low - energy spectral weight were reproduced , after proper rescaling to match the value of @xmath358 at @xmath252 .
parameters : @xmath562 . , width=283 ] .2 cm cccccc & & & & & & & @xmath405 ( k ) & @xmath47 & @xmath261 & @xmath47 & @xmath261 & @xmath89 & & & @xmath90 & & & @xmath256 & 1.640 & 1.181 & 0.0 & 0.0 & @xmath262 & 0.0 & 0.280 & 0.586 & 0.494 & @xmath263 & 0.0 & 0.0 & 0.465 & 0.392 & @xmath265 & 0.219 & 0.471 & 0.249 & 0.471 & @xmath266 & 0.916 & 0.532 & 0.871 & 0.532 & @xmath267 & 0.436 & 0.253 & 0.415 & 0.253 [ tab : d2 ] a crucial test of the present theory concerns the temperature dependence of the high - spin spectral weight along the @xmath67 axis @xmath256 , which according to experiment decreases by about 50% between low temperature and @xmath415 k.@xcite in the present theory based upon frozen oo this decrease amounts only to about 27% [ fig .
[ fig : kd2](a ) ] , and the maximal possible reduction of @xmath256 reached in the limit of @xmath575 is by 33% .
this result suggests that the frozen oo scenario in lavo@xmath4 is _ excluded by experiment _ ; further arguments supporting this point of view are given in sec .
[ sec : d2mot ] .
note also that the actual values of @xmath256 , shown in refs . and
, had to be reduced to match the classical prediction at @xmath259 .
a second classical state with @xmath15-type oo , as proposed for the low temperature @xmath197-af phase of yvo@xmath4,@xcite gives an af exchange constant along the @xmath67 axis , @xmath576 while the value of @xmath577 is the same as @xmath578 , see eq .
( [ jabcaf ] ) , so in the present classical approach it does not change when the oo changes from the @xmath197-type to @xmath15-type . in contrast , depending on the type of oo , the exchange constant along the @xmath67 axis can be either fm or af , as suggested by experiment@xcite and confirmed by the data in table [ tab : d2 ] . the actual values estimated with the same parameters as for lavo@xmath4 are : @xmath579 and @xmath580 mev .
here one finds that @xmath581 as in experiment , but the agreement with the experimental values,@xcite @xmath582 and @xmath583 mev , is much poorer .
hence , we conclude that quantum effects beyond the present classical analysis such as orbital fluctuations,@xcite orbital peierls effect,@xcite and spin - orbit coupling , @xcite play here an important role and have to be included in a quantitative theory . furthermore , while the weights in the @xmath91 planes are given by spin - spin correlations @xmath276 , as in eqs . ( [ wa21])([wa23 ] ) , for the spectral weights in the @xmath15-type oo one finds along the @xmath67 axis @xmath584 we evaluated these spectral weights ( see table [ tab : d2 ] ) for the standard value of @xmath562 and for the spin - spin correlation function @xmath247 determined within the oguchi method for an af bond along the @xmath67 axis ( appendix [ app : spincov ] ) , stabilized by the above value of the af interaction @xmath89 , and taking again @xmath567 .
it is along this axis where the theory predicts a markedly different behavior of the spectral weights from that found above for @xmath197-type
oo , see table [ tab : d2 ] . in spite of af bonds in all three directions ,
the spectral weight distribution is again anisotropic the weights are considerably higher for the @xmath91 planes due to the broken cubic symmetry in orbital space . in addition ,
the spectral weights obtained in @xmath91 planes at low temperature differ between the two af phases , as the mf terms are larger in @xmath197-af phase and make this case somewhat closer to the classical limit of @xmath416 .
particularly the prediction that @xmath585 , following from the classical @xmath15-type
oo , is easy to verify .
in fact , the experimental data are puzzling as one finds instead finite and temperature dependent spectral weight also for the low - energy regime in the @xmath197-af phase of yvo@xmath4,@xcite pointing out once again that the present calculation with frozen oo is oversimplified .
finally , we demonstrate that the scenario of strong jt interaction , quenching the orbital dynamics , can not apply to lavo@xmath4 .
we do so by investigating its consequences for the orbital transition temperature @xmath586 and for the temperature dependence of the optical intensity @xmath256 of the lowest multiplet transition .
consider first the transition temperature @xmath586 associated with the phase transtion into the state with oo .
we have already seen in secs .
iii and iv that strong orbital - lattice coupling in a perovskite structure would in fact necessarily decouple orbital and spin degrees of freedom and lead to @xmath587 , contradicting the experiment.@xcite the jt coupling between the jt - active local lattice modes @xmath588 and the pseudospin ( orbital ) variables @xmath589 ( which refer to the active @xmath590 orbitals along the @xmath67 axis ) may be written as follows , @xmath591 where @xmath588 is the appropriate linear combination of coordinates of the ligand ions next to site @xmath56 .
the above local coupling induces local distortions and an associated energy gain of @xmath592 ( the jt energy ) per site , and moreover generates a cooperative jt effect in the following way .
as the oxygens are shared by two neighboring vanadium ions in the perovskite structure of lavo@xmath4 , the jt distortions @xmath588 at nearest - neighbor sites are not independent from each other . hence the electron densities in @xmath513 orbitals at two vanadium ions at sites @xmath56 and @xmath57 in the @xmath91 plane ( see inset in fig .
[ fig : to ] ) get coupled to each other , basically because they depend on the actual displacement @xmath502 of the shared oxygen ion .
more specifically , this displacement generates interactions between the orbitals@xcite and one finds , taking care of the orthogonality constraint on the @xmath588 variables , @xmath593 with the interaction constant being given by @xmath594 . here
the coefficient @xmath595 is determined by the phonon spectrum , viz . by all branches
in which the local coordinates @xmath588 participate , @xcite and @xmath596 ( see e.g. refs . and
for how to estimate @xmath595 for the perovskite lattice ) . the jt interaction ( [ tautau ] )
reflects the cooperative nature of the jt problem .
it favors orbital alternation and thus supports the superexchange orbital interaction @xmath597,@xcite and the orbital model relevant for the @xmath15-af phase of lavo@xmath4 is then @xmath598 with @xmath599 .
the first term follows from the spin - orbital model , and @xmath600 .
while the superexchange contribution to @xmath601 is small ( adopting the values of ref . , @xmath602 mev , i.e. , @xmath603 mev ) , one finds that yet @xmath604 if one assumes @xmath605 , i.e. if one basically identifies the jt interaction with the jt energy , and further accepts the estimate for @xmath606 mev given in ref . .
then the ising term quenches the @xmath607-driven orbital dynamics and leads to a cooperative transition at @xmath586 which locks the orbitals in all three directions .
however , one should be aware that the total energy decrease produced by the jt distortion of the lattice , @xmath608 , as obtained in an _ ab initio _ calculation such as in ref .
, actually equals @xmath609 , i.e. it comprises both the local energy gain @xmath610 , which does not contribute to the ordering,@xcite as well as the jt ordering energy @xmath611 . the jt contribution to @xmath601
is therefore given by @xmath612 , and since @xmath595 is usually appreciably smaller than @xmath613 one should expect that @xmath614 is definitely smaller than @xmath608 , so that it is more likely that actually @xmath615 . for increasing jt interaction @xmath614 in : orbital - only model [ eq .
( [ omo ] ) , dashed line ] , and spin - orbital model [ eqs .
( [ omo ] ) and ( [ stau ] ) , solid line ] , compared with the nel temperature @xmath224 ( dashed - dotted line ) . both @xmath586 and @xmath224 in units of @xmath616 .
the inset shows intersite orbital correlations due to jt distortions in lavo@xmath4 .
, width=291 ] to gain more insight in the role of the jt interaction we performed exact diagonalization of eight - site chains along the @xmath67 axis , combined with a mf treatment of the orbital interactions in the @xmath91 plane , to determine the orbital transition temperature @xmath586 .
as expected , it scales with @xmath614 in the regime of large jt interaction @xmath617 ( but still @xmath618 ) as shown in fig .
[ fig : to ] , as was also found in ref . .
however , we were surprised to see that motome _
et al._@xcite have discarded this result considering it to be `` an artefact of the mf treatment '' , and argued that the coupling between neighboring sites is determined solely by @xmath616 .
the latter applies only in the limit of extremely anisotropic coupling , @xmath619 , which is not relevant here .
in fact , it is a classical result @xcite that the jt coupling of eq .
( [ jtcoup ] ) also induces intersite interactions between the orbitals as in eq .
( [ tautau ] ) which may actually dominate over the superexchange term and determine @xmath586 in the limit of large @xmath620 .
as @xmath586 is 20 - 80% above @xmath224 in other rvo@xmath4 ( r= ce , pr , nd , y , etc . ) compounds,@xcite it is not plausible that for r = la it is determined by superexchange alone .
therefore we argue that as the @xmath91 plane correlations are of ising - type , the mf result should be in fact a reasonable estimate for @xmath586 .
thus , the proximity of the orbital and the magnetic transitions in lavo@xmath4,@xcite @xmath621 ( @xmath622 ) , implies that the jt interaction is small , @xmath623 , and the jt splitting of the @xmath624-doublet is actually smaller than the superexchange energy scale itself , @xmath625 .
furthermore , as the spin and orbital exchange interactions are interrelated,@xcite spin disorder should reduce the effective orbital exchange @xmath616 .
indeed , we have verified that this follows from the full spin - orbital superexchange model@xcite which contains an extra term , @xmath626,\ ] ] describing the coupling of spins and orbitals along the @xmath67 axis .
however , @xmath586 obtained from the complete spin - orbital model , @xmath627 has about the same value as that found from @xmath628 alone ( see fig . [
fig : to ] ) , and the estimate @xmath629 remains valid . , in units of @xmath1 ( @xmath1 is the principal energy scale of the spin - orbital model ) , and ( b ) orbital correlations @xmath630 along the @xmath67 axis ( bottom ) , as obtained for the spin - orbital model eqs .
( [ horb ] ) , as given in ref . , with small @xmath631 ( solid lines ) , and the orbital model eq .
( [ omo ] ) with @xmath632 , as analyzed in ref .
( dashed lines ) . ,
width=302 ] moreover , one finds that for large @xmath614 ( e.g. , @xmath633 mev , i.e. for @xmath605 and @xmath634 mev @xmath635 , following ref . )
the temperature dependence of the optical intensity @xmath256 derived from @xmath628 is weak below 300 k ( fig .
[ fig : full ] ) , so adding orbital correlations in this range of parameters to the spin correlations of sec .
[ sec : d2oo ] can not improve the agreement with experiment .
hence , this analysis clearly shows that : ( i ) a substantial variation of @xmath256 below 300k ( @xmath12 , see fig .
[ fig : full ] ) observed in ref . is obtained only for small @xmath614 ; ( ii ) while neither spin correlations for frozen oo , nor orbital correlations that follow from @xmath628 alone would suffice , only the full spin - orbital model ( [ horb ] ) that includes coupled spin - orbital fluctuations is able to explain a large enhancement of @xmath256 at low temperature .
this was indeed demonstrated in ref . , where spin and orbital correlations were treated self - consistently , and only then the strong temperature dependence of @xmath256 could be successfully reproduced by the theory . summarizing , on comparing the results of figs .
[ fig : to ] and [ fig : full ] with the experimental data,@xcite one has to conclude that the proximity of @xmath586 and @xmath224 in lavo@xmath4 implies that @xmath614 is in fact much smaller than the total energy associated with the cooperative jt distortion @xmath636 mev estimated in ref . .
the enhanced optical conductivity along the @xmath67 axis @xcite also suggests that the local splitting @xmath637 is smaller than the dynamical orbital exchange @xmath616 , thus supporting the scenario of fluctuating orbitals@xcite in lavo@xmath4 .
therefore , the assumption of rigid oo , which was so successful for lamno@xmath4 , fails for lavo@xmath4 and the full quantum spin - orbital many - body problem has to be treated explicitly.@xcite
the main purpose of this paper was to make the experimental consequences of the superexchange spin - orbital models for correlated transition metal oxides with orbital degeneracy more transparent .
we formulated a general approach to the spectral weights in optical spectroscopy and illustrated it on several examples with different multiplet structure . while a general feature of all the superexchange spin - orbital models is a tendency towards enhanced quantum fluctuations,@xcite we gave reasons why in many situations
such fluctuations are quenched .
one then arrives at much simpler reduced models , where certain states with oo allow for a good insight into the mechanisms responsible for the magnetic interactions and for the optical spectral weights .
the common feature of all these cases is that the knowledge of only a few effective parameters , the superexchange energy @xmath1 , hund s exchange @xmath2 and the charge - transfer parameter @xmath24 , is sufficient to work out the quantitative predictions of the theory for a given type of orbital ordered state . in some of these cases the theoretical models
simplify so much that it is even possible to perform calculations with the help of a pocket calculator .
the cases of copper fluoride kcuf@xmath4 and the manganite lamno@xmath4 turned out to be simpler , and could be understood with frozen oo and quenched orbital dynamics below the structural transitions which occur at much higher temperatures than the nel temperature , @xmath638 .
however , we have also seen that particularly in @xmath0 systems , in the cubic titanates and vanadates , the orbital dynamics may not be quenched .
therefore , in some cases only the full quantum many - body problem gives proper answers for the experimental situation .
we came to these conclusions by analyzing in detail the consequences of decoupling of the spin and orbital degrees in states with rigid oo and by comparing the predictions of the theory with the experimental data , wherever available . in the undoped manganite lamno@xmath4
we could provide a consistent explanation of the magnetic and optical experimental data by deducing the values of the above effective parameters @xmath335 , and next showing that both the magnetic exchange constants @xmath90 and @xmath89 , and the anisotropy and the temperature dependence of the low - energy optical spectral weights can be reproduced by the theory in a satisfactory way . in the case of the copper fluoride
kcuf@xmath4 optical data were not available , but the constraints in the theory given by the exchange constants are so strong that we could conclude that the insulating state in this compound has charge - transfer character .
it remains to be verified by future experiments to what extent the predictions made here concerning the optical spectral weights and based on the classical picture with ordered @xmath3 orbitals apply to kcuf@xmath4 .
also for latio@xmath4 and ytio@xmath4 we investigated the classical states with oo given by certain wavefunctions which guarantee that the observed isotropic af or fm states are realized .
as in all other cases , the theory predicts in such states a rather pronounced temperature dependence and spectral weight transfers in the optical spectra near the magnetic transition .
future experiments will have to establish whether and to what extent such a scenario relaying on rigid oo could be valid
. however , already without these data we could demonstrate , by looking at the exchange constants , that there are certain indications that orbital fluctuations play a role and thus the quantum physics might dominate here over the thermal fluctuations of the spins alone .
.2 cm cccccc & & lavo@xmath4 & lamno@xmath4 & kcuf@xmath4 parameters & orbitals & @xmath0 & @xmath3 & @xmath3 effective & @xmath1 & 40 & 150 & 33 & @xmath2 & 0.13 & 0.18 & 0.12 & @xmath24 & @xmath639 & 0.6 & 1.2 microscopic & @xmath10 & 3.8 & 3.8 & 7.5 & @xmath13 & 0.50 & 0.67 & 0.90 & @xmath17 & @xmath640 & 3.5 & 4.0 [ tab : uj ] the case of the cubic vanadate lavo@xmath4 really shows that one may encounter the full complexity of the spin - orbital superexchange model when the spins and orbitals fluctuate coherently , and these fluctuations are essential to get a meaningful quantitative description of the optical data .
this case in particular demonstrates the importance of combining the magnetic and optical data . whereas the classical analysis of the spin fluctuations for frozen orbitals seems to suffice to explain the exchange constants in lavo@xmath4 , by looking at the optics one realizes that the picture of frozen oo induced by large jt coupling is here misleading , and the complete spin - orbital dynamics has to be considered instead . before concluding this paper ,
we summarize in table [ tab : uj ] the effective parameters of the spin - orbital models,@xcite and the possible values of the microscopic parameters the coulomb interaction @xmath10 , hund s exchange @xmath13 , and the charge transfer energy @xmath17 , that are consistent with these effective parameters .
the values of the superexchange constant are @xmath641 mev for kcuf@xmath4 and lavo@xmath4 , while for lamno@xmath4 the value of this parameter is much higher , @xmath642 mev .
this difference reflects a rather high value of the effective hopping parameter @xmath28 in the undoped manganite . by considering the optical data of lavo@xmath4 and lamno@xmath4 we came to the conclusion that hund s exchange @xmath13 is somewhat reduced from the respective atomic values,@xcite and we give already these reduced parameters in table [ tab : uj ] , accompanied by the corresponding values of @xmath10 . summarizing ,
we have illustrated a common approach to the optical and the magnetic data in mott insulators with orbital degeneracy , which provides the basis for a better theoretical understanding of the experimental constraints on the underlying phenomena and on the model parameters .
it is a unique feature of these systems that the superexchange interactions , and the spin , orbital , and composite spin - and - orbital correlations induced by them , are responsible for the distribution of spectral weight in the optical excitations . we hope that extending the present analysis of the high - energy excitations in the upper hubbard band by an analysis of the low - energy excitations that occur in doped systems , will allow to develop a quantitative theoretical approach designed to describe the optical spectra of doped transition metal insulators .
we thank j. van den brink , n. kovaleva , b. keimer , and c. ulrich for insightful discussions .
a. m. ole would like to acknowledge support by the polish ministry of science and information technology under project no . 1 p03b 068 26 .
here we present an analysis of the magnetic and optical data within an effective @xmath19 model for lamno@xmath4 , given by the @xmath284 term alone , i.e. , assuming @xmath239 in eq .
( [ hct4e ] ) . unlike in kcuf@xmath4
, the alternating oo in lamno@xmath4 refers to the orbitals _ occupied by electrons _ , and is characterized by a single angle @xmath167 in eqs .
( [ ood9 ] ) .
we performed an analysis of the exchange constants @xmath89 ( [ jc4 ] ) and @xmath90 ( [ jab4 ] ) as functions of the effective parameters @xmath1 and @xmath2 .
although these parameters can not be uniquely determined , we have verified that only a narrow range of the orbital angle @xmath643 gives reasonable agreement with experiment . here
we present the results obtained with @xmath644 mev and @xmath645 . both exchange constants exhibit a rather strong dependence on the orbital angle @xmath167 ( fig .
[ fig : j4u ] ) .
in contrast to the case of kcuf@xmath4 , the effective model parameters obtained at @xmath239 suffices to explain even almost quantitatively the observed exchange constants in lamno@xmath4 . this result is also consistent@xcite with the earlier analysis of ref . .
due to the strong dependence of the exchange constants @xmath90 and @xmath89 on the angle @xmath167 , one can exclude the oo of alternating directional @xmath382 orbitals , obtained with @xmath383 in eqs .
( [ ood9 ] ) . while it is frequently assumed that the @xmath0 superexchange ( [ jt4 ] ) is large , the present analysis shows that a consistent description of the magnetic properties requires instead a _ rather moderate value of _ @xmath342 in lamno@xmath4 . for the present parameters one finds @xmath646 mev,@xcite and @xmath342 increases by @xmath391% when the ct terms are included ( see sec .
[ d4:resj ] ) .
although it might be argued that @xmath393 mev is too large as the experimental value of @xmath224 in camno@xmath4 is only 110 k , @xcite one deals here _ de facto _ with different values of the @xmath10 , @xmath13 and @xmath17 parameters , namely with those for mn@xmath71 ions instead of the ones for mn@xmath69 ions , which apply in the intermediate excited states contributing to the superexchange in camno@xmath4 . yet the differences in these parameter values can not be large from the very nature of their physical origin , and so @xmath342 for lamno@xmath4 can not differ by more than a factor of 2 from the value of the @xmath0-induced superexchange in camno@xmath4.@xcite this supports our finding that a small ratio @xmath394 corresponds to realistic parameter values for lamno@xmath4 .
( [ jab4 ] ) and @xmath89 ( [ jc4 ] ) as functions of the orbital angle @xmath167 ( solid lines ) , obtained within the effective superexchange model @xmath284 ( @xmath239 ) .
experimental values@xcite of @xmath364 and @xmath89 for lamno@xmath4 ( indicated by circle and diamond ) are nearly reproduced for the oo with @xmath647 .
parameters : @xmath644 mev , @xmath645 .
, width=283 ] next we consider the temperature dependence of the spin - spin correlations and the optical spectral weights ( [ opsa ] ) . as in the full model discussed in sec .
[ d4:reso ] , one may assume frozen oo in the relevant temperature range below room temperature , and derive the temperature dependence of the optical spectral weights from that of the intersite spin - spin correlations ( [ spins ] ) , with the latter determined in a cluster method as explained in appendix [ app : spincomn ] . using the exchange interactions obtained with the present parameters at @xmath647 ( fig .
[ fig : j4u ] ) , one finds the empirical estimate@xcite @xmath648 k , which reasonably agrees with the experimental value @xmath401 k.@xcite the af bonds feel staggered mf terms for @xmath177 , so the spin - spin correlations @xmath247 had to be determined by a numerical solution , as explained in appendix [ app : spincomn ] .
both correlation functions change fast close to @xmath224 , reflecting the temperature dependence of the brillouin function , and remain finite at @xmath402 [ fig .
[ fig : op](a ) ] . for @xmath649 , and the order parameter @xmath398 ( dashed line ) ; ( b ) kinetic energies @xmath358 for the high - spin excitations , standing for the optical intensities ( [ opsa ] ) at low energy along @xmath649 axes ( solid lines ) , and total kinetic energies @xmath412 ( long - dashed lines ) , as obtained for lamno@xmath4 within the effective superexchange model @xmath284 for the oo given by an angle @xmath650 . parameters as in fig .
[ fig : j4u ] .
, width=291 ] the effective model also allows one to discuss the qualitative features of the spectral weight distribution in the optical spectra .
the theory predicts that only _
high - spin _ optical excitations are allowed at @xmath259 for the fm bonds in the @xmath91 planes , and one finds for this polarization a large kinetic energy @xmath265 [ fig . [ fig : op](b ) ] .
in contrast , the optical excitations for the af bonds along the @xmath67 axis are predominantly of _ low - spin _ character , and thus the kinetic energy @xmath256 is rather small , resulting in a large anisotropy @xmath651 of the low - energy optical intensities confirmed by experiment.@xcite when the temperature increases and the spin - spin correlations weaken , this anisotropy is reduced , but remains pronounced also at @xmath408 and still exceeds @xmath652 at @xmath415 k due to the persisting oo . as shown elsewhere,@xcite also a quantitative analysis of the present effective model gives a rather satisfactory agreement with the optical data@xcite in the entire temperature range . in contrast
, the total optical intensities have a much weaker temperature dependence and anisotropy [ fig .
[ fig : op](b ) ] .
thus , the main features on the experimentally observed intensity distribution and its temperature variation in the optical spectra@xcite are well reproduced already by the present effective model .
finally , we verify whether the used parameters @xmath644 mev and @xmath645 could be derived from the microscopic parameters of the ct model .
the value of @xmath653 ev follows from the spectroscopic values of the @xmath37 and @xmath15 racah parameters,@xcite so @xmath645 implies @xmath654 ev . knowing the value of @xmath1 , this leads to an estimated effective @xmath19 hopping element @xmath655 ev .
indeed , one finds that these microscopic parameters are in the expected range .
the effective coulomb interaction was estimated within the effective @xmath19 model for lamno@xmath4 as @xmath656 ev from spectroscopic data , @xcite so the agreement is close to perfect .
furthermore , taking the usually accepted values of @xmath657 ev and @xmath424 ev , following refs . and
, one finds a very plausible value of the effective hopping parameter @xmath658 ev , again quite close to the value derived above .
note that the experimental magnetic interactions in doped bilayer manganites were explained with a similar value of @xmath659 ev fixed by experiment.@xcite this favorable comparison emphasizes once again our main conclusion that the relevant model parameters can be derived by combining the results of magnetic and optical experiments , whenever available .
here we describe briefly a simple method which we used to determine the spin - spin correlation functions @xmath276 and @xmath247 for a pair of interacting spins @xmath280 at nearest neighbor mn ions in lamno@xmath4 .
the spin - spin correlations were obtained by performing a statistical average over the exact eigenstates for a single ( fm or af ) bond , found in the presence of mf terms acting at each spin of the bond and originating from its neighboring spins .
this , in fact , is the simplest cluster mf theory , known in the theory of magnetism as the oguchi method.@xcite as long as the mf terms vanish ( at @xmath408 ) , one finds for the various eigenstates with degeneracy @xmath660 , labeled by the total spin @xmath661 , @xmath662 depending on @xmath663 the scalar product @xmath664 varies between @xmath665 and @xmath666 .
consider first a fm bond @xmath58 in the @xmath91 plane , with the hamiltonian given by , @xmath667 where for the @xmath14-af phase with uniform order parameter in the @xmath91 plane @xmath668 the identical mf acting at both spins is @xmath669 at @xmath408 the eigenenergies of @xmath670 follow from eq .
( [ ss4ab ] ) .
for @xmath177 the order parameter @xmath398 could in principle be determined self - consistently in the present cluster approach . however , to gain a qualitative insight into the temperature dependence of @xmath276 it suffices to use a self - consistent solution of the mf equation , @xmath671 as the mf value @xmath672 is overestimated , it is appropriate to use in eq .
( [ sz4 ] ) the value of @xmath224 after an empirical reduction,@xcite leading to @xmath673 .
one finds an analytic solution for @xmath276 : @xmath674 where the terms @xmath675 refer to the subspaces of total spin @xmath663 , @xmath676 with @xmath677 and @xmath678 .
note that the term @xmath679 is absent in the numerator of eq .
( [ s4ab ] ) , because @xmath680 [ see eq .
( [ ss4ab ] ) ] . for an af bond @xmath58 along the @xmath67 axis
the hamiltonian is given by @xmath681 where the molecular field @xmath682 alternates between the sites @xmath56 and @xmath57 , so @xmath683 .
unlike for the fm bond , the field @xmath684 couples now states which belong to different values of @xmath663 ( but to the same value of @xmath685 ) . the staggered
mf plays no role for @xmath686 , while for @xmath687 the eigenstates have been found by diagonalizing the matrices , with diagonal elements following from eq .
( [ ss4ab ] ) , and offdiagonal ones
@xmath688 : @xmath689 -h_c & 0 \end{array } \right ) , \label{h4sz3}\ ] ] @xmath690 -2\sqrt{\frac{3}{7}}\;h_c & 0 & -4\sqrt{\frac{1}{7}}\;h_c \\[0.2 cm ] 0 & -4\sqrt{\frac{1}{7}\;}h_c & -3j_c \end{array } \right ) , } \label{h4sz2}\ ] ] @xmath691 -\sqrt{\frac{15}{7}}\;h_c & 0 & -\frac{16}{\sqrt{70}}\;h_c & 0 \\[0.2 cm ] 0 & -\frac{16}{\sqrt{70}}\;h_c & -3j_c & -\sqrt{\frac{21}{5}}\;h_c\\[0.2 cm ] 0 & 0 & -\sqrt{\frac{21}{5}}\;h_c & -5j_c \end{array } \right ) , } \label{h4sz1}\ ] ] @xmath692 -\frac{4}{\sqrt{7}}h_c & -\lambda & -\frac{12}{\sqrt{35}}h_c & 0 & 0 \\[0.2 cm ] 0 & -\frac{12}{\sqrt{35}}h_c & -3j_c & -2\sqrt{\frac{7}{5}}h_c & 0 \\[0.2 cm ] 0 & 0 & -2\sqrt{\frac{7}{5}}h_c & -5j_c & -2\sqrt{2}h_c \\[0.2 cm ] 0 & 0 & 0 & -2\sqrt{2}h_c & -6j_c \\[0.2 cm ] \end{array } \right ) . } \label{h4sz0}\ ] ] in this way a complete set of eigenstates @xmath693 with energies @xmath694 for @xmath695 was determined .
finally , the spin - spin correlation function @xmath247 was found using a standard formula @xmath696 where @xmath697 is the partition function .
the short - range spin - spin correlations @xmath534 for the cubic vanadates were determined using the oguchi method@xcite for a bond of interacting @xmath134 spins . as in the case of lamno@xmath4 ( see appendix [ app : spincomn ] ) , we solve exactly a single fm ( af ) bond @xmath58 with interaction @xmath89 ( @xmath90 ) , and the mf terms @xmath698 originating from neighboring spins and acting on each spin of the bond . in the present case the scalar product is given by @xmath699 for a fm bond , now along the @xmath67 axis , one finds an analytic solution .
@xcite this problem is analogous to that given by eq .
( [ h4ab ] ) . using the mf approximation ,
the order parameter @xmath398 was determined from eq .
( [ sz4 ] ) with @xmath134 , and @xmath700 was reduced to @xmath701 as appropriate for @xmath134 spins.@xcite the final result for @xmath247 reads @xmath702 where the terms @xmath675 originate from different subspaces of total spin @xmath703 , @xmath704 here @xmath705 @xmath706 and @xmath678 ; compare with eq .
( [ h4c ] ) . for an af bond @xmath58 in the @xmath91 plane
the mf hamiltonian is given by @xmath707 in analogy to an af bond in lamno@xmath4 ( see appendix [ app : spincomn ] ) , the correlation function @xmath276 can be found numerically by considering the subspaces of @xmath685 .
the molecular field , @xmath708 does not contribute to the highest eigenenergies @xmath709 in the subspace of @xmath710 [ cf . with eq .
( [ h4ab ] ) ] , while the remaining eigenstates had to be found by diagonalizing the matrices corresponding to other values of @xmath685 : @xmath711 -h_{ab } & -j_{ab } \end{array } \right ) , \label{h2sz1}\ ] ] @xmath712 -\frac{2}{\sqrt{3}}\;h_{ab } & -j_{ab } & -2\sqrt{\frac{2}{3}}\;h_{ab } \\[0.2 cm ] 0 & -2\sqrt{\frac{2}{3}}\;h_{ab } & -2j_{ab } \end{array } \right ) . } \label{h2sz0}\ ] ] by solving the above eigenproblems , we determined a complete set of eigenstates @xmath693 , with energies @xmath713 , labeled by @xmath714 .
therefore , the spin - spin correlation function @xmath276 for two interactings @xmath134 spins on an af bond follows in the present case from an equation similar to eq .
( [ s4c ] ) , with the relevant matrix elements @xmath715 now given by eq .
( [ ss2ab ] ) .
s. maekawa , t. tohyama , s. e. barnes , s. ishihara , w. koshibae , and g. khaliullin , _ physics of transition metal oxides _ , springer series in solid state sciences vol .
144 ( springer - verlag , heidelberg , 2004 ) .
m. cyrot and c. lyon - caen , j. phys .
( paris ) * 36 * , 253 ( 1975 ) ; s. inagaki , j. phys .
. jpn . * 39 * , 596 ( 1975 ) ; c. castellani , c. r. natoli , and j. ranninger , , 4945 ( 1978 ) ; * 18 * , 4967 ( 1978 ) ; * 18 * , 5001 ( 1978 ) .
h. f. pen , j. van den brink , d. i. khomskii , and g. a. sawatzky , , 1323 ( 1997 ) ; r. shiina , f. mila , f .-
c . zhang , and t. m. rice , , 144422 ( 2001 ) ; s. di matteo , n. b. perkins , and c. r. natoli , _ ibid . _
* 65 * , 054413 ( 2002 ) ; s. di matteo , g. jackeli , c. lacroix , and n. b. perkins , , 077208 ( 2004 ) ; s. di matteo , g. jackeli , and n. b. perkins , , 024431 ( 2005 ) ; , 020408(r ) ( 2005 ) .
fractional contributions due to exchange terms @xmath716 shown in the spectrum for the @xmath72 case [ see fig . [
fig : levels](a ) ] follow from the differences between the exchange elements @xmath40 in hamiltonian @xmath717 ( [ heegen ] ) ( see table i ) , and were obtained using the relation @xmath718 satisfied approximately by the experimental values for mn@xmath71 ions.@xcite , the values of @xmath719 and @xmath720 differ and depend on bond direction below the structural transition when the jt distortions remove the degeneracy of the @xmath3 orbitals.@xcite here the idealized case of @xmath721 is discussed .
d. a. tennant , t. g. perring , r. a. cowley , and s. e. nagler , , 4003 ( 1993 ) ; b. lake , d. a. tennant , and s. e. nagler , _ ibid .
_ * 85 * , 832 ( 2000 ) ; d. a. tennant , r. a. cowley , s. e. nagler , and a. m. tsvelik , , 13368 ( 1995 ) .
k. hirota , n. kaneko , a. nishizawa , and y. endoh , j. phys .
* 65 * , 3736 ( 1996 ) ; f. moussa , m. hennion , j. rodrguez - carvajal , h. moudden , l. pinsard , and a. revcolevschi , , 15149 ( 1996 ) ; g. biotteau , m. hennion , f. moussa , j. rodrguez - carvajal , l. pinsard , a. revcolevschi , y. m. mukovskii , and d. shulyatev , _ ibid .
_ * 64 * , 104421 ( 2001 ) .
although the oo in lamno@xmath4 is in principle unknown , we remark that the @xmath15-type oo with symmetric / antisymmetric combinations of @xmath3 orbitals , @xmath723 , is favored by a combination of the orbital superexchange and orbital interactions induced by the jt effect at @xmath408 .
@xcite we have verified that the amplitude of pairs of two orthogonal orbitals is larger for bonds in the @xmath91 planes than along the @xmath67 axis , and thus the fm @xmath3 term is here stronger by a factor close to three . the nearly linear behavior of @xmath89 is remarkable and one can prove it by performing an expansion in @xmath502 .
note also that for @xmath725 isotropic exchange interactions @xmath726 are recovered at @xmath506 .
a. v. mahajan , d. c. johnston , d. r. torgeson , and f. borsa , , 10966 ( 1992 ) ; h. kawano , h. yoshizawa , and y. ueda , j. phys .
jpn . * 63 * , 2857 ( 1994 ) ; y. ren , t. t. palstra , d. i. khomskii , a. a. nugroho , a. a. menovsky , and g. a. sawatzky , , 6577 ( 2000 ) ; m. noguchi , a. nakazawa , s. oka , t. arima , y. wakabayashi , h. nakao , and y. murakami , _ ibid . _ * 62 * , 9271(r ) ( 2000 ) .
having only the magnetic data@xcite for latio@xmath4 , we could not obtain meaningful values of the effective parameters @xmath1 and @xmath2 in this case , so this compound was not included in table [ tab : uj ] . | the temperature dependence and anisotropy of optical spectral weights associated with different multiplet transitions is determined by the spin and orbital correlations . to provide a systematic basis to exploit this close relationship between magnetism and optical spectra
, we present and analyze the spin - orbital superexchange models for a series of representative orbital - degenerate transition metal oxides with different multiplet structure . for each case
we derive the magnetic exchange constants , which determine the spin wave dispersions , as well as the partial optical sum rules .
the magnetic and optical properties of early transition metal oxides with degenerate @xmath0 orbitals ( titanates and vanadates with perovskite structure ) are shown to depend only on two parameters , viz .
the superexchange energy @xmath1 and the ratio @xmath2 of hund s exchange to the intraorbital coulomb interaction , and on the actual orbital state . in @xmath3 systems
important corrections follow from charge transfer excitations , and we show that kcuf@xmath4 can be classified as a charge transfer insulator , while lamno@xmath4 is a mott insulator with moderate charge transfer contributions . in some cases orbital fluctuations are quenched and decoupling of spin and orbital degrees of freedom with static orbital order gives satisfactory results for the optical weights .
on the example of cubic vanadates we describe a case where the full quantum spin - orbital physics must be considered . thus information on optical excitations , their energies ,
temperature dependence and anisotropy , combined with the results of magnetic neutron scattering experiments , provides an important consistency test of the spin - orbital models , and indicates whether orbital and/or spin fluctuations are important in a given compound .
[ _ published in : phys . rev .
b * 72 * , 214431 ( 2005 ) . _ ] |
understanding the link between supermassive black holes ( smbhs ) and their host galaxies is essential for deciphering the formation and evolution of galaxies .
galaxy / smbh co - evolution is expected given the observed correlations between smbh mass and galaxy properties ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and theoretical arguments that smbh growth is self - regulated by feedback ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a key prediction from this framework is that a galaxy experiencing rapid inflow of cold gas can evolve through various classes of starbursts and active galactic nuclei ( agn ) , such as ultra - luminous infrared galaxies ( ulirgs ) and quasars ( qsos ) , and that these phases are connected in an evolutionary sequence ( e.g. , * ? ? ?
with such a model , both a starburst and a heavily obscured agn can co - exist @xcite . testing this picture
requires not only finding signatures of obscured agn activity within starburst galaxies , but also interpreting them in the context of galaxy / smbh co - evolution .
locally , the most extreme starbursts are the ulirgs ( @xmath4 ; * ? ? ?
* ) , which are almost exclusively the result of a recent / ongoing major merger @xcite , which triggers both starburst and obscured agn activity . to evaluate smbh growth and the role of feedback during this critical phase
, it is necessary to robustly estimate the agn power during all observed phases . a primary challenge to determine the fraction of flux attributable to the agn is the reprocessing of its photons by the host galaxy s interstellar medium ( ism ) . in essence
, most diagnostics focus on finding signatures that are directly associated with the agn output : narrow - line region ( nlr ) emission , x - rays , torus emission , or radio emission . a popular approach to study , in particular , dust - obscured agn utilizes the mid - infrared ( mid - ir ) regime , loosely defined as 3 - 30@xmath5 , where typical starbursts have a spectral energy distribution ( sed ) with a minimum in continuum emission , while continuum emission from dust surrounding agn rises toward longer wavelengths owing to hot dust emission , potentially from a torus ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this idea led to the construction of diagnostic diagrams separating starburst from agn dominated sources , based on infrared space observatory ( _ iso _ ) data of local ir - luminous sources @xcite .
the sensitive infrared spectrograph ( irs ; * ? ? ? * ) aboard the _ spitzer _
space telescope @xcite enabled the expansion of this approach to larger samples covering a wider range of physical properties ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and permitted mid - ir diagnostic work on objects at higher redshifts . in particular , mid - ir diagnostics have been applied to galaxies at @xmath6 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the epoch of peak star - formation rate density ( e.g. , * ? ? ? * ) and peak quasar number density ( e.g. , * ? ? ?
* ) , which makes it a period crucial for evaluating the co - evolution of galaxies and smbhs .
it is also when lirgs and ulirgs ( see * ? ? ?
* for a review ) make a significant contribution to the global averaged luminosity density and to the cosmic infrared background @xcite .
other findings on high-@xmath7 ir - luminous galaxies suggest that high - redshift ulirgs may be unlike local ulirgs in that they may not all be late - stage mergers .
for example , @xmath7@xmath82 sources tend to be more starburst - dominated than @xmath9 sources of comparable mid - ir luminosity @xcite .
differences in both the spectral energy distributions ( seds ) and morphologies of high redshift ulirgs , as compared with local ( @xmath10 ) sources , do suggest that they are not analogous ( e.g. , * ? ? ?
* hereafter s12 ) .
furthermore , submillimetre galaxies , a subset of high - redshift ulirgs , may be a mix of early - stage quiescently star - forming mergers , late - stage merger - induced starbursts , and isolated disk galaxies @xcite .
the typically greater gas fractions @xcite imply that fueling both star - formation and black hole accretion is relatively easier , and therefore not necessarily analogous to what we see in local ulirgs . in both the local and high - redshift work
, it is not clear to what extent common diagnostics miss the crucial evolutionary stage when the agn is most deeply obscured .
it is possible that we can only detect the presence of an agn once sufficiently many unobscured lines of sight to the agn exist .
moreover , while ir diagnostics generally follow the expected agn / starburst fractions , the evolution of ir sed properties reflects a complex mix of these components which may be difficult to interpret for individual sources or samples ( * ? ? ? * hereafter v09 ) .
therefore it is desirable to search for signposts of agn powering that may suffer minimally from these complexities . to analyze this issue ,
numerical calculations have been used to capture the complex geometries and radiation sources relevant for disentangling agn activity from star formation .
originally restricted to simplified geometries ( e.g. , * ? ? ?
* ) , calculations of dust attenuation and reprocessing are now tenable for three - dimensional dynamical simulations on galactic scales including starbursts and smbh emission ( e.g. , * ? ? ?
studies of hydrodynamical simulations with dust radiative transfer in postprocessing have found that agn signatures , such as a power - law sed or warm ir colors in ulirgs , are generally associated with smbh activity @xcite but may also be caused by intense starbursts @xcite . analyzing the predicted far - ir emission from such simulations , @xcite and @xcite found that the ir signatures of star formation vary depending on whether the activity occurs in a quiescent or bursty mode , potentially complicating discriminators of agn and star formation activity .
one challenge raised by this body of work is how to interface the modeled smbh with the host galaxy when it is not feasible to model the central engine fully self - consistently , a problem similar to the one described by , e.g. , @xcite for modeling young star clusters . in this paper
we combine the high diagnostic power of mid - ir features , such as pah emission and the @xmath3 silicate feature ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , with three - dimensional hydrodynamical galaxy merger simulations to better understand existing _
spitzer _
irs data , and to prepare for future data from the _ james webb space telescope_. in this novel approach to agn diagnostics , we calculate infrared seds from simulations using dust radiative transfer , including a simple model for agn accretion and emission .
we then use mid - ir diagnostics to gain insight regarding agn activity when dust in the host galaxy may be important , and also to highlight areas in which such modeling techniques might be improved . in section [
s : simulations ] we describe simulations of two representative starbursts meant to bracket the level of obscuration in ulirgs , from which we compute and analyze the mid - ir seds as described in section [ s : diagnostics ] .
in section [ s : results ] , we show how these features depend on agn power , evolutionary stage , viewing direction , intrinsic agn sed , and assumptions about the dust and ism . in section [ s : comparison ] we construct commonly used diagnostic diagrams and evaluate their ability to estimate the agn luminosity fraction .
we consider implications of this work for future agn modeling and studies of the evolution of ir - luminous galaxies in section [ s : implications ] , and we conclude in section [ s : conclusions ] .
we explore a toy model for our indicators in appendix [ appendix ] .
[ cols="^,^,^ " , ] using a least absolute deviation method , we fit the relation @xmath11 for three quantities @xmath12 ( @xmath13 with mw dust , and both @xmath13 and @xmath14 with smc dust ) and each of our two ism models for the highly obscured simulation , and compile the results in table [ tab : lagnpredict ] . to measure the scatter in the correlations residuals
, we used the same scale estimator @xmath15 defined in section [ ss : angle ] , which is generally insensitive to outliers . in both cases , @xmath16 and each @xmath17 predict @xmath18 with roughly the same median scatter , @xmath19 .
however , we find that @xmath16 experiences a failure rate ( @xmath20 ) two to three times higher than @xmath17 in the highly obscured calculation .
both quantities predict @xmath18 equally well under the alternate ism treatment and for the less obscured candidate ( not shown ) , with @xmath21 and failure rates @xmath22 .
several simulated points at @xmath23 fail in both @xmath24 and @xmath13 . in these cases
, corresponding to the most highly obscured lines of sight in our @xmath25 calculation from figure [ fig : cameraseds ] , the mid - ir is so heavily attenuated that no @xmath26 signature from the central source survives in the observed sed ( see also section [ ss : angle ] ) .
diagnostic @xmath17 requires information about the galaxy s stellar mass ( rest - frame @xmath27 ) , limiting its immediate application to large new samples .
however , the rest - frame wavelength range @xmath17 requires will be covered by the _ james webb space telescope _ ( _ jwst _ ) , an observatory whose mid - infrared instrument ( miri ) ( e.g. , * ? ? ? * ) permits measurements of @xmath28 and @xmath29 for sources at @xmath30 .
the tight correlation we find between @xmath17 and @xmath18 reflects the possible benefit in combining information about @xmath29 and @xmath28 with near - ir color diagnotics from _ jwst _ itself ( e.g. , * ? ? ?
* ) , or by matching to sources from other ir surveys ( e.g. , * ? ? ?
* ; * ? ? ?
for the two example seds in figure [ fig : datased ] , whose known agn fractions are both @xmath31 , this procedure estimates @xmath32 and @xmath33 ( uncertainties are from the minimum theoretical scatter in the correlations ) for the high and low obscuration examples , respectively . .
[ fig : datased],width=307 ] [ ss : limitations ] a concern with the present method is that we can not be sure how our few idealized simulations fit into a cosmological context .
we have mitigated this issue by presenting two rather different ulirg candidates , by varying the relative power between the agn and star formation , and by varying the assumptions acting on dust obscuration and emission .
we are planning to build a library of model seds for a much wider variety of systems , enabling more robust studies across ir - luminous galaxy populations .
moreover , we have focused here on agn signatures at mid - infrared wavelengths , and information at longer wavelengths ( c.f .
* ) , such as the location of the far - ir peak , could further constrain the agn fractions in such sources ( e.g. , s12 ) . in an upcoming paper ( e. roebuck et al . in prep . ) , we include the far - ir peak predictions from these simulations in interpreting the properties of @xmath34 starbursts and quasars ( the s12 sample ) .
the simplest interpretation of figure [ fig : data_comparison ] is that analogues of our `` highly obscured '' and `` less obscured '' examples are both realized in observed sources , in addition to both mw and smc dust : all four models are required to cover the wide observed distributions of mid - ir color , pah emission , and silicate absorption
. a higher ratio of silicate to carbonaceous dust grains increases @xmath29 while shrinking @xmath28 ( figures [ fig : time_evolution_dust ] ) , without changing @xmath35 ( figure [ fig : combo_diagram ] ) , permitting estimates of @xmath36 under a variety of physical conditions , dust models , and optical depths . although different ism or dust models can produce broadly similar mid - ir trends , a separate measure of the agn power , such as x - ray flux , is essential to further constrain these models
furthermore , the recent discovery of high - redshift , hyper - luminous , hot - dust ulirgs @xcite suggests that we may lack a key phase in these few simulations , in which feedback initiates a transition between scenarios represented by our `` highly obscured '' and `` less obscured '' sed models .
several assumptions regarding the dust radiative transfer may lead to discrepancies between our model sources and real sources .
in particular , from figure [ fig : data_comparison ] we have identified two failures to which these assumptions likely contribute : 1 ) the mid - ir colors of the most agn - dominated highly obscured sources are redder than their real counterparts ( although bluer at @xmath37 ; see figure [ fig : combo_diagram ] ) , and 2 ) we do not simulate enough sources with little signs of pah emission or silicate absorption : @xmath38 _ and _ @xmath39 . the ir calculation in sunrise distributes the energy absorbed by small pah grains equally between emission in thermal equilibrium and in modes giving rise to the distinctive mid - ir pah lines @xcite . therefore , any change to this assumption or more sophisticated calculation , such as thermally fluctuating small grains ( e.g. , * ? ? ?
* ) , will affect the dust continuum shape and pah strength .
in addition , we do not currently model dust destruction , and our model for dust production is crude , both key areas where near - future progress may be tenable . by not allowing our radiation to destroy dust grains or alter the dust grain distribution ( or changing more generally the gas distribution owing to realistic feedback ) , we may be preventing ourselves from obtaining truly `` power - law '' agn seds with no mid - ir dust features that may arise after agn feedback disrupts the ism .
we have purposefully focused on how galaxy dust can affect the fixed smbh source sed ( with minor variations in section [ ss : torus ] ) , rather than how the physics on smaller scales play out in observations of ulirgs .
while high levels of galaxy dust can reproduce a number of properties of obscured quasars , the evolution of the central regions during these times of maximum obscuration will depend on the behavior of gas structures very close to the smbh .
this behavior may set the minimum scatter to be observed in correlations between these indicators .
using multiscale simulations , @xcite showed how gravitational instabilities propagate down to scales of @xmath40 pc during large - scale gas inflows .
they demonstrated that a number of properties of obscured agn attributed to a torus are explained by such dynamically - evolving accretion structures .
furthermore , these simulations suggest that the angular momentum vector of the small - scale `` torus '' can often be misaligned with the angular momentum of gas on galaxy scales , and the relative orientations can vary significantly in time @xcite .
thus , the attenuation from the host galaxy may vary more significantly than our current simulations suggest .
therefore we are motivated to develop methods of integrating knowledge of the impact of these different scales on the emergent seds as a means for further interpreting the evolution of ir - luminous galaxies .
recent work on merger simulations at higher resolution @xcite and with different hydrodynamics solvers ( e.g. , arepo : @xcite , c. hayward et al . in prep . ) shows that the average global evolution in simulations like the ones we analyze here , and hence conclusions made based on them , are robust .
however , such advances will enable more sophisticated treatments of feedback processes and hydrodynamics in gas with a complex phase structure and at smaller scales ; therefore future sed modeling could exploit these improvements to produce ever more detailed comparison libraries .
ccccc @xmath41 & @xmath42 & @xmath43 & @xmath44 & @xmath45 + @xmath13 , mw dust & @xmath46 & @xmath47 & @xmath48 & @xmath49 + @xmath14 , smc dust & @xmath42 & @xmath50 & @xmath51 & @xmath52 + @xmath13 , smc dust & @xmath53 & @xmath54 & @xmath55 & @xmath56 + @xmath41 & @xmath46 & @xmath57 & @xmath44 & @xmath56 + @xmath13 , mw dust & @xmath58 & @xmath47 & @xmath59 & @xmath60 + @xmath14 , smc dust & @xmath46 & @xmath61 & @xmath62 & @xmath63 + @xmath13 , smc dust & @xmath58 & @xmath61 & @xmath64 & @xmath65 +
in this work we used the detailed information contained in high - resolution hydrodynamical merger simulations to analyze the impact of galaxy - scale evolution of dust on interpretations of agn power in ulirgs derived from the mid - infrared . using the dust radiative transfer code sunrise , we calculated the sed evolution during two model starbursts meant to represent possible extremes of observed ulirg populations .
one merger is characteristic of bright starbursts at @xmath66 , while the other is a typical gas - rich merger at @xmath0 .
increasing computational resources will enable future analyses of many more galaxy and galaxy merger scenarios and more sophisticated modeling of the small - scale physics that are presently very uncertain .
in this initial work , we studied the variation of the modeled mid - ir seds as a function of viewing angle , source sed , dust grain distribution , galaxy ism clumpiness , and agn power , and evaluated the ability of common mid - ir diagnostics to predict the agn uv - ir luminosity fraction , @xmath36 .
we developed a possibly generic mid - ir estimator of @xmath36 that can be measured for new samples with next - generation instruments such as _ jwst_. our main conclusions are : 1 .
our modeling technique broadly reproduces the observed span of common mid - ir diagnostics used to disentangle agn and starburst activity .
although generally consistent with previous interpretations , none of the indicators straightforwardly predict @xmath36 because they depend non - linearly on the galaxy s properties , which can vary rapidly on timescales @xmath67 .
3 . highly obscured sources can be more agn - dominated than realized , even when the shape of their sed resembles a pure starburst .
a large enough column density reprocesses all of the direct agn light into longer wavelengths . 4 .
the mid - ir seds of sources with a powerful , highly obscured agn depend strongly on the direction from which the source is observed . along some lines of sight ,
significant dust self - absorption attenuates the ir flux and obscures the mid - ir signatures of the buried agn .
sources with a deep @xmath3 silicate absorption feature can be reproduced by models either with an extremely luminous agn and mw galaxy dust , or with an average - strength agn and smc galaxy dust .
it is possible to construct an idealized mid - ir indicator utilizing future _ jwst _ data to cleanly estimate @xmath36 while simultaneously constraining the dust grain model .
this research has made use of nasa s astrophysics data system .
the computations in this paper were run on the odyssey cluster supported by the fas science division research computing group at harvard university .
we thank the anonymous referee for valuable suggestions that improved this paper .
124 natexlab#1#1 armus , l. , et al .
2006 , , 640 , 204 . 2007 , , 656 , 148 bauer , f. , yan , l. , sajina , a. , & alexander , d. 2010 , , 710 , 212 beers , t. , flynn , k. , & gebhardt , k. 1990 , , 100 , 32 bouwens , r. j. , illingworth , g. d. , franx , m. , & ford , h. 2007 , , 670 , 928 brandl , b. , et al .
2006 , , 653 , 1129 bush , s. j. , cox , t. j. , hayward , c. c. , thilker , d. , hernquist , l. , & besla , g. 2010 , , 713 , 780 caputi , k. , et al .
2007 , , 660 , 97 chakrabarti , s. , cox , t. j. , hernquist , l. , hopkins , p. f. , robertson , b. , & matteo , t. d. 2007 , , 658 , 840 cox , t. , dutta , s. n. , matteo , t. d. , hernquist , l. , hopkins , p. f. , robertson , b. , & springel , v. 2006 , , 650 , 791 cox , t. , jonsson , p. , primack , j. , & somerville , r. 2006 , , 373 , 1013 dale , d. , et al .
2006 , , 646 , 161 . 2007 , , 655 , 863 dasyra , k. m. , et al .
2009 , , 701 , 1123 , t. , springel , v. , & hernquist , l. 2005 , , 433 , 604 dole , h. , et al .
2006 , , 451 , 417 dopita , m. , et al .
2005 , , 619 , 755 draine , b. t. , & li , a. 2001 , , 551 , 807 . 2007 , , 657 , 810 draine , b. t. , et al .
2007 , , 663 , 866 dwek , e. 1998 , , 501 , 643 eisenhardt , p. r. m. , et al .
2012 , , 755 , 173 fadda , d. , et al .
2010 , , 719 , 425 faucher - gigure , c .- a . , lidz , a. , zaldarriaga , m. , & hernquist , l. 2009 , , 703 , 1416 fazio , g. , et al .
2004 , , 154 , 10 ferrarese , l. , & merritt , d. 2000 , , 539 , l9 genzel , r. , et al .
1998 , , 498 , 579 georgantopoulos , i. , et al .
2011 , , 531 , a116 gordon , k. , & clayton , g. 1998 , , 500 , 816 gordon , k. d. , calzetti , d. , & witt , a. n. 1997 , , 487 , 625 gordon , k. d. , clayton , g. c. , misselt , k. a. , landolt , a. u. , & wolff , m. j. 2003 , , 594 , 279 groves , b. , dopita , m. , sutherland , r. , kewley , l. , fischera , j. , leitherer , c. , brandl , b. , & van breugel , w. 2008 , , 176 , 438 guhathakurta , p. , & draine , b. t. 1989 , , 345 , 230 gltekin , k. , et al .
2009 , , 698 , 198 hayward , c. , kere , d. , jonsson , p. , narayanan , d. , cox , t. , & hernquist , l. 2011 , , 743 , 159 hayward , c. c. , jonsson , p. , kere , d. , magnelli , b. , hernquist , l. , & cox , t. j. 2012 , , 424 , 951 hayward , c. c. , narayanan , d. , keres , d. , jonsson , p. , hopkins , p. f. , cox , t. j. , & hernquist , l. 2013 , , 428 , 2529 hernquist , l. 1990 , , 356 , 359 hernquist , l. , & katz , n. 1989 , , 70 , 419 hickox , r. c. , et al .
2012 , , 421 , 284 hnig , s. f. , beckert , t. , ohnaka , k. , & weigelt , g. 2006 , , 452 , 459 hnig , s. f. , kishimoto , m. , gandhi , p. , smette , a. , asmus , d. , duschl , w. , polletta , m. , & weigelt , g. 2010 , , 515 , a23 hopkins , p. , hernquist , l. , cox , t. , di matteo , t. , martini , p. , robertson , b. , & springel , v. 2005 , , 630 , 705 hopkins , p. , hernquist , l. , cox , t. , di matteo , t. , robertson , b. , & springel , v. 2006 , , 163 , 1 hopkins , p. , hernquist , l. , cox , t. , robertson , b. , & krause , e. 2007 , , 669 , 45 . 2007 , , 669 , 67 hopkins , p. , richards , g. t. , & hernquist , l. 2007 , , 654 , 731 hopkins , p. , younger , j. d. , hayward , c. c. , narayanan , d. , & hernquist , l. 2010 , , 402 , 1693 hopkins , p. f. , cox , t. j. , hernquist , l. , narayanan , d. , hayward , c. c. , & murray , n. 2013 , , -1 , 733 hopkins , p. f. , hayward , c. c. , narayanan , d. , & hernquist , l. 2012 , , 420 , 320 hopkins , p. f. , hernquist , l. , hayward , c. c. , & narayanan , d. 2012 , , 425 , 1121 hopkins , p. f. , et al .
2004 , , 128 , 1112 houck , j. , et al .
2004 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 5487 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
j.c.mather , 6276 houck , j. , et al .
2005 , , 622 , l105 imanishi , m. , nakagawa , t. , ohyama , y. , shirahata , m. , wada , t. , onaka , t. , & oi , n. 2008 , publications of the astronomical society of japan johnson , h. , & morgan , w. 1953 , , 117 , 313 jonsson , p. 2006
, , 372 , 2 jonsson , p. , groves , b. a. , & cox , t. j. 2010 , , 403 , 17 juneau , s. , dickinson , m. , alexander , d. m. , & salim , s. 2011 , , 736 , 104 katz , n. , weinberg , d. h. , & hernquist , l. 1996 , , 105 , 19 kennicutt , r. c. 1998 , , 498 , 541 kormendy , j. , & richstone , d. 1995 , , 33 , 581 lacy , m. , sajina , a. , petric , a. o. , seymour , n. , canalizo , g. , ridgway , s. e. , armus , l. , & storrie - lombardi , l. j. 2007 , , 669 , l61 laor , a. , & draine , b. t. 1993 , , 402 , 441 laurent , o. , mirabel , i. , charmandaris , v. , gallais , p. , madden , s. , sauvage , m. , vigroux , l. , & cesarsky , c. 2000 , , 359 , 887 , e. , et al . 2005 , , 632 , 169 leitherer , c. , et al . 1999 , , 123 , 3 levenson , n. , heckman , t. , krolik , j. , weaver , k. , & ycki , p. 2006
, , 648 , 111 levenson , n. , sirocky , m. , hao , l. , spoon , h. , marshall , j. , elitzur , m. , & houck , j. 2007 , , 654 , l45 li , a. , & draine , b. t. 2001 , , 554 , 778 lutz , d. , maiolino , r. , spoon , h. , & moorwood , a. 2004 , , 418 , 465 lutz , d. , spoon , h. , rigopoulou , d. , moorwood , a. , & genzel , r. 1998 , , 505 , l103 lutz , d. , et al .
1996 , , 315 , l137 magorrian , j. , et al .
1998 , , 115 , 2285 messias , h. , afonso , j. , salvato , m. , mobasher , b. , & hopkins , a. 2012 , , 754 , 120 morrison , r. , & mccammon , d. 1983 , , 270 , 119 mosteller , f. , & tukey , j. 1977 , data analysis and regression .
a second course in statistics ( addison - wesley , reading , ma ) narayanan , d. , hayward , c. c. , cox , t. j. , hernquist , l. , jonsson , p. , younger , j. d. , & groves , b. 2010 , , 401 , 1613 narayanan , d. , et al .
2010 , , 407 , 1701 nenkova , m. , sirocky , m. , nikutta , r. , ivezi , v. , & elitzur , m. 2008 , , 685 , 160 pier , e. , & krolik , j. 1992 , , 401 , 99 pope , a. , et al .
2008 , , 689 , 127 richards , g. , et al .
2006 , , 166 , 470 richards , g. t. , et al .
2006 , , 131 , 2766 rigopoulou , d. , spoon , h. , genzel , r. , lutz , d. , moorwood , a. , & tran , q. 1999 , , 118 , 2625 robertson , b. , hernquist , l. , cox , t. j. , matteo , t. d. , hopkins , p. f. , martini , p. , & springel , v. 2006 , , 641 , 90 roussel , h. , helou , g. , beck , r. , condon , j. , bosma , a. , matthews , k. , & jarrett , t. 2003 , , 593 , 733 sajina , a. , spoon , h. , yan , l. , imanishi , m. , fadda , d. , & elitzur , m. 2009 , , 703 , 270 sajina , a. , yan , l. , fadda , d. , dasyra , k. , & huynh , m. 2012 , , 757 , 13 sajina , a. , yan , l. , lacy , m. , & huynh , m. 2007 , , 667 , l17 sanders , d. , & mirabel , i. 1996 , , 34 , 749 sanders , d. , soifer , b. , elias , j. , madore , b. , matthews , k. , neugebauer , g. , & scoville , n. 1988 , , 325 , 74 schweitzer , m. , et al .
2006 , , 649 , 79 siebenmorgen , r. , haas , m. , krgel , e. , & schulz , b. 2005 , astronomy and astrophysics , 436 , l5 silk , j. , & rees , m. 1998 , , 331 , l1 smith , j. , et al .
2007 , , 656 , 770 snyder , g. , cox , t. , hayward , c. , hernquist , l. , & jonsson , p. 2011
, , 741 , 77 spoon , h. , marshall , j. , houck , j. , elitzur , m. , hao , l. , armus , l. , brandl , b. , & charmandaris , v. 2007 , , 654 , l49 spoon , h. w. w. , et al .
2004 , , 154 , 184 springel , v. 2005 , , 364 , 1105 . 2010 , , 401 , 791 springel , v. , di matteo , t. , & hernquist , l. 2005 , , 361 , 776 springel , v. , & hernquist , l. 2002 , , 333 , 649 . 2003 , , 339 , 289 stern , d. , et al .
2005 , , 631 , 163 . 2012 , , 753 , 30 sturm , e. , lutz , d. , tran , d. , feuchtgruber , h. , genzel , r. , kunze , d. , moorwood , a. , & thornley , m. 2000 , , 358 , 481 sturm , e. , et al .
2006 , , 653 , l13 surace , j. a. , & sanders , d. b. 1999 , , 512 , 162 surace , j. a. , sanders , d. b. , & evans , a. s. 2000 , , 529 , 170 tacconi , l. j. , et al .
2010 , , 463 , 781 tran , q. , et al .
2001 , , 552 , 527 tremaine , s. , et al .
2002 , , 574 , 740 veilleux , s. , et al .
2009 , , 182 , 628 vijh , u. p. , witt , a. n. , & gordon , k. d. 2003 , , 587 , 533 weingartner , j. c. , & draine , b. t. 2001 , , 548 , 296 werner , m. , et al .
2004 , , 154 , 1 witt , a. n. , thronson , h. a. , & capuano , j. m. 1992 , , 393 , 611 wright , g. , et al .
2004 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 5487 , society of photo - optical instrumentation engineers ( spie ) conference series , ed . j. mather , 653663 wu , j. , et al .
2012 , , 756 , 96 wu , y. , charmandaris , v. , hao , l. , brandl , b. , bernard - salas , j. , spoon , h. , & houck , j. 2006 , , 639 , 157 wuyts , s. , cox , t. j. , hayward , c. c. , franx , m. , hernquist , l. , hopkins , p. f. , jonsson , p. , & van dokkum , p. g. 2010 , , 722 , 1666 yan , l. , et al .
2007 , , 658 , 778 younger , j. , hayward , c. c. , narayanan , d. , cox , t. j. , hernquist , l. , & jonsson , p. 2009
, , 396 , l66
here we make plausibility arguments based on a simple toy model to validate the simulated behavior of the diagnostics we analyzed .
we assume a point source agn that has an sed composed of direct agn emission and dust re - emission from the torus .
we label this flux @xmath68 .
this is attenuated by a foreground screen such that the flux at infinity is @xmath69 . for the starburst component , @xmath70 ,
energy is absorbed by dust and re - emitted at ir wavelengths .
then , this dust emission is attenuated at larger scales with some effective optical depth .
the presence of silicate grains causes a sharp increase in the dust opacity near @xmath71 .
suppose that the attenuation curve is such that the attenuation on either side of the deep silicate feature is @xmath72 , but in the feature , the attenuation is @xmath73 where @xmath74 .
@xmath75 represents the ratio of the peak of the full extinction curve at @xmath76 to that arising without the silicate absorption ; here we analyze @xmath77 , which roughly matches the behavior we see in the simulations as expected with an extinction curve for mw - like dust @xcite .
as in section [ ss : simdiagnostics ] , @xmath78 .
we define the intrinsic monochromatic agn fraction @xmath79 .
this quantity is closely related to , but not directly proportional to , the bolometric agn fraction @xmath36 used throughout the rest of this paper and defined in section [ ss : agnfractiondefinition ] .
we estimate the conversion between these quantities in equation [ eq : bolcorrect ] . by deriving toy diagnostics as a function of @xmath80
we can controllably analyze their behavior in much the same way that we explored the simulations in sections [ s : results ] , [ s : comparison ] , and [ s : implications ] using the agnx0 , agnx1 , and agnx10 models defined in sections [ ss : alternatemodels ] and [ ss : galaxymodels ] . expanding @xmath26 , we find , @xmath81 where @xmath82 is a new parameter that characterizes the silicate absorption feature imprinted on the ir energy emitted from star formation . in the rt calculations ,
we obtain @xmath83 since @xmath84 when @xmath85 ( figure [ fig : time_evolution ] ) ; this corresponds to an effective optical depth to the star forming regions of @xmath86 .
since the sf gas is spatially extended , @xmath87 will change less with viewing angle than @xmath88 , allowing us to keep this parameter @xmath82 fixed .
the pah fraction @xmath28 is defined to be @xmath89 . add new variables @xmath90 and @xmath91 for the pah emission luminosity , and assume that the pah emission is some fixed factor @xmath92 times the continuum emission for either agn or star formation .
we expand to find , @xmath93 the true behavior is somewhat more complex than this , since in figure [ fig : time_evolution ] we see that @xmath28 varies quickly at high agn fractions and obscuration levels , but since both the numerator and denominator have a similar dependence on @xmath94 , there is minimal scatter with viewing angle at all times .
now , if we ignore sf by setting @xmath95 , we get @xmath96 , or @xmath97 for all @xmath98 .
the magnitude of this pre - factor ( @xmath99 ) implies that the depth of the silicate feature varies more strongly with optical depth , and thus viewing angle and time , than does the continuum .
its negative sign gives rise to the intuitive behavior that the silicate absorption feature becomes more apparent ( @xmath29 shrinks ) with increasing optical depth .
however , this trend fails in , e.g. , figure 6 , and the d points scattered low in figure [ fig : d_v_lagn ] . for simplicity , in the case of
mixed sf and agn , let @xmath100 and @xmath101 , but let @xmath102 vary .
when @xmath102 is very small , then the agn terms dominate , yielding @xmath103 , the same behavior as the no - sf case .
when @xmath102 is large enough that the agn term in the numerator of equation [ eq : tausi ] can be safely ignored , then we have @xmath104 .
this simplifies depending on @xmath102 : therefore , for powerful agn , @xmath29 depends strongly on @xmath102 , until @xmath102 gets too far above 1 , giving rise to the strong viewing - angle scatter we noticed in the simulations in section [ ss : angle ] . in figure
[ fig : appendixfig ] , we plot example curves showing how @xmath29 and @xmath28 depend on @xmath106 and @xmath102 in this toy model , which generally match the behavior we find in the simulations . the qualitative behavior of diagnostic @xmath13 can be derived in the same fashion . for this , we seek a similar type of model for @xmath35 .
roughly speaking ( figure [ fig : diagram ] ) , the pure - agn sed is such that @xmath107 , and the stellar sed is @xmath108 .
the added complication in this situation is the attenuation of the agn source depends on wavelength , roughly speaking as @xmath109 ( e.g. , @xcite ) .
we can use the same correction when combining @xmath29 and @xmath35 to model the @xmath13 diagnostic .
we get : @xmath110 in figure [ fig : appendixfig ] , we construct @xmath111 and @xmath13 analytically and plot their behavior as a function of @xmath106 and @xmath102 ( at 4.5@xmath5 ) .
a final difference between the simulated quantities and these toy models is that we derived these models as a function of a monochromatic agn fraction , while the quantity @xmath36 used in the rest of this paper is the bolometric agn fraction .
the correction depends on the sed shape of the various components , which we parametrize as @xmath112 , @xmath113 .
the agn bolometric correction factor @xmath114 is fixed since our input template has a fixed sed shape over the wavelengths used by sunrise , while the sf correction @xmath115 varies ; we find that @xmath116 in our rt calculations , and so we derive : @xmath117 figure [ fig : appendixfig ] shows @xmath26 , @xmath111 , and @xmath13 from these idealized formulae for a sampling of agn fractions and optical depths , confirming much of the behavior we observed in the results of the rt calculations .
, @xmath35 , and @xmath13 as a function of near - ir optical depth and agn fraction .
these simple models consist of an agn source buried behind a simple obscuring screen ( `` galaxy dust '' ) , plus an extended region of star formation that has a fixed spectral shape that does not depend on the optical depth to the agn . in the upper left panel
, @xmath118 has a strong dependence on @xmath102 , even at low levels of obscuration , giving rise to the large viewing angle scatter of @xmath29 in the simulations of section [ ss : angle ] : since the agn is a point source , its final sed depends only on attenuation along a single line of sight , which can vary widely , while indicators of activity over a larger spatial region ( such as @xmath28 ) have their attenuation averaged over many lines of sight .
several other phenomena from the simulations are demonstrated by these models : 1 ) generally , the silicate feature is enhanced as the agn fraction increases and as @xmath102 increases , up until the sf gas begins to dominate the attenuated agn sed , at which point the dependence of @xmath29 on @xmath102 reverses ; 2 ) the near - ir color @xmath35 effectively selects powerful agn when @xmath102 is small enough ( in our units , @xmath119 ) ; and 3 ) at larger @xmath102 , @xmath120 decreases and @xmath121 increases in such a way that the combination of these diagnostics , @xmath13 , remains roughly constant for a fixed agn fraction .
this constancy is shown in the bottom left panel , and gives rise to the tight correlation between agn fraction and @xmath13 in the bottom right panel .
these simple arguments roughly verify the rt calculations and motivate the behavior of the empirically defined indicator @xmath13 ( section [ ss : idealindicators ] ) .
[ fig : appendixfig],width=624 ] | we analyze the link between active galactic nuclei ( agn ) and mid - infrared flux using dust radiative transfer calculations of starbursts realized in hydrodynamical simulations . focusing on the effects of galaxy dust , we evaluate diagnostics commonly used to disentangle agn and star formation in ultraluminous infrared galaxies ( ulirgs ) .
we examine these quantities as a function of time , viewing angle , dust model , agn spectrum , and agn strength in merger simulations representing two possible extremes of the ulirg population : one is a typical gas - rich merger at @xmath0 , and the other is characteristic of extremely obscured starbursts at @xmath1 .
this highly obscured burst begins star - formation - dominated with significant pah emission , and ends with a @xmath2 period of red near - ir colors . at coalescence , when the agn is most luminous , dust obscures the near - infrared agn signature , reduces the relative emission from polycyclic aromatic hydrocarbons ( pahs ) , and enhances the @xmath3 absorption by silicate grains .
although generally consistent with previous interpretations , our results imply none of these indicators can unambiguously estimate the agn luminosity fraction in all cases .
motivated by the simulations , we show that a combination of the extinction feature at @xmath3 , the pah strength , and a near - infrared slope can simultaneously constrain the agn fraction and dust grain distribution for a wide range of obscuration .
we find that this indicator , accessible to the _ james webb space telescope _ , may estimate the agn power as tightly as the hard x - ray flux alone , thereby providing a valuable future cross - check and constraint for large samples of distant ulirgs . |
one of the most studied quantities to understand the fission process is the fragment mass and distribution , which is very closely related to the topological features in the multi - dimensional potential energy surface @xcite .
structures on the distribution of mass and may be interpreted by shell effects on potential energy of the fissioning system , determined by the strutinsky prescription and discussed by dickmann @xcite and wilkins @xcite . in order to investigate the fragments with very low excitation energy , using the time of flight method , signarbieux @xcite measured the fragment mas ( @xmath7 ) distribution for high values of fragment kinetic energy . because in that kinetic energy region there is no neutron emission , the time of flight technique permits separate neighboring fragment masses . in this work one
calculates the deformations of those fragments which must correspond to the most compact scission configurations , i.e. to the highest values of coulomb interaction energy between the two fragments .
in the process of thermal neutron induced fission of @xmath8 , a composed nucleus @xmath9 with excitation energy equal to neutron separation energy ( @xmath10 ) is formed first .
then , this nucleus splits in two complementary light and heavy fragments having @xmath11 and @xmath12 as mass numbers , and @xmath13 and @xmath14 as kinetic energies , respectively .
the q - value of the this reaction is given by the relation @xmath15 where m(z , a ) is the mass of nucleus with @xmath16 and @xmath7 as proton number and mass number , respectively .
this available energy at scission configuration is spend in prescission total kinetic energy ( @xmath17 ) , fragments interaction coulomb energy , @xmath18 , fragments deformation energy , @xmath19 where ( @xmath20 ) and ( @xmath21 ) are the light and heavy fragment deformation energy , respectively , and in fragments intrinsic excitation energy , @xmath22 where ( @xmath23 ) and ( @xmath24 ) are the light and heavy fragment intrinsic excitation energy , respectively .
then the balance energy at scission configuration results @xmath25 if there is no neutron emission , the light and heavy fragments reach the detectors with their primary kinetic energies equal to @xmath26 and @xmath27 , respectively .
the total primary fragments kinetic energy will be @xmath28 the maximal value of total kinetic energy is reached when the sum of tde and tdx is minimal , i.e. @xmath29 the most compact scission configuration occurs when maximal value of coulomb energy is equal to the available energy , i.e. @xmath30 in this case , from eq .
[ eq : tkeqdx ] one obtains the relations @xmath31 and @xmath32 not always this situation is possible to occur .
nevertheless we can assume that for each mass fragmentation the maximal value of total kinetic energy is obtained for similar condition , i.e. @xmath33 , @xmath34 and @xmath35 .
mo calculated by a drop liquid model with pairing and shell correction @xcite .
see text.,scaledwidth=50.0% ] tc calculated by a drop liquid model with pairing and shell correction @xcite .
see text.,scaledwidth=50.0% ] sn calculated by a drop liquid model with pairing and shell correction @xcite .
see text.,scaledwidth=45.0% ] the total energy of a nucleus is calculated at first approximation by a liquid drop model type ( @xmath36 ) , using the mass formula of myers and swiatecki @xcite .
the shell correction ( @xmath37 ) is calculated by the strutinsky s method @xcite , using nilsson hamiltonian @xcite : @xmath38 , \label{eq : nilssonv}\ ] ] where @xmath39 and @xmath40 are the nilsson s constants .
the pairing correction is calculated using the bcs method @xcite .
then , the relation for the total energy of the nucleus ( z , n ) results : @xmath41 where @xmath42 is the energy of a nucleus @xmath43 having deformation @xmath44 , and @xmath45 the energy in its spherical shape . the constant of the harmonic oscillator was the suggested by nilsson @xcite : @xmath46 as one said , the total fragments kinetic energy is close to the available energy for light and heavy complementary fragments with masses around @xmath47 and @xmath48 , respectively .
let us relate this result to the deformation for nuclei in this mass neighborhood .
the energies of nuclei @xmath49mo and @xmath49tc as a function of their corresponding deformations ( @xmath50)are presented on figs . [ fig:106108mo ] and [ fig:106108tc ] , respectively .
the assumed nilsson s constants @xcite for these nuclei are @xmath51 , @xmath52 , @xmath53 and @xmath54 .
as we can see , those nuclei have a prolate shape with to @xmath55 in their ground state .
if the fragment deformation changes from @xmath56 to @xmath55 the deformation energy will decreases by around 2 mev , while a change from @xmath55 to @xmath57 increases of deformation energy by 4 mev .
this result suggests that these nuclei are prolate and soft between @xmath56 to @xmath55 and became stiff for higher prolate deformations .
the energy as a function of deformation for nuclei @xmath58sn are presented on fig .
[ fig:130132sn ] .
the assumed nilsson s constants for these nuclei are @xmath59 , @xmath60 , @xmath61 and @xmath62 .
one can see that @xmath63sn is softer than @xmath64sn . for a deformation from @xmath65 to @xmath66
the nucleus @xmath63sn spends around 5 mev while the nucleus @xmath64sn , for the same deformation , spends 10 mev .
the neutron number number @xmath67 and proton numbers around @xmath68 correspond to spherical hard nuclei .
the above characteristics of light fragments , corresponding to masses from @xmath69 to @xmath70 , and their complementary fragments , corresponding to mass from @xmath71 to @xmath72 , makes possible that their maximal values of the total kinetic energy of complementary fragments ( @xmath73 ) are close to the available energy .
mo@xmath3 , @xmath4sn@xmath5 as a function of their deformation . @xmath74 and @xmath75 are the light and heavy fragment deformation .
, scaledwidth=50.0% ] for the the case of @xmath76u(n@xmath77 , f ) , the total kinetic energy of the couple @xmath2mo@xmath3 , @xmath4sn@xmath5 is almost equal to the available energy .
this results means that the corresponding scission configuration is composed by fragments in their ground state , i.e. @xmath78 for each one . on the fig .
[ fig : edecscission ] we can see the several equipotential energy of the scission configuration composed by those fragments given by the relation @xmath79 where @xmath75 and @xmath74 are the heavy and light fragment deformation energy , respectively , calculated using the nilsson model@xcite and @xmath18 is the interaction coulomb energy between the two fragments separated by 2 fm . on this curve
one obtains that for@xmath80 and @xmath81 the coulomb energy is equal to the available energy to 204 mev .
the results are similar to complementary fragments corresponding to the deformed transitional nuclei with @xmath11 between 100 and 106 ( @xmath82 between 60 y 64 ) and to the spherical nuclei with @xmath12 around 132 ( @xmath83 y @xmath84 ) . for the complementary fragments @xmath2mo@xmath3 and @xmath4sn@xmath5 , the maximal value of ce corresponds to ground state nuclei or close to that .
this case is unique .
other configurations will need deformation energy , which will be higher for the harder nuclei . on the fig .
[ fig:130132sn ] is presented the deformation energy for the spherical nuclei @xmath63se , @xmath85se and @xmath64se , respectively .
we can see that the double magic nucleus @xmath64se need 2 mev more than @xmath63se for going from the spherical state @xmath56 to the slightly deformed @xmath86 .
the fact that @xmath63se is no so hard as @xmath64se explain which the highest values of coulomb interaction energy correspond to values close to the available energy for @xmath76u(n@xmath87,f ) as well as for @xmath88u(n@xmath87,f ) .
from calculations of scission configurations from thermal neutron induced fission of @xmath8 and @xmath89 , one can conclude that the highest value of coulomb interaction energy between complementary fragments correspond to fragmentations ( @xmath2mo@xmath3 , @xmath4sn@xmath5 ) and ( @xmath2mo@xmath6 , @xmath4sn@xmath5 , respectively . for both cases
the calculated maximal value of coulomb interaction energy are equal to the available energy of the reaction for spherical ( @xmath90 ) heavy fragments and prolate ( @xmath81 ) complementary light fragments , which correspond to their ground states .
moreover the light fragments are soft between @xmath91 and @xmath81 and hard if they go to more prolate shapes ; while the heavy fragment @xmath4sn@xmath5 is no so hard as @xmath4sn@xmath92 .
the calculated maximal value of coulomb interaction energy is equal to the measured maximal value of total kinetic energy of fragments . the prescission kinetic energy and intrinsic excitation energy of fragments
are assumed to be null .
these results suggest that fission process take time to explore all energetically permitted scission configurations .
99 p. m@xmath93ller , d. g. madland , a. j. sierk , a. iwamoto , _ nature _ * 409 * ( 2001 ) 785 . f. dickmann and k. dietrich , _ nucl .
a _ * 129 * ( 1969 ) 241 . b. d.wilkins , e.p .
steinberg and r.r .
chasman , _ phys .
c _ * 14 * ( 1976 ) 1832 . | using a time of flight technique , the maximal values of kinetic energy as a function of primary mass of fragments from low energy fission of @xmath0 and @xmath1 were measured by signarbieux from calculations of scission configurations , one can conclude that , for those two fissioning systems , the maximal value of total kinetic energy corresponding to fragmentations ( @xmath2mo@xmath3 , @xmath4sn@xmath5 ) and ( @xmath2mo@xmath6 , @xmath4sn@xmath5 ) , respectively , are equal to the available energies , and their scission configurations are composed by a spherical heavy fragment and a prolate light fragment both in their ground state . _ keywords _ : low energy fission ; @xmath0 ; @xmath1 ; fragment kinetic energy ; cold fission + pacs : 21.10.gv ; 25.85.ec ; 24.10.lx + + |
electron transport in quantum wires ( qws ) now attracts considerable interest because of the fundamental importance of the electron - electron ( _ e - e _ ) interaction in one - dimensional ( 1d ) systems .
@xcite in addition , it is expected that the _ e - e _ interaction may produce important effects in the transport that would be attractive for applications .
however , up to now there is no clear and unambiguous knowlege of which effects of _ e - e _ interaction are observable under the realistic conditions of a qw with leads and what the reasons are for the deviations from conductance quantization ( observed experimentally @xcite ) , and very little is known about the electron transport under far from equilibrium conditions . in order to understand the transport properties ,
it is necessary to know the spatial distribution of the electric potential and the electron density in the structures under investigation .
many arguments and facts @xcite show that the leads play an essential role in the conductance , if the transport is investigated by measuring the electric current in an external circuit and by measuring the voltage drop between the leads .
it is obvious that the interaction of the qw with the leads is not weak . in the present paper
we investigate this interaction , considering the qw and the leads as a unified system .
an important problem that arises is how the electric field and the electron density are distributed at equilibrium ( i.e. , in the ground state ) , and how they are redistributed when an external voltage is applied .
we show that even in the equilibrium state the qw acquires a charge and a contact potential difference between the qw and the leads .
this phenomenon is similar to the well known contact potential difference in classical conductor systems .
the essential difference between the low - dimensional system under consideration and classical conventional 3d conductors is that the contact potential is not screened over a finite length but spreads over a length determined by the geometrical size of the structure .
@xcite the contact potential difference is determined by the difference in the chemical potentials @xmath0 of electrons in the qw and in the leads , when these subsystems are considered independently . by analyzing the chemical potentials we find that three cases are possible , depending on the wire radius and the background charge density : ( i ) the qw is charged positively , ( ii ) the qw acquires a negative charge , or ( iii ) the qw remains neutral as a whole .
we develop a model of a qw with leads in which the lead - wire interaction is taken into account .
it is based on a hartree - fock approach for the electrons in the qw and the representation of leads ( which are considered as electron reservoirs ) in a way that takes into account their 3d nature , but requires only a 1d calculation . using this model we investigate the distribution of the electron density and the electric potential at equilibrium , as well as under far from equilibrium conditions .
in particular , we emphasize the role of the exchange interaction effect on these quantities .
if in the absence of coupling the chemical potential of the wire exceeds that of the reservoir , @xmath1 , the coupled system expels electrons from the qw into the leads , when equilibrium is established . as a result a potential well appears in the qw .
the wire acts as a charge donor to the reservoir .
if @xmath2 , electrons are attracted by the wire , and a potential barrier arises between the qw and the leads .
the wire acts in this case as an acceptor . in the case
where @xmath3 , the electron density is redistributed only within the qw and only friedel oscillations arise near the contacts .
the friedel oscillations appear in all cases ; however , they are superimposed on a much more slowly varying potential created by the acquired charge .
the slowly varying potential has an amplitude that is large compared to the friedel oscillations and thus dominates the scattering processes .
the exchange interaction strongly affects the potential shape and somewhat enhances the friedel oscillation amplitude .
when an external voltage is applied , the chemical potentials in the electron reservoirs are shifted relative to each other , disturbing the electron flows in the qw . as this takes place , the electron density , the potential , and the exchange energy
are changed self - consistently .
the importance of the electrostatic potential distribution in quantum wires with leads , especially for the investigation of time - dependent transport , and for nonlinear transport , has been emphasized previously , @xcite but quantitative calculations have to our knowledge not been reported thus far .
under far from equilibrium conditions , when the applied voltage exceeds the fermi energy in the qw , the electron density is substantially redistributed between the qw and the reservoirs , giving rise to a very strong variation in the potential landscape . in turn
the potential produces a variation in the electron density .
the connection between the electron density and the potential is very important for the understanding of nonlinear transport .
the need of a self - consistent treatment has been emphasized by landauer .
@xcite within the scattering approach , in the weakly nonlinear regime , it has been investigated by christen and one of the present authors @xcite and by ma , wang and guo .
@xcite the calculation of the present paper allows us to investigate the strongly nonlinear transport in the system under consideration .
if the applied voltage is high enough , the self - consistent connection of charge and potential gives rise to an instability of the electron density distribution and ultimately leads to multistability of the electron states in the qw .
this means that several stable states with different spatial distributions of the electron density and the potential are possible at a given applied voltage .
the paper is organized as follows . in sec .
[ contact - potential ] the chemical potential difference between the decoupled qw and the 2d electron reservoir is analyzed .
section [ model ] describes the model of the qw with leads .
section [ eq - state ] contains the results of the numerical calculations of the electron density and potential distribution in the qw with leads . in sec .
[ multi - stability ] the multistability of the electron states is described that appears for far from equilibrium conditions .
to be specific we consider a qw connecting two regions of a 2d electron gas .
we assume that there are no nearby gates and that all electric field lines emanate and terminate either on the wire or on the 2d electron gas .
qw structures of this kind are produced by etching of heterostructures with a 2d electron gas .
such structures are widely used in experiments .
@xcite first we investigate such a qw separately from 2d electron reservoirs .
the uncoupled wire is charge neutral .
the electron charge is concentrated inside the qw , while the compensating positive charge of the impurities is really located in the immediate vicinity of the qw or at its surface .
the decoupled qw and the electron reservoir have their own chemical potentials @xmath4 and @xmath5 which are generally not equal each other .
we are interested in the chemical potential difference @xmath6 between the qw and the reservoir . according to seitz s theorem @xcite the chemical potential in the qw
is determined by the fermi energy of noninteracting electrons and the self - energy @xmath7 which takes into account the _ e - e _ interaction , @xmath8 here @xmath9 is the first subband energy caused by transverse confinement ( @xmath10 is the qw radius , @xmath11 is the effective mass of electrons ) .
@xmath12 contains the contributions arising from the exchange and correlation interaction as well as from the electron interaction with the positive background charge .
the exchange and correlation energy was investigated in the recent paper of calmels and gold @xcite using the self - consistent theory of singwi , tosi , land and sjlander @xcite for the case where only the lowest subband is occupied . the hartree energy
is easily estimated if we assume that the positive charge is located at the surface of the qw .
these calculations lead to the following expression for the chemical potential @xmath4 of the qw in terms of the dimensionless parameters @xmath13 and @xmath14 ( with @xmath15 the effective bohr radius and @xmath16 the 1d electron density ) : @xmath17 where @xmath18 is the effective rydberg . in eq .
( [ mu1d ] ) the first term is the lowest subband energy , the second term is the kinetic energy , and the third term represents the exchange and correlation energy .
there are two expressions for the exchange - correlation term depending on whether @xmath19 or @xmath20 : @xmath21 and @xmath22\ , \qquad \mbox{if}\quad r_s>1\,.\ ] ] the last term in eq .
( [ mu1d ] ) is the hartree energy of the electron interaction with the positive background ; @xmath23 is a numerical factor that depends on the radial distribution of the electron density . if the electron density is distributed uniformly , @xmath24 .
the chemical potential in the 2d reservoir can easily be obtained from the known expression @xcite for the electron energy as a function of the density parameter @xmath25 ( where @xmath26 is the 2d electron density ) . in the high density case , @xmath27 , one obtains @xmath28 where @xmath29 is the thickness of the 2d layer .
using eqs .
( [ mu1d ] ) and ( [ mu2d ] ) , the chemical potential difference @xmath30 is calculated as a function of the qw radius and the density parameter @xmath31 .
the results of these calculations are illustrated in fig .
[ delta - mu ] , where the @xmath32 dependence on the wire radius @xmath10 is shown for various values of @xmath31 .
here we consider @xmath31 as an independent parameter because the background charge density depends on external factors , such as the charge absorbed at the wire surface .
figure [ delta - mu ] shows that in sufficiently thin wires the chemical potential is higher than in the reservoir .
however , with increasing radius of the qw the chemical potential in the wire can become lower than in the reservoir .
the chemical potential difference @xmath6 is caused by all energy components contributing to the chemical potentials in the qw and in the reservoir .
as an example it is instructive to consider the estimations for the specific case where the qw diameter is equal to the 2d layer thickness , @xmath33 , and the background charge density per unit area , @xmath26 , is the same in the reservoir and in the qw .
the latter means that @xmath34 . in this case
the energies contributing to the chemical potentials are estimated as follows .
the confinement energy in the qw is approximately twice that in the reservoir .
hence the confinement energy causes @xmath4 to rise with respect to @xmath5 .
the ratio of kinetic energy in the qw to that in the reservoir is @xmath35 .
since the qw is supposed to be a 1d system , the product @xmath36 must be small .
hence the fermi energy @xmath37 in the qw is noticeably smaller than the fermi energy @xmath38 in the reservoir .
this results in lowering @xmath4 relative to @xmath5 .
the ratio of the exchange - correlation energies in the case of @xmath19 is estimated as @xmath39 .
the exchange - correlation energy in the qw is seen to be larger than that in the reservoir .
taking into account that the exchange - correlation energy is negative , we conclude that it lowers @xmath4 with respect to @xmath5 .
if the wire is now coupled to the 2d reservoir , we can distinguish three cases .
for @xmath40 , electrons are transferred from the qw to the reservoir , so that the positive background charge dominates the electron charge in the wire . if @xmath41 , the reservoir supplies electrons to the qw producing an excess negative charge . when @xmath42 , the electron density is not redistributed between the wire and reservoir .
@xcite the electron density redistribution continues until an equilibrium state is attained in the whole system such that there exists a uniform electrochemical potential . as this takes place , a charge and
a built - in electric field @xmath43 appear in the qw . if the reservoir conductivity is high , the built - in field satisfies the condition @xmath44 .
such a charge transfer is similar to what happens when a contact potential difference appears in classical 3d systems .
however , an essential difference is in the distance over which the contact field is screened . in the 3d case
the contact field is screened over a finite length ( e.g. , the debye length or the thomas - fermi screening length ) .
in the mesoscopic structure considered here , the contact field is produced by the charge , one part of which is situated in the 1d wire and the other on the surface of the reservoir adjacent to the wire .
one can easily see , that owing to the 3d nature of the electric field , any distribution of charges in a qw can not screen the contact field over a finite distance .
thus the question arises of how the charge density and the electric field are distributed in the qw and over the reservoir surface . in the case of a high enough electron density , the interaction effects are not strong , and this problem can be solved analytically using the thomas - fermi approximation .
@xcite in the present paper , we study this problem using a numerical solution of the schrdinger equation coupled to the poisson equation within the hartree - fock approximation .
finding the self - consistent electron density and the potential in a qw coupled to electron reservoirs is a rather complicated problem because the electron density redistribution between the qw and the reservoirs produces a strong variation of the electron energy in the qw .
the energy variation is estimated as @xmath6 , which is shown above to be of the order of the fermi energy .
in essence , the qw and the reservoirs should be considered as a unique quantum system . to our knowledge , such a problem has not been studied to date . in the present paper , we investigate it using a simplified model based on the hartree - fock approximation , which allows one to take adequately into account the charge accumulated in the qw both for the equilibrium state and under the far from equilibrium conditions appearing when an external voltage is applied .
this approach allows one also to study the exchange interaction effect on the charge accumulated and the electric potential .
however , it does not take into account the electron correlation energy .
the ratio of the correlation energy to the exchange energy depends on the electron density . for 1d conductors
this ratio can be estimated according to ref . .
the correlation energy is negligible when @xmath45 . under this condition
the hartree - fock approximation is justified .
it is instructive to estimate numerically the number of electrons in a gaas qw when @xmath46 .
the fermi energy and the electron density are respectively @xmath47 4 mev and @xmath48 @xmath49 .
our computation procedure works well when the qw length is not too large in comparison with the fermi wavelength .
if one puts the qw length equal to @xmath50 m , the total number of electrons in the qw is estimated as about 18 .
thus the system contains about ten electrons in a qw open to reservoirs .
the number of electrons really existing in the wire and their density distribution is determined by the qw length , the background charge , the applied voltage , and the _ e - e _ interaction energy .
the parameters of the qw estimated above are easily realizable in experiment . in order to investigate cases with different relative positions of the chemical potentials in the qw and reservoirs ,
we introduce a positive background charge density @xmath51 in the qw , which is considered as a parameter of our model . by varying @xmath51
it is possible to realize any relative position of the chemical potentials of the uncoupled system .
charges on the qw surface are not taken into account in our present consideration .
another simplification is that only the lowest subband in the qw is considered . a 1d sketch of the energy diagram of a qw with leads and the electron flows ( in a far from equilibrium situation ) is shown in fig . [ diagram ] .
here @xmath52 is the confinement energy in the qw and @xmath53 are the chemical potentials in the reservoirs .
the positions of @xmath54 and @xmath55 relative to the conduction band bottom of the corresponding reservoirs are fixed because the electron gas in the reservoirs is incompressible .
however , @xmath54 and @xmath56 are shifted relative to each other in the presence of an applied voltage @xmath57 .
the following energies contribute to the potential shape of the structure : the confinement energy in the qw ; the potential produced by the external voltage source ; the hartree and exchange energies in the qw . in the reservoirs ( @xmath58 , @xmath59 being the wire length ) the electron density is taken to be so high that the _ e - e _ interaction energy can be ignored in comparison with the kinetic energy .
moreover , the reservoirs are assumed to be ideally conducting and hence can be treated as equipotentials . inside the qw
the _ e - e _ interaction potential @xmath60 is determined by both the direct interaction of electrons with each other and the indirect interaction via image charges induced by electrons in the reservoirs . because of this , the interaction potential depends on the coordinates @xmath61 of the interacting electrons separately , rather than simply on their difference . in the qw
the one - electron wave functions @xmath62 are characterized by quantum numbers : @xmath63 ( @xmath64 indicates electrons incident on the qw from the left reservoir and @xmath65 corresponds to electrons incident on the qw from the right reservoir ) ; @xmath66 is a wave number in the left ( @xmath64 ) or right ( @xmath65 ) reservoir ; @xmath67 is a spin variable . in this paper
we ignore changes in @xmath67 and suppose that the states with opposite spins are equally occupied .
the effects of possible spin polarization in a qw will be considered elsewhere .
the electron transport in the reservoirs , close to the transition between the reservoir and the qw , is modeled by a 1d schrdinger equation without interaction .
this model allows one to simulate adequately the transmission probability between the reservoirs and the qw .
the value of the transmission probability calculated in this way as compared with that for a true 2d to 1d transition at the interface of the reservoir and the wire has been investigated by direct comparison of the transmission probabilities of these two geometries .
@xcite the difference depends on the energy but does not exceed 15% even close to the transmission threshold .
the physical reason for the success of the purely 1d model is that near the chemical potential level , of all the electron waves in the reservoir only those couple effectively to the wire that have a wave vector that is nearly parallel to the wire axis ( collimation effect @xcite ) .
we also emphasize that in the problem under investigation the electron density of the 1d schrdinger problem is needed only within the wire .
it is unimportant for the reservoirs because they are treated as equipotentials and therefore the calculated electron density distribution in the reservoirs does not directly affect the _ e - e _ interaction in the qw .
the calculated electron wave functions in the reservoirs affect only the transmission probability through the contacts . in the reservoirs , @xmath62 is thus @xmath68 + r_r \exp[-ik_r ( rx+l/2)]\ , & \quad \mbox{if}\quad r x<-l/2\\ t_r \exp[ik'_r ( rx - l/2 ) ] \ , & \quad \mbox{if}\quad r x > l/2\,,\\ \end{array } \right .\ ] ] where @xmath69 , and @xmath70 is the applied voltage . in the qw , @xmath62
is determined by the equation @xmath71 \psi_{r , k_r , s } = \varepsilon_r(k_r ) \psi_{r , k_r , s}\,.\ ] ] in eq .
( [ schrodinger ] ) the potential energy has the following components .
@xmath72 is an effective potential that simulates the electron confinement in the qw . in the simplest case , we can assume that @xmath73 for @xmath74 and @xmath75 for @xmath58 .
@xmath76 is the hartree energy , @xmath77 \,,\ ] ] with @xmath78 being the positively charged background density , @xmath79 being the electron density , @xmath80 and @xmath81 the electron distribution function in the reservoirs . in eq .
( [ hartree ] ) we assume for simplicity that the radial component of the background charge density is the same as the electron density . the _ e - e _ interaction potential @xmath60 that appears in eq .
( [ hartree ] ) depends on the spatial configuration of the leads . in what follows
the numerical calculations are carried out for the case where the leads are represented as two plates perpendicular to the qw .
this configuration is convenient for further calculations because in this case a relatively simple analytical expression is obtained for @xmath60 .
@xcite this form of the interaction potential allows one to take into account not only the direct coulomb interaction of electrons but also their interaction via image charges induced on the lead surfaces .
the interaction potential is @xmath82\ , \sinh\left[y\left(1/2 - \xi ' \right)\right ] & { \rm if } & \xi < \xi ' \\
\sinh\left[y\left(1/2 - \xi \right)\right]\ , \sinh\left[y\left(1/2 + \xi ' \right)\right]\ , & { \rm if } & \xi > \xi '
\,,\\ \end{array } \right.\]]where @xmath83 and @xmath84 is the fourier transform of the radial density , which is taken to be @xmath85 $ ] . using the analytical expression for @xmath60 , instead of direct solution of the 3d poisson equation ,
highly facilitates computations .
@xmath86 is the exchange energy operator , @xmath87 2 where @xmath88 f(k_r)\,.\]]for the reservoir configuration that we consider here , the `` external potential '' is a linear function of @xmath89 : @xmath90 .
the energy @xmath91 in eq .
( [ schrodinger ] ) is expressed in terms of the wave vector @xmath66 and the applied voltage @xmath70 , @xmath92where we assume that the energy reference is fixed at @xmath93 .
in addition , we require continuity of the wave functions determined by eqs .
( [ schrodinger ] ) and ( [ reservoirwavefunction ] ) and their derivatives at the reservoir - wire interfaces @xmath94 .
the distribution functions @xmath81 in the reservoirs are taken in the form of the fermi functions with the temperature @xmath95 considering the fact that the fermi level in the right reservoir is shifted down by @xmath96 with respect to the left one . inside the qw
no distribution functions are assigned .
the electron distribution over the energy is determined by the electron flows from the left and right reservoirs and the interaction processes inside the qw .
the external voltage produces a variation of the electron flows , as a consequence of which the electron density is changed . ultimately , this results in the self - consistent variation of both the electron states and their occupation for both the left and right moving particles .
the wave functions that we consider in this work are characterized by a continuous quantum number @xmath66 .
@xcite hence , @xmath97 should be considered as a function of two variables @xmath89 and @xmath66 .
equation ( [ schrodinger ] ) is an integro - differential equation with respect to the variable @xmath89 and an integral equation with respect to the variable @xmath66 .
we develop a numerical scheme for the solution of this equation on a grid spanning the two variables .
the computation method is described in the appendix .
it is worth noting that in the case where the voltage is applied , the wave functions are found without using any expansion in terms of the undisturbed wave functions .
the numerical computations were performed using the 32-processor computer system parsytec cc .
first , we consider the equilibrium state that appears in the absence of an applied voltage @xmath98 . in order to realize the three cases ( @xmath99 , and @xmath100 ) described in sec .
[ contact - potential ] , we vary the density of the positive background charge @xmath78 . in doing this
it is convenient to compare @xmath78 with the characteristic density @xmath101 where @xmath102 is the equilibrium level of the chemical potential in the system .
this quantity has a simple physical meaning in the case where the exchange and correlation interaction is absent .
it is the background density that determines which of the three cases is realized in the hartree case . if @xmath103 , electrons are not redistributed between the qw and the leads in the equilibration process , if @xmath104 , electrons flow from the qw to the reservoirs , and if @xmath105 , electrons are transferred from the reservoir to the qw . of course , turning on the exchange and correlation interaction shifts the value of the background density at which electrons are not redistributed .
nevertheless , as a reference , the value @xmath106 remains convenient .
the electron density distribution @xmath79 for the three cases is illustrated in figs .
[ n - b15 ] , [ n - b05 ] , and [ n - b10 ] . shown here
are also the graphs of the potential energy @xmath107 that includes the confinement energy and the hartree energy , @xmath108 but not the exchange energy . the exchange energy is not included in @xmath107 because it is a functional of @xmath97 rather than a direct function of @xmath89 .
it is useful to note that @xmath76 essentially coincides with the electric potential .
let us consider first the case where the exchange interaction is not taken into account .
it is illustrated by the dashed lines in figs .
[ n - b15 ] , [ n - b05 ] , [ n - b10 ] . in this case
@xmath107 gives the full single - particle potential shape in the qw .
if @xmath104 ( this corresponds to @xmath1 ) , fig .
[ n - b15 ] shows that the potential shape lies below the @xmath109 energy and hence the interaction energy is negative .
this means that a positive charge is accumulated in the qw .
it is responsible for the appearance of a potential well .
the case @xmath2 is realized when @xmath105 . the electron density distribution and the potential shape are shown in fig .
[ n - b05 ] .
the interaction energy is seen to be positive and the potential shape in the qw lies above @xmath109 .
this means that a negative charge is accumulated in the qw .
it produces a potential barrier that hinders electrons in passing through the qw . if @xmath103 , the electron density is essentially not redistributed between the qw and the reservoirs
however , friedel oscillations of the electron density appear near the contacts , fig .
[ n - b10 ] .
the friedel oscillations are also observed if there is carrier transfer , but they are superimposed on the much stronger variation of the potential due to the charging of the qw .
let us now consider the exchange interaction effect .
the electron density distribution and the potential shape , calculated by taking into account the exchange interaction , are shown in figs .
[ n - b15 ] , [ n - b05 ] , [ n - b10 ] as solid lines . the exchange interaction is seen to result in an increase of the electron density .
this is a consequence of the fact mentioned in sec .
[ contact - potential ] that the exchange interaction decreases the chemical potential in a qw .
that is why more electrons come into the qw when the equilibrium state is established .
correspondingly , the negative charge in the qw increases , which results in the growth of the energy @xmath107
. however , this does not yet mean that the electron states with energy lower than @xmath107 are necessarily states decaying in the qw , since the exchange interaction lowers the effective barrier between the qw and the reservoir . since it would be incorrect to consider the exchange energy as a function of @xmath89
, we calculate an average value of the exchange energy per particle incident on the qw with the energy @xmath110 from the left reservoir , @xmath111(here @xmath112 denotes averaging over the qw length ) . the average exchange energy @xmath113 is shown in fig .
[ ex - eh ] as a function of @xmath110 .
the average hartree energy @xmath114 is also given in this figure .
it is seen that @xmath113 exceeds @xmath114 for all energies .
hence the joint effect of the exchange interaction and the hartree interaction consists in an effective lowering of the barrier , so that electrons with energy below the confinement energy @xmath109 can transit through the qw without decay .
another effect produced by the exchange interaction is also seen from figs .
[ n - b15 ] , [ n - b05 ] , [ n - b10 ] .
the exchange interaction strongly enhances the friedel oscillation amplitude .
this result agrees qualitatively with the analytical calculation of the interaction effect on the transmission through a barrier in 1d systems .
@xcite in our case the friedel oscillations are generated at the contacts of the qw with the reservoirs .
their amplitude in the potential energy is quite pronounced but smaller than the fermi energy @xmath115 in the qw . for the discussion which follows it is important to remark that the friedel oscillations are superimposed on the smooth variation of the potential produced by the charge accumulated in the qw .
even if the exchange interaction is fully taken into account , this smooth component has an amplitude that is larger than that of the friedel oscillations .
due to the smooth variation of the potential the qw becomes nonuniform .
the model that we have developed above allows us to find the electric current arising when an external voltage is applied .
the current is calculated as the sum of the partial currents of the states @xmath116 , taking into account their occupation .
the linear regime is realized when @xmath117 . in this case
we have obtained the dc conductance as a function of the chemical potential @xmath102 .
the results of these calculations for zero temperature and for several densities of the background charge @xmath78 are given in fig .
[ l - cond ] .
the conductance oscillations with varying chemical potential are a consequence of the nonadiabatic reservoir - wire interface .
the rapid variation of the potential at this interface leads to backscattering and , if the electron wave is coherent over the entire wire length , to resonances .
the oscillations have the same origin as the resonances observed in over the barrier transmission of noninteracting particles . a similar effect
also appears in the transport of noninteracting electrons through a narrow , ballistic , nonadiabatic constriction in a 2d electron gas .
@xcite our calculations show that the _ e - e _ interaction changes the effective potential barrier that electrons have to overcome in passing from one reservoir to another .
first , we discuss the results obtained within the hartree approximation , when exchange interaction is neglected .
these results are represented by the lines 1 - 3 in fig .
[ l - cond ] for several different background densities @xmath78 . with increasing @xmath78
the oscillations become more frequent , which means that the effective wave number of the electrons is increased .
exactly the same behavior is demonstrated in figs .
[ n - b15 ] , [ n - b05 ] , [ n - b10 ] , namely , with increasing @xmath78 the potential in the qw is shifted downward causing the kinetic energy to increase .
a similar effect occurs when the exchange interaction is turned on .
it is demonstrated in fig .
[ l - cond ] by curve 3 ( obtained by ignoring the exchange interaction ) and by line 4 ( obtained by including exchange interaction ) , the background charge being the same in both cases .
the exchange interaction is seen to make the conductance oscillations more frequent .
the reason for this effect is that the exchange interaction results in an effective lowering of the potential energy of the electrons and consequently in an increase of their kinetic energy . in order to assess the exchange interaction effect on the effective potential , it is instructive to see how the exchange interaction , affects the spectral density of electrons , i. e. , @xmath118 integrated over the qw length
this is illustrated in fig .
the exchange interaction allows the electrons with energy below the confinement energy @xmath109 to pass through the qw .
it is interesting to note that , despite the fact that friedel oscillations are present in the qw , no noticeable suppression of the conductance is observed .
the effect of conductance suppression by a periodic potential associated with friedel oscillations was considered for infinite 1d systems with a @xmath119 potential in refs and .
this phenomenon is connected with the fact that a periodic component of the potential suppresses the transmission of the electrons with energy near the fermi level across the qw ( a gap appears at the fermi level ) .
the absence of this effect in our system is a consequence of two facts .
first , the qw has a finite length .
second ( and no less essential ) , the qw becomes inhomogeneous owing to the electron density redistribution between the qw and the leads . as a consequence the kinetic energy at the chemical potential level and the friedel oscillation period become dependent on the position in the qw
this is why the resonant interaction of electrons at the chemical potential level with the friedel oscillations is destroyed and the electron passage is not suppressed .
a significant redistribution of the electron density between the qw and the reservoirs occurs under far from equilibrium conditions when the applied voltage exceeds the fermi energy .
electrons are injected from the left reservoir ( cathode ) while the electrons entering the qw from the positive reservoir ( anode ) are scattered back inside the qw . as a consequence ,
the electron density decreases in the qw ( roughly speaking to one - half of the equilibrium density ) though the positive background charge is unchanged . because the positive charge is dominant , a potential well appears in the qw , with the potential shape being distorted by the external potential , as illustrated in fig .
[ diagram ] . therefore the kinetic and potential energies are greatly changed .
the change in the potential energy produces variations in the wave functions ( including even a possibility for resonant states to appear ) and the electron density distribution . in this way
feedback arises between the electron density and the potential in the qw , which is an important mechanism in nonlinear transport .
it is that mechanism which is realized in the model proposed . a complete numerical analysis of the nonlinear transport properties in the wide range of applied voltages within the hartree - fock approximation meets some difficulties caused by the long computation time .
in this paper we restrict our consideration to the hartree approximation , which is reasonable at high enough voltage because the exchange energy decreases when the kinetic energy of electrons is increased .
@xcite the calculations were carried out using the method of pseudotime evolution to the steady solution @xcite described in the appendix .
it turns out that in some range of applied voltage an instability of the evolution process appears .
the instability origin is not connected with the computation process but is caused by real behavior of the system .
the mechanism of the instability is as follows . when the applied voltage is high enough ( compared to the fermi energy ) , the electron flow injected from the negatively charged reservoir is the only flow in the qw . let a velocity fluctuation appear in some portion of the wire . to be definite ,
let us assume that the velocity is increased above its stationary value .
since the total electron flow is limited by the contact , it is not disturbed by this fluctuation .
hence , the continuity of the current requires that the electron density decreases .
this leads to a growth of the positive ( net ) charge , because electrons can not completely neutralize the background charge .
the excess positive charge causes the potential energy of the electrons to decrease . under the condition of ballistic transport ,
this results in a new increase of the velocity , and so on until some nonlinear process stabilizes this instability . in our model
this is achieved by a redistribution of the overall electron density and a reshaping of the potential distribution in the qw .
in such a way the potential shape is switched from one state to the other under the condition that both states are characterized by the same potential difference across the qw ends . in order to describe the transition from one shape to the other as a continuous process it is necessarily to characterize the nonequilibrium state of the system by a parameter other than the applied voltage .
this parameter should distinguish states with different potential shapes and the same potential difference across the ends .
as such a parameter , the mean kinetic energy @xmath120 of electrons in the qw can be used , @xmath121 where @xmath122 is the kinetic energy operator .
this conclusion is similar to what is known in the theory of hot electron instabilities in semiconductors . the electron heating by the electric field results in s- and n - shaped current - voltage characteristics . under these conditions ,
the electron temperature uniquely determines the state of the system .
@xcite in our case the kinetic energy of the electrons is a direct analog of the electron temperature .
we have developed the algorithm that allows one to solve our problem in the case where the mean kinetic energy is fixed rather than the applied voltage .
this algorithm is described in the appendix . in this computation scheme
all quantities ( including @xmath70 ) are determined by @xmath120 .
we have found that this algorithm gives stable results and the nonequilibrium state of the system under investigation is uniquely determined by @xmath120 . in particular , the dependence of @xmath120 on the applied voltage is shown in fig .
[ multi-1 ] .
the kinetic energy is seen to have several values for a given voltage @xmath70 , while @xmath70 is uniquely defined by @xmath120 .
correspondingly , several nonequilibrium states , with different distributions of the electron density and the potential , are possible at a given voltage .
it is obvious that the states are not all stable with respect to time - dependent fluctuations .
the multistability , and in particular bistability , phenomenon described above may be useful for understanding the negative differential conductance observed in quantum wires .
in this paper a qw of a finite length coupled to reservoirs is considered as a unified system .
we have found that the electron density is substantially redistributed between the wire and the electron reservoirs when this system is formed . as a consequence of this process , a qw can acquire a net charge .
the charging of the wire is caused by the chemical potential difference between the wire and the leads which exists if these two subsystems are decoupled .
this phenomenon is similar to the contact potential difference in a classical multiconductor system .
the structures of the charge density and the electric potential distributions differ according to the chemical potential difference : ( i ) a positive charge is accumulated in the wire and a potential well is developed there ; ( ii ) the wire is charged negatively giving rise to a potential barrier ; ( iii ) the wire remains uncharged as a whole . in all cases ,
friedel oscillations are present which are generated at the nonadiabatic contacts of the qw with reservoirs .
they are superimposed on the relatively smooth profile of the potential produced by the charge accumulated in the wire .
this smooth potential has a large amplitude .
the friedel oscillation amplitude is strongly enhanced if the exchange interaction is included .
variation of the electron density in a qw due to electron redistribution between the wire and the reservoirs produces a significant effect on dc conductance .
this effect is connected with the change of the kinetic energy of the electrons due to two factors : ( i ) the variation of the accumulated charge and the potential variation associated with this charge ; ( ii ) the variation of the exchange energy .
the exchange energy substantially lowers the effective potential barrier that electrons have to overcome when passing from one reservoir to another .
the friedel oscillation potential does not suppress the conductance because the qw becomes inhomogeneous as a consequence of the electron density redistribution between the qw and the leads .
the electron density variation due to the voltage applied across the leads and the change of the effective potential shape , associated with this variation , in the wire is an important mechanism for nonlinear transport .
the most interesting consequence of the charging effect in the wire is the instability that arises under a high enough applied voltage .
the instability shows itself as a spontaneous increase of the kinetic energy of the injected electrons at a given applied voltage . in a qw with a nonadiabatic connection to the reservoirs , the development of instabilities results in the appearance of multistable states , i.e. , in the existence of several stable states at a given voltage .
we conclude by emphasizing that the charging effect analyzed here is a general phenomenon that might be important in many mesoscopic systems containing electronically different compounds .
examples of current interest are carbon nanotubes , @xcite hybrid normal - superconducting systems , and atomic quantum point contacts .
the present work has been supported by intas ( grant no .
96 - 0721 ) . v.a.s . and s.v.p acknowledge the russian fund for basic research ( grant no . 99 - 02 - 18192 ) , the russian program `` physics of solid - state nanostructures '' ( grant no .
97 - 1054 ) and the russian program `` surface atomic structures '' ( grant no .
is funded by the swiss national science foundation .
the multiprocessor computer system parsytec cc used for the numerical calculations has been acquired by the institute for mathematical modeling of the russian academy of sciences with financial support from the ec esprit program ( grant no .
the problem of finding the wave functions for a qw coupled to reservoirs can be reduced to the solution of eq .
( [ schrodinger ] ) in the inner region @xmath123 and matching @xmath124 and its derivative at the boundaries @xmath125 to the wave functions ( [ reservoirwavefunction ] ) in the outer regions . combining the matching equations for @xmath124 and @xmath126
, one can exclude the coefficients @xmath127 and @xmath128 to get finally the following nonuniform boundary conditions for @xmath124 : @xmath129 where @xmath66 and @xmath130 are defined in eq .
( [ reservoirwavefunction ] ) the wave functions @xmath116 are considered as functions of two continuous variables : the space coordinate @xmath89 and the energy @xmath131 . thus @xmath132 , where @xmath89 and @xmath133 are varied respectively in the regions @xmath123 and @xmath134 , with the upper boundary @xmath135 being equal to @xmath136 ( @xmath137 is the bolzmann constant , @xmath95 is the temperature ) . in this region
the uniform grid @xmath138 is constructed .
the equation for @xmath139 on the grid is obtained from eq .
( [ schrodinger ] ) with use of the integro - interpolative method for the node presentation on @xmath140 and the trapezium formula when calculating the integrals on @xmath66 for the nodes on @xmath141 .
this results in a set of nonlinear finite - difference equations that can be symbolically presented in the form @xmath142 \psi } = { \rm \bf f}\;,\ ] ] where @xmath143 is the wave function vector to be found and @xmath144}$ ] is the nonlinear operator .
the matrix equation ( [ m_psi ] ) is inhomogeneous as a consequence of the boundary conditions ( [ boundary_cond ] ) .
the equations ( [ m_psi ] ) are solved by the iteration method . however , the commonly used successive approximation scheme ( @xmath145 { \bf \psi}^{(l+1)}={\rm \bf f}$ ] , with @xmath146 being the iteration number ) turns out to be badly convergent .
we use the method of pseudotime evolution to the steady solution .
@xcite more specifically , we use the two - layer iteration scheme of this method . in this scheme
the approximating matrix @xmath147 is introduced , which is calculated with the use of the iteration process @xmath148-{\rm\bf \hat{m}}^{(l)},\ ; l = 0 , 1 , 2 , \dots\,,\ ] ] where @xmath149 is a pseudotime parameter .
the choice of @xmath149 allows one to attain the best convergence of the iteration process . as the starting value of @xmath150 we use
@xmath151 $ ] , i.e. , the @xmath152 matrix for noninteracting electrons . during the iteration process
, @xmath153 is calculated with the use of the equation @xmath154 the pseudotime @xmath149 is determined by the @xmath155 operator spectrum .
the optimal convergence is attained when @xmath156 where @xmath157 and @xmath158 are the lowest and highest eigenvalues of @xmath155 .
the iteration process is ended when the following condition is fulfilled @xmath159 where @xmath160 is an element of the @xmath161 matrix . in the present paper , @xmath119
was chosen to be @xmath162 .
the above method is successful when the system under investigation has a unique solution .
however , at some fixed values of the applied voltage the computation shows an instability .
in the course of the pseudotime evolution process the calculated quantities ( such as the potential , the electron density , the kinetic energy ) are randomly switched between several values .
this is connected with the fact that the state of the system is not uniquely determined by the calculation scheme where the applied voltage is fixed .
a unique description of the system is achieved by using the mean kinetic energy @xmath120 of the electrons as the parameter that defines the nonequilibrium state of the system .
we have developed computation algorithm that allows one to vary @xmath120 continuously , in other words , we solve the problem using @xmath120 as the fixed parameter instead of the applied voltage . an essential question appearing in this algorithm
is how the applied voltage @xmath70 should be defined when @xmath120 is given .
the equation defining @xmath70 is obtained from eq .
( [ schrodinger ] ) . multiplying this equation by @xmath163 , integrating it over @xmath66 and over @xmath89 , and summing over @xmath164
one gets an equation of the following form : @xmath165 = b[{\bf \psi}]\,,\ ] ] where @xmath166 $ ] and @xmath167 $ ] are functionals of the electron wave functions . solving this equation with respect to @xmath70 one gets @xmath70 as a functional of @xmath143 , with @xmath120 being a parameter , @xmath168\;.\ ] ] when solving the problem with @xmath120 as a parameter , eq .
( [ v_a - psi ] ) should be taken into account together with eq .
( [ m - evolution ] ) .
this system of equations is solved using the above pseudotime evolution method and two - layer iteration scheme .
the set @xmath169 approximating @xmath70 is defined as @xmath170 - v_a^{(l)}\;,\quad l = 0 , 1 , 2 , \dots\;.\ ] ] as the starting value of the @xmath169 set , the arbitrary value of @xmath70 in the stability region close to the instability threshold can be used . in this generalized procedure
the pseudotime @xmath149 is chosen taking into account the spectral properties of the total matrix @xmath171 .
i. maximov , q. wang , m. graczyk , p. omling , l. samuelson , w. seifert , and i. shorubalko , in _ proceedings of the 7th international symposium `` nanostructures : physics and technology , '' st .
petersburg , russia _ ( ioffe institute , st petersburg , 1999 ) p.458 - 459 .
novoselov , yu.v .
dubrovskii , v.a .
sablikov , d.yu .
ivanov , e.a .
vdovin , yu.n .
khanin , v.a .
tulin , d. esteve , and s. beaumont , in _ proceedings of the 7th international symposium `` nanostructures : physics and technology , ''
st petersburg , russia _ ( ioffe institute , st .
petersburg , 1999 ) p.77 - 80 .
our analysis shows that localized states may appear far from equilibrium , if the background charge density is much higher than @xmath172 defined in eq .
( [ n_0 ] ) .
this case should be investigated separately . | we investigate the distribution of the electron density and the potential in a quantum wire coupled to reservoirs , treating this structure as a unified quantum system and taking into account the coulomb interaction of electrons .
the chemical potential difference that exists between a decoupled , isolated quantum wire and the reservoirs gives rise to charge transfer in the coupled system .
we show that the quantum wire can be charged positively or negatively or remain neutral as a whole , depending on such factors as the wire radius and the background charge density in the wire .
the magnitude of the charge and its sign are to a large extent determined by the exchange interaction of the electrons in the wire . using a hartree - fock approach ,
we develop a model of a quantum wire which includes the reservoirs .
this model allows us to find the self - consistent distribution of the electron density and the potential in the wire both at equilibrium and in the presence of transport .
the linear conductance is investigated as a function of the chemical potential .
the nonadiabatic transition from the reservoirs to the wire leads to conductance oscillations caused by multiple scattering of electron waves .
the period of the oscillations depends on the charge acquired by the wire and the exchange energy .
we find that the exchange interaction strongly enhances the friedel oscillations near the contacts .
however , they do not noticeably suppress the conductance because the wire has a finite length and is charged . under far from equilibrium conditions , which appear
when the applied voltage exceeds the fermi energy in the wire , the system becomes unstable with respect to fluctuations of the electric potential and the electron density .
the instability results in the appearance of multistable electron states . 2 |
in the last years , much effort has been paid to experimentally detect the quark - gluon plasma ( qgp ) in heavy - ion collisions . in these experiments , precise theoretical inputs about the nature of the transition and the equation of state ( eos ) of qgp
are indispensable to unambiguously identify the signal of qgp . here
, lattice qcd provides us with the most powerful method to compute thermodynamic properties of qgp from the first principles of qcd . in the quenched approximation of qcd ,
the study of finite temperature qcd is already well matured : computations of @xmath0 and eos have been made on lattices with several temporal lattice sizes @xmath1 , using both the standard plaquette action and various improved actions .
continuum extrapolation to the limit of large @xmath1 has been performed with different actions to confirm good agreement within an accuracy of a few percent@xcite . for the pressure ,
precise comparisons of the results from different calculation methods , i.e. the integral method@xcite and the derivative method@xcite , have also been made@xcite .
the expected absence of the pressure gap at the first order deconfinement transition point has also been demonstrated through a new non - perturbative calculation of anisotropy coefficients@xcite .
the quenched study is now entering a new stage : thermal effects on the pole mass have been studied by the qcdtaro collaboration using an anisotropic lattice@xcite .
several groups are also trying to compute hadron spectral functions at finite - temperature applying the maximum entropy method@xcite .
on the other hand , the study of full qcd at finite temperatures is still at the stage of development .
until recently , eos with dynamical quarks has been computed only for the staggered fermion .
the milc collaboration calculated eos in 2 flavor qcd with the standard staggered fermion on @xmath2 and 6 lattices@xcite .
the bielefeld group computed eos for @xmath3 using an improved staggered fermion at @xmath2@xcite .
this year , the bielefeld group extended the computation to @xmath4 and 3 and studied the @xmath5-dependence@xcite .
a new development of this year is the first calculation of eos for a wilson - type fermion@xcite . using a renormalization group ( rg ) improved gauge action@xcite coupled with 2 flavors of clover - improved wilson quarks@xcite ,
the cp - pacs collaboration performed a systematic study of the pressure and the energy density of qgp . in this report , we concentrate on the development in finite temperature qcd with dynamical quarks , focusing on the topics of eos , and attempt a comparison of the results from different lattice fermions . in section [ sec :
phase ] , we discuss about the phase structure , critical properties around the chiral phase transition of 2 flavor qcd , and the transition temperature .
we then discuss new calculations of eos in section [ sec : eos ] .
a brief summary is given in section [ sec : summary ] .
= 8.0 cm [ fig : phase ] it has been difficult to study eos with wilson - type quarks due to several difficulties : one is the complicated phase structure due to the explicit violation of chiral symmetry and due to the appearance of the parity - flavor broken phase ( aoki phase)@xcite .
therefore , a systematic study surveying a wide range of the parameter space is required .
the second reason is that , when the standard plaquette gauge action and the standard wilson quark action are used , lattice artifacts are large on coarse lattices used in most finite temperature simulations , for example , unexpected strong phase transition is observed at intermediate quark masses in 2 flavor qcd@xcite .
therefore , we have to use improved actions to suppress these lattice artifacts .
the cp - pacs collaboration studied 2 flavor qcd with an rg - improved gauge action combined with a clover - improved wilson quark action@xcite .
the first step is to clarify the phase structure for this combination of actions .
figure [ fig : phase ] is the result for the phase diagram of qcd obtained on an @xmath2 lattice .
the solid line @xmath6 is the location of the chilal limit at zero temperature
. the pion mass decreases as @xmath7 is increased from small @xmath7 , and vanishes on the line @xmath6 .
when @xmath1 is fixed as in most simulations , the temperature @xmath8 becomes higher as @xmath9 is increased .
the dashed line @xmath10 is the pseudo - critical line for the finite temperature phase transition for @xmath2 .
the region to the right of @xmath10 ( larger @xmath9 ) is the high temperature qgp phase , and that to the left of @xmath10 ( smaller @xmath9 ) is the low temperature hadron phase .
the crossing point of the @xmath6 and the @xmath10 lines is the chiral phase transition point@xcite . in the region above the @xmath6 line , a parity - flavor symmetry of the wilson - type quark action
is broken spontaneously@xcite . at zero temperature ,
the boundary of the parity - flavor broken phase ( the @xmath6 line ) is known to form a sharp cusp touching the free massless fermion point @xmath11 at @xmath12 . at finite temperatures ,
the parity - flavor broken phase retracts from the large @xmath9 region .
the boundary of the parity - flavor broken phase at @xmath13 ( the @xmath14 line ) is shown in fig .
[ fig : phase ] for @xmath2 . as shown in this figure , the chiral transition point locates close to the cusp of the parity - flavor broken phase , the difference being at most of @xmath15 .
this is consistent with the picture that the massless pion , i.e. the goldstone boson associated with spontaneous chiral symmetry breaking , appears only in the cold phase .
= 8.0 cm an important step towards clarification of the qcd transition in the real world is the study of the finite temperature chiral phase transition in @xmath4 qcd .
when the chiral transition in @xmath4 qcd is second order , the transition is expected to be universal to that in a 3-dimensional o(4 ) spin model : with the identifications @xmath16 for the magnetization , @xmath17 for the external magnetic field and @xmath18 for the reduced temperature , where @xmath19 is the chiral transition point , we expect the same scaling behavior as the o(4 ) spin model .
therefore , confirmation of the o(4 ) scaling in qcd provides us with a test of the order of the chiral transition .
the o(4 ) scaling was first tested with staggered fermion by karsch and laermann@xcite .
the study was extended to a wider range of the quark mass and lattice sizes by the jlqcd collaboration@xcite and by the bielefeld group@xcite .
however , an agreement of the critical exponents between the o(4 ) spin model and qcd with 2 flavors of staggered fermions was not obtained for @xmath2 . for the case of wilson fermion ,
et al._@xcite investigated the scaling relation ; @xmath20 using the rg gauge action combined with the standard wilson quark action .
they identified the subtracted chiral condensate defined by an axial ward - takahashi identity@xcite @xmath21 as the magnetization of the spin model . here
, the quark mass @xmath22 is defined by an axial vector ward - takahashi identity@xcite , and the tree - level renormalization coefficient @xmath23 was adopted .
they found that the scaling relation ( [ eq : scfn ] ) is well satisfied with the o(4 ) critical exponents and the o(4 ) scaling function .
figure [ fig : o4 ] is the new result for the case of the clover - improved wilson quark action coupled with the rg gauge action , obtained on a @xmath24 lattice@xcite .
the vertical axis is @xmath25 and the horizontal axis is @xmath26 , where @xmath9 and @xmath27 are the critical exponent obtained in the o(4 ) spin model@xcite .
the dashed line is the o(4 ) scaling function@xcite .
they fitted the data to the scaling function adjusting @xmath19 and the scales of two axes .
as seen from fig .
[ fig : o4 ] , qcd data is well described by the o(4 ) scaling ansatz also for this combination of actions .
the best fit gives @xmath28 for the chiral transition point , with @xmath29 .
this result suggests that the chiral transition is of second order for @xmath4 .
this year , the milc collaboration published their final results of a study of scaling properties in qcd with standard staggered quarks@xcite .
they found that the qcd data are inconsistent with the o(4 ) scaling , in accord with the previous results for the critical exponents@xcite . at the same time , a comparison of the results for @xmath212 suggests that the discrepancy with the o(4 ) scaling function may be removed when @xmath1 is sufficiently large .
a different approach to clarify the origin of the unexpected results was proposed by kogut , laga and sinclair@xcite .
they carried out a simulation at precisely zero quark mass by adding the additional , irrelevant four - fermi interaction to qcd . from their results obtained on @xmath2 and 6 lattices
, they also suggested importance of simulations at larger @xmath1 .
for staggered quarks , a comparison with an o(2 ) spin model is also important because the symmetry of the staggered quark action is o(2 ) at finite lattice spacings .
calculated the scaling function for the o(2 ) spin model@xcite .
they found that the o(2 ) scaling function is quite similar to the o(4 ) scaling function .
this implies that the o(2 ) scaling also can not explain the behavior of the staggered quarks at small @xmath1 .
[ cols="^,^,^,^,^,^,^",options="header " , ] = 7.6 cm = 7.5 cm [ fig : tcks ] in fig .
[ fig : tcmrho ] , we summarize the results of the pseudo - critical transition temperature @xmath30 in @xmath4 qcd with various actions ; plaquette gauge combined with the standard staggered and wilson quark@xcite , symanzik gauge with p4-improved staggered quark@xcite , plaquette gauge with clover quark@xcite , symanik gauge with clover quark@xcite , rg gauge with clover quark@xcite , and plaquette and rg gauge with domain - wall quark@xcite .
the scale is fixed by the vector meson mass @xmath31 at the simulation point .
filled and open symbols are respectively the results on @xmath2 and 6 lattices .
diamonds and stars are new ; the diamonds from the p4-improved staggered quark by the bielefeld group@xcite , and the stars from the rg gauge with clover quark by the cp - pacs collaboration@xcite .
we note that the new results are slightly larger than the previous results from the standard staggered quarks presented by circles .
the bielefeld group and the cp - pacs collaboration also attempted chiral extrapolations .
the results agree with each other within errors in the chiral limit , giving @xmath32 mev for the chiral transition temperature .
the bielefeld group also investigated the @xmath5-dependence of @xmath30 on a @xmath24 lattice@xcite .
figure [ fig : tcks ] shows their results for 2 ( square ) and 3 ( triangle ) flavors of degenerate staggered quarks , normalized by the string tension . from this figure , the dependence on @xmath5 seems to be small for @xmath43 .
table [ tab : tc ] is a summary of the transition temperature in the chiral limit .
the quenched results are extrapolated to the continuum limit . however , those for full qcd , @xmath33180 mev , are the results for @xmath2 , and confirmation at lager @xmath1 is required .
= 8.0 cm = 8.0 cm
the equation of state ( eos ) is one of the most basic information in a phenomenological study of qgp . in this section ,
we discuss the latest results for eos from full qcd . among various methods to compute the pressure , the integral method@xcite is commonly used in recent investigations .
this method is based on the equation , @xmath34 , valid for large homogeneous systems , where @xmath35 is the free energy density . because the derivatives of the partition function can be expressed by expectation values of operators , which are computable by a monte - carlo simulation
, we obtain the pressure by integrating this expectation value in the parameter space . for the case of the wilson quark
, we have @xmath36 with @xmath37 on the integration path .
the starting point of the integration path should be chosen such that @xmath38 there .
the cp - pacs collaboration carried out a systematic calculation of eos in @xmath4 full qcd , using the rg gauge and the clover quark actions on @xmath24 and @xmath39 lattices@xcite .
simulations at @xmath40 for subtraction and also to fix the scale and the line of constant physics , were made on a @xmath41 lattice .
the derivatives of the action in @xmath9 and @xmath7 were computed using a u(1 ) noise method .
they found that the derivative in the @xmath7-direction gives a smaller error in the final pressure .
therefore , they chose the integral path in the @xmath7-direction .
figure [ fig : prs - k ] ( left ) is the @xmath7-dependence of @xmath42 for various @xmath9 . in order to check the reliability of this method
, they also show the data as a function of @xmath9 in fig .
[ fig : prs - k ] ( right ) .
the short lines on the symbols are the slope of the pressure in terms of @xmath9 , which are measured directly from @xmath43 .
we see that the slopes well describe the pressure curves which are computed from an independent measurement of @xmath44 .
this means that the value of the pressure is independent of the choice of the integration path in the parameter space .
= 7.8 cm the data in this form is not yet useful for phenomenological applications . we wish to calculate the temperature dependence for each physical system , i.e. , on a line of constant physics .
the cp - pacs collaboration has chosen the condition @xmath45 constant , determined at @xmath40 , to define the lines of constant physics .
the relation between the simulation parameters and the physical parameters is then given by fig .
[ fig : pirhoconst ] .
the solid lines are the lines of constant physics .
the dashed lines are @xmath46 constant lines , where @xmath30 is the pseudo - critical temperature on the same line of constant physics and the scale was fixed by @xmath31 .
the dashed line @xmath47 itself was determined by fitting the pseudo - critical transition points ( diamonds ) obtained on an @xmath2 lattice .
interpolating the data in fig .
[ fig : prs - k ] yields fig .
[ fig : prs - ttc ] in which the pressure is given as a function of temperature , for each line of constant physics . filled ( open ) symbols are the results of @xmath2 ( 6 )
. different shapes of the symbol correspond to different values of @xmath48 , i.e. , different quark masses .
this figure shows that the pressure is almost independent of the quark mass in a wide range of @xmath49 . on the other hand ,
the @xmath1-dependence is sizeable : there exist large lattice artifacts in the @xmath2 results .
= 8.0 cm [ fig : prs - ttc ] = 8.0 cm [ fig : eng - ttc ] the energy density is obtained by combining the results for the pressure and the results for the quantity ; @xmath50 . \label{e3p}\end{aligned}\ ] ] here , the beta functions , @xmath51 and @xmath52 , can be determined by measuring the differentials of @xmath9 and @xmath7 in terms of the @xmath53 , on each @xmath54 constant line .
figure [ fig : eng - ttc ] is the final results for the energy density as a function of the temperature .
they indicate that the quark mass dependence for the energy density is also small .
previous results from the staggered quark showed a peak just above @xmath0 on an @xmath2 lattice@xcite .
we do not see such a peak in fig .
[ fig : eng - ttc ] from the improved wilson quark .
also the peak disappears at @xmath55 for the staggered fermion , thus , the peak seems to be a lattice artifact .
we find that the magnitude and @xmath56-dependence of eos on @xmath55 lattices are quite similar between improved wilson and staggered quarks .
the quark mass dependence was also investigated with the staggered quark ; similar to the case of the improved wilson quark , the difference between @xmath57 and @xmath58 is found to be small within errors@xcite .
short lines on the right axes of figs .
[ fig : prs - ttc ] and [ fig : eng - ttc ] are the stefan - boltzmann ( sb ) values at @xmath2 , 6 and in the continuum limit .
both @xmath2 and 6 results do not indicate a clear approach towards these values in the high temperature limit . a clear deviation from the sb limits at finite @xmath1
was also observed in a quenched study of eos with the rg - improved action@xcite .
further study is necessary to clarify this phenomenon .
however , quenched studies show that eos extrapolated to the continuum limit @xmath59 is consistent with the continuum sb value in the high temperature limit . from figs . [
fig : prs - ttc ] and [ fig : eng - ttc ] , we note that the results for @xmath55 are already close to the continuum sb limit
. an accurate continuum extrapolation may be possible already with data at @xmath60 or 10 .
= 7.4 cm [ fig : prsbie ] = 7.4 cm [ fig : engbie ] so far , we have discussed simulations with 2 dynamical quarks , which are identified as light up and down quarks . in this subsection , we review results on the effect of the strange quark .
the bielefeld group compared eos for @xmath61 and @xmath62 on a @xmath24 lattice , using the p4-improved staggered quark action with the symanzik improved gauge action@xcite .
@xmath1-dependence is known to be small for this action in the stefan - boltzmann limit .
they computed eos at @xmath63 for @xmath4 and 3 , and at @xmath64 and @xmath65 for @xmath66 .
figure [ fig : prsbie ] shows the results of pressure .
a clear @xmath5-dependence is visible ; the pressure becomes larger as @xmath5 is increased .
clear @xmath5-dependence was observed also in the energy density .
they also observed peaks of the energy density near @xmath30 , similar to the result of the milc collaboration at @xmath2 .
they attempted to estimate the energy density in the chiral limit by subtracting the contribution of @xmath67 in the calculation of the @xmath68 , since this term vanishes when the quark mass is zero .
quark mass dependence was assumed to be small in the pressure .
figure [ fig : engbie ] shows the resulting estimate of @xmath69 in the chiral limit .
the peak disappeared in this figure .
therefore , in contrast to the case of fig .
[ fig : eng - ttc ] , noticeable quark mass dependence was suggested for the energy density .
further study is necessary to clarify this point .
we reviewed progress of the numerical study of finite - temperature qcd with dynamical quarks in the past year , focusing on the topics of eos .
for the transition temperature , recent results for @xmath4 qcd , computed with improved staggered and improved wilson fermion actions on @xmath2 lattices , agree well with each other . in the chiral limit
, the transition temperature is about 170 mev .
this value is higher than the previous results from the standard staggered fermion at @xmath2 and 6 .
confirmation of this result requires further study at lager @xmath1 .
the first calculation of eos with wilson - type fermion was performed for @xmath4 at @xmath2 and 6 , using an improved wilson fermion .
the results of eos for @xmath55 are quite similar between the improved wilson and the staggered fermions , and are close to the continuum stephan - boltzmann value at high temperatures .
this may suggest that a small @xmath60 or 10 is already sufficient for a precise calculation of eos . on the other hand ,
a clear @xmath5-dependence was reported for eos .
this implies the importance of including dynamical strange quark in the simulations . in summary
, we have obtained two keywords for the next step of full qcd thermodynamics : larger @xmath1 with improved actions towards the continuum limit , and simulations with the dynamical strange quark .
studies in these directions are now starting .
i would like to thank a. ukawa and k. kanaya for helpful suggestion and comments on the manuscript .
i also thank f. karsch for communication on their results and useful discussions .
this work is supported by jsps . | recent developments in finite - temperature qcd with dynamical quarks are reviewed focusing on the topics of critical temperature , the equation of state , and critical behaviors around the chiral phase transition . |
understanding the nature of the first stars is one of the key questions of modern cosmology . despite extensive observational searches ,
truly pristine stars have so far escaped detection .
currently , the most metal - deficient star known has a total metallicity of @xmath9 @xcite .
the persistent lack of more metal - poor stars seems to confirm the idea that primordial stars were all more massive than @xmath10 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and that their formation was possibly quenched at early times ( e.g. * ? ? ?
* ; * ? ? ?
however , the chemical signatures of these extinct stellar generations could be retained in the photospheres of ancient ( @xmath11 gyr ) low - mass second - generation stars , which formed in pre - enriched environments , @xmath12 , where metals and dust grains dispersed by the first stars enabled efficient gas cooling and fragmentation ( e.g. * ? ? ?
these stellar fossils should be observable in the oldest stellar components of our galaxy and in its ancient dwarf satellites .
high- and medium - resolution spectroscopic studies of galactic halo stars have revealed the existence of a population of carbon - rich stars ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these objects are usually defined to have carbon - to - iron ratio [ c / fe]@xmath13 @xcite , and they can be divided into two main populations : carbon - rich stars that exhibit an excess in heavy elements formed by slow ( or rapid ) neutron capture processes , cemp - s ( cemp - r ) stars , and carbon - rich stars that do not exhibit such an excess , cemp - no stars .
the available data are consistent with the idea that cemp - s stars belong to binary systems @xcite , and have acquired their carbon - excess from an asymptotic giant branch ( agb ) companion ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus , the chemical abundances measured in these stars are not representative of the interstellar medium ( ism ) out of which they formed . on the other hand
, cemp - no stars are _ not _ preferentially associated to binary systems @xcite .
hence there is no observational evidence supporting the idea of mass transfer as the origin of their chemical abundances , which was suggested by some authors ( e.g. * ? ?
* ; * ? ? ?
furthermore , both their _ frequency _ and _ carbon - excess _ increase with decreasing [ fe / h ] ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and 8 out of the 9 halo stars discovered at [ fe / h]@xmath14 are cemp - no stars @xcite .
these findings favor the idea that cemp - no stars are a peculiar stellar population and that their chemical abundances likely reflect their birth environment .
the unusual chemical compositions of the most iron - poor and carbon - rich stars can be successfully matched by models of _ primordial faint sn _ that experienced mixing and fallback , hence releasing small amounts of iron and large amounts of carbon and other light elements ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. a reatively good agreement with observations is also obtained by models of zero- ( or very low- ) metallicity massive `` spinstars '' , which experience mixing and mass loss because of their very high - rotational velocities ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recently , chemical evolution studies have further support the idea of the link between primordial _
faint _ sn and cemp - no stars , showing that the observed fraction of carbon - enhanced to carbon - normal stars at [ fe / h]@xmath4 is successfully reproduced _ if faint sn dominated the early metal - enrichment _ @xcite .
thus , we can work under this simple hypothesis to predict the frequency of pristine carbon - enhanced stars in _ ancient _ and _ metal - poor _ dwarf galaxies .
carbon - enhanced metal - poor stars have been found in a significant fraction in the faintest satellites of the milky way , the so - called ultra - faint dwarf galaxies , with total luminosities @xmath15 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
different groups have proposed these galaxies to be the living relics of the first star - forming minihaloes , which formed in the milky way environment prior the end of reionization ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these theoretical predictions are consistent with recent observations of star formation histories in ultra - faint dwarf galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in particular , by interpreting the observed fe - luminosity relation and metallicity distribution function ( mdf ) of dwarf galaxies in a cosmological context , @xcite predicted these ancient ultra - faint dwarf galaxies to be the best objects to look for the chemical imprints of the first stellar generations .
this picture is supported by the recent discovery of several cemp - no stars at [ fe / h]@xmath4 in segue 1 @xcite , which is one of the faintest ultra - faint dwarfs .
however , carbon - enhanced metal - poor stars seem to be rare in the more luminous `` classical '' dwarf spheroidal ( dsph ) galaxies , @xmath16 , where measurements are available for larger stellar samples .
the deficiency of cemp(-no ) stars is especially mysterious in the sculptor dsph galaxy .
the observed color - magnitude - diagram ( cmd ) and mdf of sculptor are consistent with this galaxy being dominated by ancient stars , @xmath17 gyr old @xcite .
however , no cemp stars have been found among the ten carefully studied stars at [ fe / h]@xmath4 @xcite .
only recently , the first cemp_-no _ star has been discovered at [ fe / h]@xmath3 @xcite . the observed chemical abundance pattern of this star is consistent with the idea that the carbon excess is the result of a _ pristine _ population of faint sn polluting its birth - environment .
however , the lack of cemp - no stars at lower [ fe / h ] is not .
the important question is , are carbon - enhanced _ metal - poor _ stars missing in classical dsph galaxies , or are they hidden ?
we argue they may be hidden . to sustain this thesis , in this paper we use a statistical , data - calibrated cosmological model for the formation of the milky way and its dwarf satellites , to predict the frequency of cemp(-no ) stars in sculptor and other local group dwarf galaxies .
we present a simple , global scenario that self - consistently explains the variation with galaxy luminosity of the observed : i ) carbon - enhanced to carbon - normal star ratios ; ii ) metallicity distribution functions ; and iii ) star formation histories .
we link the properties of local group dwarf galaxies with first stellar generations and early galaxy formation processes , and present model predictions aimed at identifying the hidden cemp - no stars in the classical dsph galaxy sculptor . for consistency with previous works @xcite , along with via lactea ii , aquarius , and clues simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ) , we adopt a lambda cold dark matter ( @xmath18cdm ) cosmology with @xmath19 , @xmath20 , @xmath21 , and @xmath22km / s / mpc .
furthermore , in the overall paper we assume the solar abundance values by @xcite . and [ fe / h ] in _ the stellar halo _ ( squares ) , _ ultra - faint _ dwarf galaxies ( circles , hexagones , pentagones ) and _ classical _ dsph galaxies ( triangles ) . [ c / fe ]
measurements are corrected to account for internal mixing processes ( see text ) .
_ cemp - no stars _ are shown as filled symbols , upper limits with arrows .
stars with [ c / fe]@xmath13 and open symbols are _ cemp - s / r stars_. filled gray squares are cemp stars with no available measurements of r- or s - process elements . colors / symbols identify stars in dwarf galaxies with increasing total luminosity : from red to blue ( see labels and text ) .
references : _ halo stars : _ @xcite . _ segue 1 : _
_ ursa major ii _ and _ coma berenice : _ @xcite .
_ leo iv : _ @xcite .
_ bootes : _ @xcite .
_ draco : _ @xcite . _
sextans : _ @xcite .
_ carina : _ @xcite .
_ sculptor : _
_ fornax : _ @xcite . ]
current observations of cemp stars , [ c / fe]@xmath13 , are shown in fig.1 , where we present a sample of stars from the literature with measured carbon - to - iron ratio , [ c / fe ] , and iron - abundance , [ fe / h ] , in both the galactic halo and nearby dwarf galaxies .
most of these data are based on one dimensional local thermodynamic equilibrium ( 1d / lte ) stellar model analysis .
measurements for _ halo stars _ are from @xcite , who selected among all available data ( see references therein ) a sample of 505 stars with [ fe / h]@xmath23 and [ c / fe ] measurements .
this sample includes dwarf and giant stars ( @xmath24 ) .
@xcite corrected [ c / fe ] to account for the depletion of surface carbon abundance , which is expected to occur on the upper red giant branch ( rgb ) , @xmath25 .
the correction depends on several observed quantities : @xmath26 , [ c / fe ] , [ fe / h ] and also , to a lesser extent , on [ n / fe ] .
hence , we should keep in mind that the corrected values presented in fig .
1 have intrinsic errors , which are estimated by the authors to be always within the @xmath27 uncertainties of the [ c / fe ] measurements , i.e. @xmath28 dex . in more distant dwarf galaxies only the brighter rgb stars are typically observed .
hence , in our selection of dwarf galaxy stars , if the internal mixing was not already accounted for by the authors , we used the online tool by @xcite to self - consistently correct the [ c / fe ] measurements .
when not available from observations we simply assumed [ n / fe]@xmath29 , in agreement with @xcite .
the data shown in fig . 1 represents the largest sample of [ c / fe ] measurements that have been homogeneously corrected to account for internal mixing processes .
+ from fig .
1 we can note that stars at [ fe / h]@xmath30 have only been found in the galactic halo .
moreover , we can see that the frequency of cemp - no stars among these `` hyper - iron - poor stars '' is extremely high : 8 out of 9 stars at [ fe / h]@xmath31 are cemp - no stars .
these objects have peculiar chemical abundance patterns consistent with the yields predicted for primordial faint sn ( e.g. * ? ? ?
. however , we also note that these extreme stars are part of a common , global trend , which involves stars in both the galactic halo and dwarf galaxies .
we see that on average [ c / fe ] is higher in stars with lower [ fe / h ] , and it declines with [ fe / h ] .
the same trend affects the incidence of cemp - no stars with respect to the overall stellar population . at [ fe / h]@xmath32 , the fraction of cemp - no stars in the galactic halo , ( @xmath33 , placco et al .
2014 ) , is consistent with the _ overall _ fraction of cemp - no stars in dwarf galaxies , where 24 stars at [ fe / h]@xmath34 have been observed , and 10 of them are found to be cemp - no stars .
this gives @xmath35 .
however , when we consider individual dwarf galaxies , the fractions are highly variable . the carbon measurements in the least luminous ultra - faint dwarf galaxies , @xmath36 ( shown in fig . 1 with red circles ) ,
are from high - resolution spectroscopic studies ( see caption for references ) .
although less than 20 stars have been observed in these 4 systems , the overall frequency of cemp - no stars is very high , and at [ fe / h]@xmath34 , 5 out of 6 stars are cemp - no stars , @xmath37 . with the only exception of seven stars in bootes ,
one of which is a cemp - no at [ fe / h]@xmath38 @xcite , only low resolution spectroscopic studies are available for dwarf galaxies with luminosities between ultra - faint and classical dsph galaxies , e.g. _ bootes _ , @xmath39 , and _ draco _ , @xmath40 .
thus , slow and rapid n - capture elements have not been measured in these stars , preventing us from distinguishing between cemp - no and cemp - r / s stars . as discussed in sec . 1
, however , there is strong observational evidence that cemp - no stars should be the dominant cemp population at [ fe / h]@xmath4 ( e.g. * ? ? ?
in these dwarf galaxies the fraction of cemp(-no ) stars at [ fe / h]@xmath34 is high : 4 out of 6 stars in bootes ( @xmath41 ) , and @xmath42 out of @xmath43 stars in draco ( @xmath44 ) .
the classical dsph galaxies _ carina _ and _ sextans _ , @xmath45 , have been followed up at high - resolution , but only stars at [ fe / h]@xmath46 have carbon measurements ( see fig . 1 ) . in the _ sculptor _ dsph galaxy , @xmath47 , many carbon measurements are available from both low- @xcite , and high - resolution spectroscopic studies @xcite . in total
, 10 stars have been found at [ fe / h]@xmath34 . after correcting for internal mixing , we find that the star scl@xmath48 observed by @xcite at [ fe / h]@xmath49 can be possibly identified as a cemp - no star , [ c / fe]@xmath50 .
this was not claimed by the authors , who used the luminosity dependent criteria developed by @xcite to identify cemp stars.however , the carbon - excess is small and the errors big , making this classification very uncertain .
thus , the only reliable cemp - no star discovered in sculptor , [ c / fe]@xmath51 , has an unusually high iron - abundance , [ fe / h]@xmath3 @xcite , which makes this star stand out with respect to the general trend in dwarf galaxies ( fig .
no cemp - no stars have been found in fornax , @xmath52 , although carbon measurements are only available for stars at [ fe / h]@xmath53 . for other ultra - faint dwarfs ( e.g. segue 2 , willmann , hercules ) or classical dsph galaxies ( e.g. leot , leoi , leoii , ursa minor ) carbon measurements are not yet available . in conclusion , while @xmath54 in the galactic halo is consistent with the _ overall _ fraction of cemp - no stars at [ fe / h]@xmath34 in dwarf galaxies , the individual milky way companions show that @xmath54 strongly decreases when the luminosity of the galaxies increases .
does the fraction of cemp - no stars depend on galaxy luminosity ? and why does the only cemp - no star observed in a more luminous dsph galaxy
have an unusually high [ fe / h]@xmath3 ?
are these observational findings reconcilable with the idea that the birth environment of cemp - no stars was polluted by primordial faint sn , and thus consistent with what we see in the galactic halo ?
we use the data - constrained cosmological code gamete ( galaxy merger tree and evolution , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) to link the properties of ancient stars in the local group with the early star formation and metal - enrichment processes .
gamete describes the formation of the galaxy and its dwarf satellites in a @xmath18cdm framework , self - consistently accounting for the key physical processes driving the formation and evolution of high - redshift dwarf galaxies : ( i ) the transition from _ massive _ , and hence short - lived population iii ( pop iii ) stars , to _ normal _ population ii ( pop ii ) stars ; ( ii ) the gradual quenching of star formation in dwarf galaxies with increasing total masses due to the enhanced photo - dissociating and photo - ionizing radiation ; ( iii ) the steady metal - enrichment of the milky way environment , or _ galactic medium _ , due to supernovae - driven outflows from star forming haloes .
our model is a statistical tool that enables us to study the most likely assembly and metal - enrichment histories of local group galaxies .
the star formation and chemical evolution of present - day galaxies is traced across cosmic time by exploiting a sample of merging histories of a milky way - size dark matter halo ( @xmath55 ) , reconstructed via a monte carlo algorithm based on the extended press - schechter formalism ( see * ? ? ?
* ; * ? ? ?
* for more details ) .
newly collapsed haloes are assumed to have a gas - to - dark matter mass ratio equal to the baryonic cosmic fraction , @xmath56 and a chemical composition equal to the milky way environment at their formation epoch .
hence they are all primordial , @xmath57 , before the onset of supernovae explosions .
the main and new features of the model are summarized below , including the underlying key assumptions relevant for this work .
we refer the reader to previous papers for more details .
+ * star formation ( sf ) . * the sf is traced along the merger trees by adopting physically motivated hypotheses .
* there exists a minimum halo mass to form stars , @xmath58 , whose evolution accounts for the suppression of sf in progressively _ more massive _ objects due to the increasing photo - dissociating and photo - ionizing radiation @xcite .
when the milky way environment is fully reionized , @xmath59 @xcite , we assume that gas accretion is suppressed in haloes with virial temperatures @xmath60 k. * the sf rate , @xmath61 , which is regulated by the sf efficiency , @xmath62 , depends on the free - fall time , @xmath63 , and mass of cold gas in each galaxy , whose gradual accretion is described by a numerically calibrated infall rate @xcite . * in _ minihaloes _ with @xmath64 k , the sf efficiency is assumed to be reduced as @xmath65^{-1}$ ] to account for the ineffective cooling by molecular hydrogen , h@xmath66 @xcite .
* pop ii stars with masses @xmath67\msun$ ] form according to a larson initial mass function , @xmath68 , if the gas metallicity exceeds the critical value , @xmath69 , which sets the minimal conditions for the formation of the first low - mass stars
. this value can be either @xmath70 , if gas fragmentation is driven by metal - line cooling ( e.g. * ? ? ? * ; * ? ? ?
* ) , or @xmath71 , if it is due to thermal emission by collisionally excited dust grains ( e.g. * ? ? ?
* ; * ? ? ?
following the most recent findings we set @xmath72 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and we explore the case @xmath73 in sec .
* massive pop iii stars form if @xmath74 .
to work under the hypothesis that the early metal - enrichment is dominated by faint sn ( e.g. * ? ?
* ; * ? ? ?
* ) , we adopt the simplest prescription and assume that all pop iii stars have a characteristic stellar mass of @xmath75 and evolve as _ faint supernovae _ ,
which experience mixing and fallback ( e.g. * ? ? ?
* chemical enrichment . *
the contribution to the chemical enrichment by different stellar populations is traced in the code by accounting for the mass and metallicity dependent stellar lifetimes , and by including all chemical elements from c to zn @xcite . for pop iii stars evolving as faint supernovae
we adopt the yields by @xcite for the case , as their yields are rescaled with stellar masses ( e.g. fig.2 by @xcite ) @xmath75 and kinetic explosion energy @xmath76erg ] .
following @xcite we assume that stars with @xmath77 do not contribute to metal enrichment , and we adopt their yields systematically halved in iron @xcite for massive @xmath78 stars that evolve as type ii supernovae ( snii ) , @xmath79erg . for low and intermediate mass asymptotic giant branch ( agb ) stars , @xmath80
, we adopt the yields by @xcite for metallicities @xmath81 and by @xcite for @xmath82 .
all relevant equations describing the chemical enrichment of star - forming galaxies can be found in @xcite .
the contribution of supernovae type ia ( snia ) has been included by adopting the yields and explosion energy ( @xmath83erg ) by @xcite , and the bimodal delay time distribution observationally derived by @xcite . at each time - step and for each star - forming halo of the merger tree
we compute the rate of snia by following @xcite and we fix the normalization constant , @xmath84 , to reproduce the actual rate of snia in the milky way , @xmath85yr@xmath86 @xcite . the chemical evolution of the gas is simultaneously traced in the star - forming haloes and in the surrounding _ galactic medium _ , or milky way environment , by including the effect of sn - driven outflows , which are regulated by the sn wind efficiency , @xmath87 . during sn explosions ,
metals present in the interstellar medium ( ism ) are ejected out of the galaxy at a rate @xmath88 , where @xmath89 is the metallicity of the gas , and @xmath90 the gas ejection rate , which depends on the cumulative sn explosion energy , and on the binding energy of the host halo ( e.g. * ? ? ?
* ; * ? ? ?
heavy elements dispersed into the milky way environment are assumed to be instantaneously mixed in this medium , so its total metallicity , @xmath91 , steadily increases across cosmic times . + , @xmath92,@xmath93 of the dsph candidates in 50 hierarchical merger histories .
observed points are from @xcite .
stars and labels underline individual dwarf galaxies . ] * universality of the free parameters . *
the free parameters of the model are the star formation and sn - wind efficiency ( @xmath62 , @xmath94 ) , the critical metallicity ( @xmath95 ) , and the fraction of stars that can give rise to sn ia ( @xmath84 ) .
these are fixed to simultaneously reproduce the global properties of the milky way at @xmath96 , along with the mdf of galactic halo stars ( e.g. * ? ? ?
* ; * ? ? ?
these free parameters are assumed to be _ the same _ for _ all _ the haloes of the merger tree and in _ all _ hierarchical merger histories of the milky way .
hence , the properties of different dwarf galaxies we will present in the following , _ are not the result of a fine tuning of the free parameters , but a consequence of the cosmological context in which we study the formation and evolution of these small systems . _ + * selection and properties of satellite candidates . *
galaxies that can survive as satellites of the milky way are selected among star - forming haloes of the merger trees that at any given redshift have dark matter masses that correspond to low - sigma density fluctuations , @xmath97 .
this assumption is supported by n - body simulations ( e.g. * ? ? ?
after selection , its subsequent evolution is followed in isolation down to @xmath96 @xcite . in fig .
2 we show the predicted fe - luminosity relation for all selected candidates in 50 assembly histories of the milky way .
we assume @xmath98 to convert the stellar mass into total stellar luminosity , @xmath99 , and compare results with observations .
we can see that model results match very well the observational data , including the nearly flat @xmath100[fe / h]@xmath101 value that is observed in ultra - faint dwarf galaxies . in fig .
2 the colored boxes identify galaxies in different luminosity ranges , using the same color - code as in fig .
[ fig : data ] .
these same colors will be used through the entire paper to distinguish among dwarf galaxies with different luminosities . in our model ,
ultra - faint dwarf galaxies are predicted to be h@xmath66-cooling minihaloes that form at @xmath102 , before photo - dissociation has suppressed gas - cooling and star formation in these small systems @xcite . in contrast , classical dsph galaxies are predicted to form at later times through the merging of smaller progenitors . on average
we find that the more luminous is a galaxy , the more massive is predicted to be its host halo , and the lower its final assembly redshift @xcite .
more luminous galaxies , therefore , keep accreting gas from the galactic medium while it is increasingly polluted with the heavy elements ejected by low - mass star - forming galaxies .
+ we finally underline that in our model we do not account for mass transfer from binary companions .
thus , we can only investigate the incidence of cemp(-no ) stars that formed in carbon - enhanced environments .
we start by discussing the predictions of our cosmological model for the mean properties of dwarf galaxies in different luminosity ranges , which have been selected among all candidates in 50 different milky way realizations ( fig .
[ fig : fel ] ) .
the main results are summarized in fig .
[ fig : mdfs ] where we show , for dwarf galaxies with increasing luminosities ( from top to bottom ) , predictions for the average : i ) fraction of cemp stars @xmath103 ( left ) , ii ) normalized mdfs ( middle ) , and iii ) low - fe tails of the mdfs ( right ) .
the fraction of cemp stars with respect to the total , @xmath104})_i / n_*({\tt[fe / h]})_i$ ] , is calculated by averaging among the total number of dwarf galaxies within a given luminosity range , @xmath105 , where @xmath106})_i$ ] is the fraction of cemp stars in the i - th galaxy , with a given [ fe / h ] value .
similarly , we computed the mean mdfs by averaging among all @xmath107 dwarf galaxies in a given @xmath108 range . for reference ,
segue 1 , coma berenice , and ursa major ii have luminosities consistent with galaxies in the top panel ( red ) , bootes in the third ( yellow ) , and sculptor in the second from bottom ( blue ) .
several interesting features can be noted in fig .
[ fig : mdfs ] : ( i ) independent of galaxy luminosity , we find that @xmath109 at [ fe / h]@xmath110 , and it rapidly decreases towards higher [ fe / h ] , with a steeper decline in more luminous dwarf galaxies .
( ii ) the shape of the normalized mdf dramatically changes with galaxy luminosity . in the least massive ultra - faint dwarf galaxies it is flat and covers a broad [ fe / h ] range .
as we move towards bigger galaxies , the mdf becomes more peaked , the peak is gradually shifted towards higher [ fe / h ] values , and the low fe - tail turns into a smaller fraction of the total .
( iii ) stars with [ fe / h]@xmath4 are predicted to be found in _ all _ dwarf galaxies , but their relative contribution to the overall stellar populations strongly decreases when the luminosity of the galaxy increases , as shown by labels in the panels .
( iv ) on average , the cumulative number of stars at [ fe / h]@xmath111 is roughly of the same order of magnitude , @xmath112 , in all dwarf galaxies with the only exception of the faintest companions .
all these features are simply a result of the hierarchical galaxy formation process and the gradual _ metal enrichment _ and _ reionization _ of the environment out of which these galaxies form , which imprint their physical properties . in our scenario ,
the faintest dwarf galaxies are associated with h@xmath66-cooling minihaloes , which virialize from the milky way environment at @xmath113 , when @xmath114 [ fe / h]@xmath115 .
the chemical enrichment proceeds very smoothly in these small systems , which transform gas into stars very inefficiently ( sec . 3 , see also @xcite ) .
roughly a constant number of stars are formed at different evolutionary phases ( or [ fe / h ] ) , and the gas can reach high [ fe / h ] before being either completely evacuated by the cumulative effect of sne explosions , or photo - heated by the increasing external ionizing radiation @xcite .
this determines the characteristic shape of their mdf that is predicted , and observed , to be flat and to extend over a broad range of [ fe / h ] .
more luminous galaxies from via the assembly of these basics building blocks and more massive progenitors , which form afterwards . in this bottom - up scenario all dwarf galaxies are expected to have , on average , some ultra - faint dwarfs among their parent haloes .
hence , they are predicted to share similar number of stars and mdf tails at the lowest [ fe / h ] .
however , the more luminous a galaxy , the less prominent is the contribution of the few thousand stars formed in these lowest - mass minihaloes into the final mdf .
this is emphasized by the numbers reported in the right panels of fig . [
fig : mdfs ] .
we can see that extremely metal - poor stars represent @xmath116 of the stellar population in ultra - faint dwarfs , while they are only @xmath117 in dsph galaxies that have luminosity similar to sculptor or higher .
for this reason , the low - fe tails seem to gradually `` disappear '' in the normalized mdfs as the luminosity of the galaxies increases . ) as a function of stellar ages ( _ left _ ) and redshift ( _ right _ ) . ]
since more massive dwarf galaxies complete their assembly at lower redshifts , they can continue accreting gas from a milky way environment that is increasingly polluted by heavy elements . the average [ fe / h]@xmath118 value at the formation epoch of the main halo , where the bulk of the stars form , sets the lower limit of [ fe / h ] in the normalized mdf .
thus , on average , the more luminous a galaxy , the more pronounced is the shift of its mdf towards higher [ fe / h ] .
furthermore , galaxies hosted by more massive dark matter haloes have bigger gas reservoir , and they can more efficiently convert gas into stars because of their higher @xmath119 k , which make them ly@xmath120 cooling haloes ( e.g. * ? ? ?
this causes a gradually more pronounced peak in their mdfs .
the fraction of cemp stars in dwarf galaxies with different luminosities is connected to all these previously presented effects .
looking at the right panels of fig .
[ fig : mdfs ] , we can see that @xmath121 follows @xmath122 at [ fe / h]@xmath14 , in all dwarf galaxies .
these _ cemp - no _ stars are predicted to form in ( progenitor ) h@xmath66-cooling minihaloes that have been predominantly polluted by primordial faint sne .
because of the small amount of iron ( @xmath123 ) , and the huge amount of carbon ( @xmath124 ) produced by faint sne , long - lived pop ii stars can already start to form in these small systems@xmath125 .
] when the ism is self - enriched up to [ fe / h]@xmath126 , which corresponds to [ c / fe]@xmath127 , and @xmath128 .
after the formation of pop ii stars , normal snii rapidly start to contribute to the chemical enrichment of both the ism , gradually decreasing the [ c / fe ] level while [ fe / h ] rises , and the milky way environment , increasing @xmath91 up to @xmath95 . in analogy to what is found in the galactic halo @xcite we predict that carbon - normal stars can start to form when [ fe / h]@xmath129 , which corresponds to @xmath130 , and therefore to the disappearance of pop iii stars @xcite .
thus , we predict that @xmath109 at [ fe / h]@xmath131 in all dwarf galaxies , and that this fraction rapidly decreases with increasing [ fe / h ] , because of the larger contribution of snii in both self - enriched and newly formed galaxies .
cemp - no stars populating the mdf at @xmath132 [ fe / h]@xmath23 , predominantly form in environment polluted by primordial faint sn and snii .
thus they are all predicted to form in the very early stages of galaxy evolution .
interestingly , we can see that these cemp - no stars efficiently form in dwarf galaxies like bootes ( yellow ) and draco ( green ) , which are in the `` transition region '' between ultra - faint dwarfs and classical dsph galaxies ( fig .
[ fig : fel ] ) .
this is because these galaxies are associated to the most massive among minihaloes , with @xmath133 k , which host pop iii stars as their least luminous companion , but have higher star formation efficiencies ( sec .
more massive , classical dsph galaxies , have higher probabilities to have among their progenitor haloes pop ii galaxies that virialized from an environment pre - enriched up to [ fe / h]@xmath134 by normal snii @xcite .
carbon - normal stars efficiently form in these galaxies , causing the steeper decline of @xmath103 with [ fe / h ] . in stars with higher iron - abundance ,
[ fe / h]@xmath135 , we find that the carbon - enhancement might also come from low - metallicity ( @xmath136 ) agb stars , which evolve on longer time - scales .
this causes the small increase of the @xmath103 values at [ fe / h]@xmath137 , which is visible in the left panels of fig .
[ fig : mdfs ] .
note that in our model we do not account for binary systems , so the carbon in these stars is not accreted from a companion , but it reflects the chemical composition of the birth environment .
the products of agb stars can be efficiently retained in the ism of low mass galaxies after snii have contributed to the chemical enrichment , and the galaxy can quietly evolve for a short period of time ( e.g. * ? ? ?
the production of s - process elements from @xmath138 agb stars is still unclear ( e.g. * ? ? ?
* ) and we do not account for it in our work .
however , these stars are most likely expected to be cemp - s stars , although their properties are probably different from those formed in binary systems , which directly accrete material from an agb companion . finally , we should emphasize that , although @xmath103 is steeper in more massive dwarfs , at [ fe / h]@xmath4 the results are all consistent within @xmath139 errors .
as already discussed , this is because these different dwarf galaxies are expected to _ share similar ancestors at high - redshift , namely low mass h@xmath66-cooling minihaloes , which represent the birth environments of stars at [ fe / h]@xmath4_. another way to consider these findings is to look at the predicted average star formation histories ( sfhs ) of dwarf galaxies , which are shown in fig .
[ fig : sfh ] as a function of stellar ages , and formation redshift .
we can see that all dwarf galaxies are predicted to host stars @xmath17 gyr old , which is what is observed in the local group ( e.g. * ? ? ?
* ) . at very high redshifts , @xmath140 , we predict dwarf galaxies to share similar star formation rates , @xmath141yr , independent of their total luminosity .
it is during these cosmic epochs that the low [ fe / h ] tails of the mdfs are built up in the star - forming progenitor minihaloes .
ultra - faint dwarf galaxies are predicted to stop forming stars prior the end of reionization , @xmath142 , because of heating by the external uv background that prevents gas cooling , and `` sterilizes '' those systems that still have some leftover gas to fuel star formation @xcite .
this is consistent with observations of sfh in ultra - faint dwarfs ( e.g. * ? ? ?
note that some of these `` sterilized '' minihaloes are predicted to evolve in isolation and survive until @xmath143 , when they can be observed as very - metal - poor damped ly@xmath120 systems @xcite .
these systems might re - start forming stars at later times ( e.g. * ? ? ?
* ; * ? ? ?
* ) and explain observations of hi - rich dwarf galaxies that show recent episodes of star formation ( e.g. leot : @xcite ) .
more luminous galaxies continue to efficiently form stars to much later times , and hence have more extended and complex sfhs . in the right panels of fig .
[ fig : sfh ] we see that @xmath144 is expected to gradually increase towards lower redshifts , reaching the maximum at the main assembling epoch of the host halo , and decreasing afterwards , when the galaxy is assumed to evolve in isolation .
the sfh that has been measured in the sculptor dsph galaxy @xcite , is qualitatively consistent with the one that we predict for dwarf galaxies with similar luminosity ( blue ) : it has a peak at @xmath145 gyr , and then declines with cosmic time , lacking stars @xmath146 gyr old .
in this global picture we can therefore explain the variation , with galaxy luminosity , of both the mdfs and star formation histories of dwarf galaxies .
we now test the results of our cosmological model against available observations for cemp stars in nearby dwarf galaxies .
we focus on three classes of dwarf galaxies that reside in different luminosity ranges , as shown in fig .
[ fig : fel ] : i ) the `` classical '' dsph galaxy sculptor , @xmath147 ; ii ) the most luminous of the ultra - faint dwarf galaxies , bootes , @xmath148 ; and iii ) the least luminous ultra - faint dwarfs , @xmath149 : the combination of seguei , coma berenice , ursa major ii , and leoiv . as pointed out in sec
. 2 , the available data for bootes are mostly blind to cemp sub - classes ( s , r , and no ) .
this represents a possible caveat for our comparison with cemp - no models .
still , most of the cemp stars found in this system have [ fe / h]@xmath4 , which is the typical [ fe / h ] range of cemp - no stars ( e.g. fig . 1 of * ? ? ?
. given the low number of [ c / fe ] measurements in dwarf galaxies ( see fig . [
fig : scl]a , [ fig : dwarfs]a , d ) , we computed the uncertainty of @xmath103 by using the results of @xcite for poisson statistics .
we derived the 1@xmath150 upper ( lower ) confidence limits as @xmath151})$ ] in case of non - detection ( @xmath152 ) , and as @xmath153})$ ] for single detection ( @xmath154})$ ] ) , where @xmath155})$ ] represents the total number of stars with available carbon measurements in different [ fe / h ] ranges .
so the fewer the measurements , the larger the uncertainties .
we used a monte carlo technique to randomly select from the mean theoretical mdfs , a number of stars equal to the total number of [ fe / h ] measurements in different dwarf galaxies , @xmath156 .
we then constructed the average @xmath157 errors on the mdf by iterating this procedure 100 times .
these errors , which are shown in fig .
[ fig : scl]b and fig .
[ fig : dwarfs]b , c together with the dispersion among different dwarf galaxies , provide an estimate of the data - driven uncertainties , i.e. they are the errors we always have to deal with while comparing theoretical models with small numbers of measurements .
the predicted average properties of sculptor - like dwarf galaxies are shown in fig .
[ fig : scl ] .
we note that there is only one detection of a cemp - no star in sculptor at @xmath158[fe / h]@xmath159 @xcite , from which we get @xmath160 . in the other [ fe / h ] bins , with zero detections , @xmath161
is simply set by the number of [ c / fe ] measurements .
so the decline of @xmath161 towards higher [ fe / h ] is simply a consequence of the larger number of stars observed at increasing [ fe / h ] . from fig .
[ fig : scl]a we can see that the theoretical predictions are consistent with @xmath161 . as we already discussed in sec . 4 ,
because of the early chemical enrichment by primordial faint sn , @xmath103 rapidly declines with [ fe / h ] , ranging from unity at [ fe / h]@xmath162 down to values @xmath163 , at [ fe / h]@xmath164 .
carbon - normal stars begin to form when [ fe / h]@xmath165 , and in sculptor they are predicted to be the majority of the stellar population ( @xmath166 ) already at [ fe / h]@xmath167 . bins .
the solid line shows the average value in the model , and the shaded area the @xmath168 dispersion among different sculptor - like dwarf galaxies ( as fig .
[ fig : mdfs ] ) . upside down triangles
are upper limits for the observed fraction of cemp stars based on available data .
labels indicate the number of [ c / fe ] measurements in each [ fe / h ] bin ( references in fig .
[ fig : data ] ) . _
middle panel : _ the observed ( points with errorbars ) and predicted mdf ( histograms with shaded uncertainty ) .
we show the @xmath157 dispersion among sculptor - like dwarf galaxies in 50 milky way possible assembling histories ( light gray ) , and among 100 monte carlo sampling of the average mdf to the number of stars observed ( dark gray ) .
labels indicate the number measurements in each [ fe / h ] bin @xcite . _
bottom panel : _ conditional probability to observe a star at a given [ fe / h ] that it is also carbon - enhanced .
labels indicate the number of [ c / fe]@xmath13 measurements in each [ fe / h ] . ]
this result has two implications : i ) to catch second - generation stars imprinted _ mainly _ by primordial faint sne in sculptor , we should look among stars at [ fe / h]@xmath169 , as in the galactic halo ; ii ) to increase the probability of finding cemp stars in sculptor we should follow up [ fe / h]@xmath4 stars , for which we predict @xmath170 .
our ability to find these cemp stars will remain naturally limited by the absolute number of stars that exist at these low [ fe / h ]
. our model predicts , and this is supported by observations , that stars at [ fe / h]@xmath4 are _ intrinsically _ rare in luminous , sculptor - like galaxies , only representing @xmath171 of the total stellar population ( see fig .
[ fig : mdfs ] ) .
this point is highlighted in fig .
[ fig : scl]b , where we compare the predicted and observed mdfs , that are normalized to the total number of stars
. noticeably , stars at [ fe / h]@xmath4 are almost invisible in the normalized function , which is dominated by more fe - rich stars
. clearly , the paucity of stars at this low [ fe / h ] makes it very challenging to search for cemp - no stars in sculptor , where only the brighter stars at the tip of the rgb can be followed up . between @xmath172[fe / h]@xmath4 , for example , we should roughly double the number of stars _ found _ in sculptor to be able to observe one cemp - no star among them .
thus , we can ask a different question : what is the joint probability to observe a star that has a given [ fe / h ] value , and that is also carbon - enhanced ?
[ fig : scl]c shows this joint probability distribution .
this has been computed by combining the two independent functions : the fraction of cemp stars _ predicted _ by the model , and the _ observed _ mdf normalized to the total number of stars , @xmath173 .
the probability , @xmath174 , is maximal at [ fe / h]@xmath3 , where we predict the highest absolute number of cemp stars .
_ this result naturally explains why the first carbon - enhanced star in the sculptor dsphs has been serendipitously found at such a high [ fe / h]_. in conclusion , the observed mdf , the derived upper limits for @xmath103 , and the iron - abundance of the only cemp star detected in sculptor , are all consistent with the hypothesis that faint sne were the main contributors to primordial metal enrichment .
we now explore the predictions of our model for less luminous ultra - faint dwarf galaxies , where many cemp stars have been found ( fig .
[ fig : data ] ) . in fig .
[ fig : dwarfs ] we show the results for bootes - like dwarf galaxies and for the least luminous ultra - faint dwarfs with @xmath149 . as for sculptor
we select these galaxies on the bases of their total luminosity among all the available candidates predicted by the model ( fig .
[ fig : fel ] ) . by inspecting fig .
[ fig : dwarfs]a and fig .
[ fig : dwarfs]d , we can see that because of the low total number of stars observed in these small systems ( @xmath175 in bootes , and @xmath176 in the faintest ultra - faint dwarf galaxies combined all together ) the upper / lower limits on @xmath177 are extremely high / low , and hence naturally consistent with model predictions . in these ancient systems ,
the fraction of rgb stars with respect to the total is estimated to be @xmath178 .
this implies that in both cases @xmath116 of rgb stars have been already followed - up . in segue
i all rgb stars have already been followed up . ]
we note that in the model , the dispersion among different galaxies is huge in the case of the faintest dwarf galaxies ( fig .
[ fig : dwarfs]d ) , and at [ fe / h]@xmath38 the @xmath179 error is consistent with @xmath180 .
this is because these faint galaxies are associated with the least massive h@xmath66-cooling minihaloes that are able to form stars , and which evolve in isolation .
these systems can either virialize from a primordial birth - environment , @xmath74 , and hence @xmath181 for [ fe / h]@xmath182 because they host primordial faint sn , or from a medium that has been pre - enriched up to @xmath183 by the products of snii . in the latter case they will only host pop ii stars , and hence @xmath184 .
these different formation paths cause the large spread in @xmath103 at [ fe / h]@xmath4 . according to our model
, the large fraction of cemp - no stars observed in the faintest dwarfs at [ fe / h]@xmath4 ( @xmath185 ) suggests that these systems are _ truly pop iii galaxies _ , which have experienced primordial star formation .
however , we clearly need better measurement statistics .
a quest for more data also emerges while comparing the predicted and observed mdfs ( fig .
[ fig : dwarfs]e ) . in the faintest dwarf galaxies ,
the errors induced by the low number of measurements ( 15 stars ) are larger than the dispersion of the model among different dwarf galaxies and milky way merger histories .
in contrast , these errors are equal ( lower ) than the intrinsic model uncertainties in bootes ( sculptor ) , where @xmath175 stars ( @xmath186 ) have been observed . fig .
[ fig : dwarfs]c , f show the joint probability to observe a star in a given [ fe / h ] range , that is also carbon - enhanced .
if we compare this function with the one derived for sculptor ( fig .
[ fig : scl]c ) , we note a clear trend : the fainter the galaxy is , the higher is the _ overall _ probability to observe a star that is also carbon - enhanced , which is the integral of @xmath174 over the [ fe / h ] range .
this explains why it is much easier to find carbon - enhanced stars in these small systems .
moreover , @xmath174 is maximum at lower [ fe / h ] in less luminous dwarf galaxies .
for bootes we find that @xmath174 is maximum at @xmath187[fe / h]@xmath188 , while for the faintest dwarfs at [ fe / h]@xmath189 . in both cases the predicted [ fe / h ] ranges coincide with [ fe / h ] values of observed cemp - no stars in these small systems . finally , we should note that for the faintest dwarf galaxies the model predicts a non - negligible fraction of cemp stars at [ fe / h]@xmath53 ( upper panel ) . as explained in sec . 4 ,
these stars are predicted to form in environment polluted by the products of agb stars with @xmath190 , which can be retained by these small galaxies when sn have already exploded .
interestingly , a cemp - s star at [ fe / h]@xmath137 that does not show evidence for a binary companion has been recently detected in segue 1 @xcite .
so , although more data are required to solidly assess if this star is in a binary system , this observation supports the idea that the physical mechanism we propose may contribute to the formation of cemp - s in these small systems .
the results of our cosmological model show that cemp - no stars , similar to those found in the galactic halo and ultra - faint dwarfs , are not necessarily missing in classical dsph galaxies , such as sculptor , but they could be _ hidden _ among rare [ fe / h]@xmath4 stars , which represent @xmath191 of the overall stellar population .
the possible existence of a cemp - no star at [ fe / h]@xmath49 ( see section 2 ) might be the first confirmation that this is indeed the case .
sculptor - like dsph galaxies , furthermore , are predicted to be among the best systems to find the most iron - poor stars , because their mdfs might extend down to [ fe / h]@xmath192 , if faint sn dominated the early metal enrichment fig .
[ fig : mdfs ] .
the key question is : how much do we need to enlarge the current stellar sample to unveil the lowest - fe tail of the sculptor mdf and catch the most pristine cemp stars ?
[ fig : low_tail ] shows how many stars at [ fe / h]@xmath182 are predicted to emerge in sculptor - like dwarf galaxies by increasing the total number of [ fe / h ] measurements .
[ fig : low_tail]a exhibits results for the actual number of observed stars , @xmath186 stars , which roughly correspond to @xmath193 of all rgb stars in sculptor .
model predictions agree quite well with observations .
[ fig : low_tail]b shows how the low - fe tail of sculptor will appear by following up all stars down to magnitudes @xmath194 , which corresponds to @xmath195 stars in the observed color magnitude diagram @xcite .
we find that in this case we might be able to observe @xmath196 stars with [ fe / h]@xmath197 , the @xmath198 of which are predicted to be cemp - no stars .
these observations can be achieved with present - day instrumentation ( e.g. eso vlt with x - shooter , flames , and uves , or keck / dimos ) to add crucial information to our understanding of the properties of first stars and early galaxy formation processes .
see for example @xcite , who used keck / dimos to obtain medium - resolution spectra ( @xmath199 , @xmath200 ) for @xmath201 stars in the central region of sculptor and derive their c and fe abundances .
future generation of telescopes will allow us to measure [ fe / h ] for stars below the main sequence turn - off in nearby galaxies , like sculptor .
for example , the limit given for mosaic spectroscopy on the eso - elt ( evans et al .
2015 ) is @xmath202 at @xmath203 .
this will dramatically increase the number of stars that can be observed at high spectral resolution in nearby galaxies .
the total number of rgb and ms stars in sculptor with @xmath204 is @xmath205 , and from these we can expect @xmath206 stars at [ fe / h]@xmath197 ( fig .
[ fig : low_tail]c ) , and @xmath207 stars with [ fe / h]@xmath208 , where the fraction of cemp - no stars is expected to be @xmath209 . with these observations , furthermore , we might be able to accurately constrain the critical metallicity value . as already shown in @xcite ( fig .
7 ) , decreasing @xmath95 in the most feasible range , @xmath210 , mainly affects the number of long - lived stars at [ fe / h]@xmath111 .
thus , the differences between @xmath95 models become observable in sculptor when a significant number of stars are followed - up , and the lowest - fe tail starts to emerge ( fig .
@xmath4 that can be observed in sculptor by increasing the sample of [ fe / h ] measurements . from top to bottom we show results for : i ) the current statistics ( @xmath186 rgb stars ) , where the available data are shown as points with poissonian errorbars ( fig . 5 for references ) ;
ii ) stars with @xmath211 ; iii ) stars with @xmath204 ; iv ) all stars .
shaded area show the @xmath212 errors obtained from the monte carlo random selection technique ( see text ) .
solid histograms with shaded area are results for our fiducial model , @xmath213 .
red dashed histograms show the same results for the case @xmath214 . ]
we used a robust , data - constrained merger - tree model for the possible formation histories of the milky way and its dwarf satellites to predict the frequency of cemp - no stars in nearby dwarf galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we have shown that the model can successfully explain the variation of the _ average _ mdfs and sfhs observed in dwarf galaxies with increasing luminosities ( e.g. * ? ? ?
* ; * ? ? ?
* ) by accounting for star - formation in h@xmath66-cooling mini - haloes , @xmath215 , with a self - consistent treatment of the reionization and metal enrichment of the milky way environment . by assuming that primordial faint sn , with mixing and fallback , dominated the early chemical enrichment , as suggested by observations of galactic halo stars ( e.g. *
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , we analyze the _ average _ properties of dwarf galaxies in different luminosity ranges , and show that : * cemp - no stars should exist in _ all _ dwarf galaxies within the observed luminosity range , @xmath216 .
* independent of galaxy luminosity the relative fraction of cemp - no stars increases towards lower [ fe / h ] , reaching @xmath209 at [ fe / h]@xmath131 .
* as the galaxy luminosity increases , the overall probability to observe cemp - no stars decreases , and the [ fe / h ] range in which they are most likely to be found is shifted towards higher values . * in classical sculptor - like dsph galaxies , the probability to find cemp - no stars is almost an order of magnitude lower than in the faintest milky way companions , @xmath149 , and it is maximal , @xmath217 at [ fe / h]@xmath3 .
the results explain why it is easier to discover cemp - no stars in ultra - faint dwarf galaxies than in more luminous classical dsphs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and also why the only cemp - no star observed in sculptor has been found at an unexpectedly high [ fe / h ] @xcite . in particular
, our model shows that cemp - no stars at [ fe / h]@xmath218 form in environments predominantly polluted by primordial faint sn , which therefore have extremely high [ c / fe ] values . on the other hand , those at @xmath219[fe / h]@xmath220
are imprinted by both primordial faint sn and low metallicity , @xmath190 , snii , in agreement with recent findings by bonifacio et al .
( 2015 ) . at [ fe / h]@xmath221
, we find that cemp stars can also form in an ism polluted by the products of agb stars with @xmath190 .
hence , they may potentially be enriched by s - process neutron capture elements ( i.e. cemp - s ) , which are not included in our model .
our findings for cemp - no stars , are a consequence of both the extremely low fe - production , and high c , from faint sn , and of the cosmological context in which the hierarchical assembly of dwarf galaxies occurs .
h@xmath66-cooling minihaloes , or ultra - faint dwarf galaxies , @xmath6 , are thus predicted to be the high - redshift progenitors of more massive `` classical '' dsph galaxies , and the environment of formation for stars with [ fe / h]@xmath4 .
the low - fe tails of dwarf galaxies are thus expected to form in these common building blocks , some of which experienced pop iii star formation @xcite .
such low - fe tails can extend down to [ fe / h]@xmath222 if built - up upon the chemical products of primordial faint sn .
as the galaxy luminosity increases our model shows that the average mdfs become more _ peaked _ , and shifted towards _ higher _ [ fe / h ] values , as is observed .
thus , stars at [ fe / h]@xmath4 become a _ lower fraction _ of the total stellar populations : from @xmath5 for ultra - faint dwarf galaxies , @xmath6 , to @xmath223 for @xmath224 .
this is why in more massive dwarf galaxies the overall probability to observe a cemp - no star decreases , and the [ fe / h ] range in which they are more likely to be found increases . in sculptor ,
@xmath225 , we find that the fraction of cemp - no stars monotonically decreases from @xmath226 at [ fe / h]@xmath227 down to @xmath228 at [ fe / h]@xmath229 . however , the probability to observe one of these stars is higher at @xmath158[fe / h]@xmath230 because more stars can be observed in this range .
a key prediction of our work is that galaxies with different luminosity are expected to share , on average , similar mdf tails _ and _ fractions of cemp - no stars at the lowest [ fe / h ] . at the moment , this hypotheses is supported by the observations of a very similar fraction of cemp stars at [ fe / h]@xmath4 in the galactic halo and in the _ overall _ sample of nearby dwarf galaxies , @xmath231 .
however , more data need to be collected for the milky way companions .
this also emerges from our analysis of the model uncertainties , which in the faintest dwarf galaxies are dominated by the low number statistics .
finally , we should be clear that our model can not exclude alternative scenarios for the formation of cemp - no stars , such as metal pollution by massive rotating primordial stars ( e.g. * ? ? ?
* ; * ? ? ?
furthermore , we did not explore the effects of different primordial initial mass functions on the fraction of cemp - no stars in dwarf galaxies .
instead , we simply worked under the hypothesis that pop iii stars evolve as faint sn with mixing and fallback , since these pristine stars should dominate the early chemical enrichment to successfully explain observations of cemp - no stars in the galactic halo ( e.g. * ? ? ? * ; * ? ? ? * ) .
our findings show that to further test the predominant role of primordial faint sn in the early universe we need to at least double the current stellar sample in sculptor , by measuring [ fe / h ] and [ c / fe ] in stars in different region of the galaxy down to magnitudes @xmath194
. deeper observations ( @xmath204 ) should enable the discovery of peculiar cemp - no stars in sculptor similar to those found in the galactic halo at [ fe / h]@xmath111 .
this will provide crucial information not only on the nature of first stars and critical metallicity value , but also on the formation of the galactic halo , and hence on the underlying hierarchical galaxy formation process .
we thank the anonymous referee for his / her useful insights , and very constructive report .
s. salvadori is grateful to p. bonifacio and m. de bennassuti for useful discussions .
she warmly thanks j. norris and e. kirby for careful reading the first version of the paper , and c. chiappini for key insights on early results .
finally , she is grateful to the netherlands organization for scientific research , which supports her research through a veni grant 639.041.233 . | we investigate the frequency and origin of carbon - enhanced metal - poor ( cemp ) stars in local group dwarf galaxies by means of a statistical , data - calibrated cosmological model for the hierarchical build - up of the milky way and its dwarf satellites .
the model self - consistently explains the variation with dwarf galaxy luminosity of the observed : i ) frequency and [ fe / h ] range of cemp stars ; ii ) metallicity distribution functions ; iii ) star formation histories .
we show that if primordial faint supernovae dominated the early metal enrichment , then cemp - no stars enriched by the first stellar generations should be present in _ all _ dwarf galaxies , with similar number of stars and cemp fractions at [ fe / h]@xmath0 .
we demonstrate that the probability to observe a star that is carbon - enhanced within a given [ fe / h ] range strongly depends on the luminosity of the dwarf galaxy and , on average , it is an order of magnitude lower in `` classical '' sculptor - like dsph galaxies ( @xmath1 ) than in the least luminous ultra - faint dwarfs ( @xmath2 ) . in addition , we explain why it may be easier to find cemp - no stars at [ fe / h]@xmath3 in classical dsph galaxies than in ultra - faint dwarfs .
these are consequences of the dramatic variation in the fraction of stars at [ fe / h]@xmath4 with galaxy luminosity : @xmath5 for galaxies with @xmath6 , and @xmath7 for @xmath8 .
we present model predictions for the low fe - tail and cemp fraction of stars in dwarf galaxies , with particular emphasis on the sculptor dsph , that can be used to shed light on the properties of the first stars .
[ firstpage ] galaxies : high - redshift , dwarf , local group ; stars : abundances ; cosmology : theory . |
observations of carbon monoxide ( co ) provide a robust window into circumstellar chemistry . due to the large percentage of carbon ( @xmath9 to @xmath10 ) and oxygen ( @xmath11 ) tied up in co in the early stages of young stellar object ( yso ) formation @xcite
, co is a valuable tracer of chemical inheritance from the envelope to the disk in solar - type ysos .
the use of high - resolution spectroscopy to observe co at infrared wavelengths enables high - precision investigations of isotope fractionation in both oxygen and carbon toward ysos in embedded and disk - phase evolutionary stages , and is consequently an important tool for connecting the solar system to currently forming protoplanetary systems .
examples include observational studies toward diffuse clouds and inferring isotopologue - specific photodissociation by co self - shielding as a potentially important mechanism of carbon isotope fractionation in co in the interstellar medium @xcite .
further , high - resolution near - infrared observations of the co rovibrational band toward protoplanetary disks @xcite and protostellar envelopes have recently revealed signatures in co isotopologue ratios that are consistent with co self - shielding , with precision sufficiently high for comparison to meteoritic data @xcite .
observations of isotope ratios toward ysos and , where possible , their parent clouds , are valuable in assessing theorized similarities between circumstellar envelopes , disks and dense clouds local to their origin .
this is important not only for evaluating predictions of galactic chemical evolution ( gce ) models that are constrained by solar system abundances , but also for characterizing the range of carbon isotope ratios toward ysos varying with evolutionary stage , and establishing accurate astrochemical constraints on protoplanetary environments .
observational studies can help demonstrate whether there are significant isotopic differences or trends between diffuse clouds , dense molecular clouds , and ysos .
these in turn help construct explanations for apparent chemical anomalies , such as the @xmath0c/@xmath2c solar system - interstellar medium ( ism ) discrepancy and solar system oxygen isotope anomaly .
this paper focuses on the carbon isotopic systems in ysos as traced by high - resolution observations of co isotopologues , with additional mention of significant findings in oxygen isotopes .
current measurements of the @xmath0c/@xmath2c ratio for the solar system are @xmath12 to 89 , derived from photospheric models applied to the solar co spectrum @xcite , and bulk solar system materials @xcite .
values of @xmath0c/@xmath2c for the solar system have been consistently higher than those of @xmath13 to 69 observed toward the local ism , measured using millimeter to sub - millimeter rotational transitions of various carbon - bearing molecules , co , cn , ch@xmath14 , and h@xmath15co @xcite .
the apparent discrepancy in @xmath0c/@xmath2c between the solar system and ism has been attributed to gce leading to enrichment of @xmath2c relative to @xmath0c over the @xmath16 gyr lifetime of the solar system ( e.g. * ? ? ?
* ; * ? ? ?
an important property of gce models is the assumption that the solar system records the chemical state of the molecular cloud material @xmath16gyr ago .
that is , the assumption that the solar nebula was isotopically representative of its parent cloud at its galactic radius at the time of formation should also hold for present - day solar - type ysos .
the challenge for observational comparisons of solar system isotope ratios in general is the need for high precision in molecular abundance measurements .
while techniques using strong rotational lines in the millimeter / sub - millimeter range often involve model - dependent comparisons between optically thick and thin lines , high - resolution infrared spectrographs have enabled significant breakthroughs in observations of nearby ysos .
infrared absorption spectroscopy of optically thin co lines allows for precise column density measurements along pencil - beam lines - of - sight , enabling the derivation of precise column densities and isotope ratios from the line profiles of as many as four currently observable co isotopologues ( @xmath0c@xmath1o , @xmath2c@xmath1o , @xmath0c@xmath17o , @xmath0c@xmath18o ) .
robust isotope ratios so derived can be compared with statistical significance to sub - millimeter ism values , solar abundances , and meteoritic oxygen .
a high degree of precision ( within @xmath19 ) on the absolute values of the column densities has been demonstrated for the four most common co isotopologues @xcite .
the study reported here builds on this work .
our targets span a range in @xmath0c@xmath1o ice optical depth , from nearly saturated to undetected , providing an opportunity to investigate a potential observational connection between the co ice and gas reservoirs by comparing the gas - phase [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] to the co ice column along a single line - of - sight .
since solids tend to concentrate the heavy isotope , relative ice - gas partioning may lead to a paucity of @xmath2c ( or excess of @xmath0c ) in the gas phase , with the opposite signature sequestered in the ice .
data for all targets were obtained as part of a european southern observatory large program @xcite to observe @xmath20 ysos and protoplanetary disks with the cryogenic infrared echelle spectrograph ( crires ) at the very large telescope ( vlt ) in chile .
the crires instrument is an adaptive optics ( ao)-assisted spectrometer that operates at very high resolving power ( @xmath5/@xmath6@xmath5 @xmath21 95,000 ) .
llll[ht ] doar24e s , n & oph & 125 & stage ii / ir companion + irs 43 & oph & 125 & stage i + irs 51 & oph & 125 & extincted stage ii + irs 63 & oph & 125 & stage i + rno 91 & l43 & 125 & stage i + vv cra
n , s & cra & 130 & stage ii / ir companion + hl tau & tau & 140 & stage i + reipurth 50 & ori & 470 & stage i / fu ori + lln 19 & vel & 700 & stage i [ target_table ] targets tracing high column densities of molecular gas were selected based on evidence of deep co absorption lines at 4.7@xmath4 m in spectra obtained with vlt - isaac @xcite and keck - nirspec ( e.g. , * ? ? ?
selected targets were observed by crires in the co overtone ( @xmath22 ) band at 2.3 @xmath4 m and the fundamental ( @xmath23 ) band at 4.7 @xmath4 m . because the overtone transitions are weaker than those of the fundamental band by a factor of @xmath24 , the former are typically optically thin for @xmath0c@xmath1o , thereby enabling the precise measurement of column densities for this molecule .
the spectra of the fundamental rovibrational bands were obtained using the @xmath25 slit , resulting in a resolving power of @xmath26 ( corresponding to 3.2kms@xmath27 ) .
several settings in the fundamental band were obtained for all targets in order to probe a range of rotational levels for @xmath2c@xmath1o , @xmath0c@xmath17o and @xmath0c@xmath18o , spanning @xmath28 to at least @xmath29 .
observations were timed such that the doppler shifts of the targets relative to the earth s atmosphere were at least 20kms@xmath27 due to a combination of the intrinsic target velocity and reflex motion of the earth . as a consequence
, telluric co lines are typically not affecting the absorption lines of the targets .
lllll[ht ] doar24e s , n & 16 26 23.36 & -24 20 59.8 & @xmath30 & 2007 - 09 - 02 + & & & @xmath31 & 2010 - 03 - 18 + irs 43 & 16 27 26.94 & -24 40 50.8 & @xmath32 & 2008 - 08 - 05 + & & & @xmath31 & 2008 - 08 - 04 + irs 51 & 16 27 39.83 & -24 43 15.1 & @xmath33 & 2008 - 08 - 02 + & & & @xmath34 & 2008 - 08 - 03 + irs 63 & 16 31 35.65 & -24 01 29.3 & @xmath35 & 2007 - 04 - 25 + & & & @xmath31 & 2008 - 08 - 07 + rno 91 & 16 34 29.28 & -15 47 01.5 & @xmath36 & 2010 - 03 - 18 + & & & @xmath31 & 2010 - 04 - 08 + vv cra n , s & 19 03 06.74 & -37 12 49.7 & @xmath37 & 2007 - 08 - 31 + & & & @xmath31 & 2007 - 08 - 31 + hl tau & 04 31 38.44 & 18 13 57.7 & @xmath38 & 2010 - 01 - 30 + & & & @xmath39 & 2010 - 01 - 28 + reipurth 50 & 05 40 27.70 & -07 27 28.0 & @xmath40 & 2007 - 10 - 11 + & & & @xmath31 & 2007 - 10 - 17 + lln 19 & 08 48 48.00 & -43 32 25.0 & @xmath41 & 2007 - 04 - 25 + & & & @xmath42 & 2007 - 04 - 25 [ obs_table ] table [ target_table ] summarizes the region , distance and evolutionary stage of the observed targets , listed in order of increasing distance from the local solar neighborhood ( @xmath43kpc ) .
specific observing parameters are listed in table [ obs_table ] .
selected targets probe a range of circumstellar environments , as noted .
because the lines - of - sight traced by infrared absorption may pass through the upper layers of an inclined disk as well as through envelope material and a foreground molecular cloud unrelated to the yso , some portion of each of these environments may be represented in the integrated spectra , depending on the target . for instance
, near - infrared absorption spectra observed toward the extincted protoplanetary disk , irs 51 , will be an integrated sight line consisting mostly of the foreground ophiuchus ridge and a relatively small contribution of the protoplanetary disk .
this geometry has been predicted by interferometric observations @xcite and is supported by our observations that reveal cold - only gas lines . ) .
spectra for doar24e n , s show few co absorption lines , while vv cra n shows a rich co forest of narrow rovibrational absorption lines .
sample lines are marked .
the low - level emission from hot co gas toward doar24 s and vv cra n is possibly representative of inner disk disk material . the weak emission line at @xmath44 m for doar24e
n is hi pfund@xmath45 .
the main portions of the spectra used for the line analyses are presented.,width=321 ] the crires spectra were reduced using standard procedures for infrared echelle spectroscopy , including flat - field correction , adjustments to account for detector non - linearity , and linearization of the spectral traces in dispersion- and cross - dispersion directions .
the spectra were optimally extracted , wavelength - calibrated using the telluric absorption lines referenced to an atmospheric model spectrum , and transformed to the local standard - of - rest frame .
relative flux calibration was carried out by dividing the targets by observations of early - type telluric standards observed immediately surrounding observations of the science target at the same airmass .
the division was optimized by modifying the telluric spectra to account for any difference in grating angle , airmass , and spectral resolving power . to aid in the measurement of any intrinsic line broadening
, the resolving powers for each science spectrum were measured from telluric spectra to within an accuracy of @xmath46kms@xmath27 , by fitting the `` reference forward model '' ( rfm , see also section [ sec : derivation ] ) .
further details on crires data processing is described in @xcite . ) .
hl tau and lln 19 have a rich co forest of narrow rovibrational absorption lines due to the various co isotopologues present .
sample lines are marked .
few lines of the rarer isotopologues ( @xmath0c@xmath17o and @xmath0c@xmath18o ) are seen toward the more deeply embedded object , lln 19 . low - level emission toward vv cra s and hl tau are possibly representative of hot co gas of the inner disk.,width=321 ] 1).the rich co forest of narrow rovibrational absorption lines are due to the various co isotopologues present .
sample lines are marked .
rno 91 shows low - level emission possibly representative of the inner disk.,width=321 ] transitions .
sample lines are marked.,width=321 ]
the fundamental @xmath48 and first - overtone @xmath49 rovibrational bands are shown for all targets in figures [ specm1a ] through [ speck2 ] , in ascending order of co ice optical depths at @xmath50 m . a forest of narrow co rovibrational absorption lines is seen to varying degrees in all spectra . a magnified spectral region of hl tau is shown as a representative example of individual line profiles ( figure [ spec_zoom ] ) .
broad , complex @xmath48 co emission is often seen in the spectra , and likely originates in the innermost ( @xmath51au ) regions of the disks @xcite .
however , the emission does not interfere with the analysis , as it is much broader than the absorption , and is generally weak or absent in the rare isotopologue transitions .
the potentially affected @xmath48 @xmath0c@xmath1o lines are not used for column density measurements .
broad absorption observed toward several objects at @xmath49 ( figures [ speck1 ] , [ speck2 ] ) likely originates in the stellar photosphere and is clearly separated from the narrow interstellar co lines .
temporal variation in [ @xmath0co]/[@xmath2co ] was found toward vv cra s , with a [ @xmath0co]/[@xmath2co ] abundance ratio of @xmath52 when using a @xmath48 observation four months prior to the first - overtone @xmath49 , as compared to @xmath53 when coeval observations were used ( errors are statistical 1@xmath54 values ) . the possibility for temporal spectral variability
should therefore be considered for doar24e and irs 63 , targets for which only non - coeval observations are available .
total abundances for each isotopologue were derived by measuring the column density of co in each observed rovibrational level of the ground vibrational state which , in turn , was obtained from the intrinsic optical depths at line center @xmath55 derived from a best - fit absorption line profile to each rovibrational transition .
individual column densities were used to make a best fit to a boltzmann population , then summed over all levels .
this method is described in detail in @xcite .
llllllll[ht ] doar24e s & 2.79 & 3.12 & @xmath56 & @xmath57 & @xmath58 & @xmath59 &
+ doar24e n & 2.79 & 3.12 & @xmath56 & @xmath60 & @xmath61 & @xmath62 & + vv cra n & 2.82 & 3.49 & @xmath63 & @xmath64 & @xmath65 & @xmath66 & @xmath67 + vv cra s & 2.82 & 3.49 & @xmath68 & @xmath69 & @xmath70 & @xmath71 & @xmath72 + hl tau & 3.26 & 3.43 & @xmath73 & @xmath74 & @xmath75 & @xmath76 & @xmath77 + lln 19 & 3.22 & 4.68 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & + rno 91 & 3.26 & 3.72 & @xmath82 & @xmath83 & @xmath84 & @xmath85 & @xmath86 + reipurth 50 & 3.44 & 4.54 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath91 + irs 63 & 3.57 & 3.39 & @xmath92 & @xmath93 & @xmath94 & @xmath95 & @xmath96 + irs 43 & 3.03 & 3.45 & @xmath97 & @xmath98 & @xmath99 & @xmath100 & @xmath101 + irs 51 & 3.03 & 3.37 & @xmath102 & @xmath103 & @xmath104 & & [ tau_table ] the technique of using the line profiles to directly determine column densities takes advantage of the high spectral resolution of crires to spectrally resolve the velocity dispersion of the individual co lines .
this in turn leads to greater precision than using unresolved lines which require a curve - of - growth analysis to derive intrinsic line widths @xcite . with crires ,
the resolving power is sufficiently high to partially resolve the intrinsic line profiles ( fwhm@xmath105 ) , albeit marginally in some cases : vv cra s , rno 91 , and irs 43 .
therefore , it is possible that the derived column densities toward these targets may be more uncertain .
llll[ht ] @xmath0c@xmath1o & @xmath106 & @xmath107 & @xmath108 + @xmath0c@xmath1o & @xmath109 & @xmath110 & @xmath111 + @xmath0c@xmath1o & @xmath112 & @xmath113 & @xmath114 + @xmath0c@xmath1o & @xmath115 & @xmath116 & @xmath117 + @xmath0c@xmath1o & @xmath118 & @xmath119 & @xmath120 + @xmath0c@xmath1o & @xmath121 & @xmath122 & @xmath123 + @xmath0c@xmath1o & @xmath124 & @xmath125 & @xmath126 [ line_tab ] the measured instrumental broadening , intrinsic velocity dispersions and optical depth ranges for each observed isotopologue are shown in table [ tau_table ] .
line - by - line optical depths and doppler shifts for observed targets are available in the online version of this paper , with a sample partial line list shown in table [ line_tab ] .
complete data tables are available in the online version of the paper .
instrumental broadening values were modeled individually for each spectrum in its entirety , following the method in @xcite for using an atmospheric model for telluric spectra .
the reference forward model ( rfm ; http://www.atm.ox.ac.uk/rfm/ ) code was used to derive a detailed fit to the individual telluric standard star spectra , with consideration of meteorological conditions at the time of observation .
the atmospheric molecular parameters were taken from the hitran 2008 database @xcite .
there are small relative differences in resolving power between the standard stars and science targets due to differences in seeing or adaptive optics ( ao ) correction , but these differences were accurately measured and corrected for by minimizing the continuum residuals after telluric division .
optimized intrinsic velocity dispersions were derived from the weighted mean of dispersions measured in individual @xmath0c@xmath17o lines , assuming that both the intrinsic and instrumental profiles are well - approximated by gaussian profiles .
the @xmath0c@xmath17o lines were used because this rare isotopologue is always optically thin , and therefore is not subjected to saturation line broadening , which may affect the more common @xmath2co isotopologue .
optical depths of the optically - thin @xmath0c@xmath1o @xmath49 lines were similarly computed using the derived intrinsic broadening from the @xmath48 @xmath0c@xmath17o lines .
it can be noted that , even when the resolving power of the two spectral bands differed for a single target , we obtained good fits to lines in both bands under the reasonable assumption that the intrinsic profile width is independent of transition and isotopologue .
further , it is reasonable to use the same velocity dispersion parameter for a target with a two - temperature distribution of the data , since thermal contributions are very small contributors to the total intrinsic dispersion ( 0.077 and 0.242 km s@xmath27 for @xmath0c@xmath1o at 20 and 200 k , respectively ) .
the rotational analyses for all targets are shown in figures [ roplots1 ] and [ roplots2 ] .
sub - level column densities , @xmath127 , derived from the fits to the absorption lines , are plotted against the energy , @xmath128 , of the rotational transition state , @xmath129 . for targets where cold gas dominates ,
a single - temperature model is sufficient to fit the data , while targets with absorption that includes higher-@xmath129 transitions are best fit with a two - temperature model , representing cold and warm environments along the line - of - sight , and allowing for different excitation temperatures for each isotopologue along a given line - of - sight .
derived isotopologue temperatures from the fits are shown in table [ temp_table ] .
high temperatures ( @xmath130 to 250 k ) are roughly consistent with those of the inner parts of typical yso envelopes , while low temperatures ( @xmath131 to 30 k ) may trace outer envelopes or molecular cloud material in front of the ysos . for lines - of - sight with no warm component ,
it is likely that the gas is from a foreground cloud not related to the yso itself ( doar24e n and s ) , vv cra s , and irs 51 .
total column densities and isotopologue ratios are shown in table [ ratio_table ] . due to the dominance of higher-@xmath129 lines in targets with both low- and high - temperature lines ,
the higher - temperature ratios are generally better constrained within two - temperature environments @xcite .
statistical errors on the column densities are generally found to be small , reflecting the very high signal - to - noise ratio of the data .
for example , average signal - to - noise per pixel in the fundamental bands is @xmath132 . with an average crires fwhm of 2.0 pixels ,
we retrieve average line errors of @xmath133 ( 1/(@xmath134 ) ) .
first - overtone bands average @xmath135 .
we therefore expect to know the column densities to a fraction of a percent , which is represented in our results .
these errors do not take into account any potential systematic errors in extracting the line strengths and profiles due to inherent non - gaussianity of the line - fitting procedure , which we can not quantify with certainty .
however , to roughly estimate the magnitude of systematic retrieval errors , we carried out a small number of forward models with 2-dimensional radiative transfer models ( radlite , * ? ? ? * ) simulating synthetic co isotopologue absorption spectra using a basic one - dimensional infalling envelope from @xcite .
the synthetic spectra were convolved with the appropriate instrumental resolution and isotopologue ratios retrieved from the synthetic spectra using the same methods as for the observed data . in each case
, we were able to reproduce the input ratios to better than 10% .
we therefore estimate that model - dependent systematics add at most a 10% error to the retrieved isotopologue ratios , in addition to the statistical error .
llcccc[ht ] doar24e s & single & @xmath136 & @xmath137 & @xmath138 &
+ doar24e n & single & @xmath139 & @xmath140 & @xmath141 & + vv cra n & high & @xmath142 & @xmath143 & @xmath144 & @xmath145 + vv cra n & low & @xmath146 & @xmath146 & @xmath147 & @xmath148 + vv cra s & single & @xmath146 & @xmath149 & @xmath150 & @xmath151 + hl tau & high & @xmath152 & @xmath153 & @xmath154 & @xmath155 + hl tau & low & @xmath156 & @xmath157 & @xmath158 & @xmath159 + lln 19 & single & @xmath160 & @xmath161 & @xmath162 & + rno 91 & high & @xmath163 & @xmath164 & @xmath165 & + rno 91 & low & @xmath166 & @xmath167 & @xmath168 & @xmath169 + reipurth 50 & high & @xmath170 & @xmath171 & @xmath172 & @xmath173 + reipurth 50 & low & @xmath174 & @xmath137 & @xmath175 & @xmath146 + irs 63 & high & @xmath176 & @xmath177 & @xmath178 & @xmath179 + irs 63 & low & @xmath180 & @xmath140 & & + irs 43 & high & @xmath181 & @xmath182 & @xmath183 & @xmath184 + irs 43 & low & & @xmath185 & @xmath167 & @xmath186 + irs 51 & single & @xmath187 & @xmath137 & & [ temp_table ] the doar24e system is a hierarchical multiple for which we have spectra of two components , with the rotational analyses nearly identical for both components .
this lack of significant spatial variability strongly suggests that we are tracing spatially extensive and homogenous foreground material , that in turn is the dominant contributor smoothly distributed on scales of at least a few 100 au .
the lack of isotopic variability between the doar24e components gives us further confidence on our retrieval method for deriving precise isotope ratios .
due to high extinction leading to a faint @xmath188 band continuum for the embedded yso irs 43 , we detect high-@xmath129 lines for all co isotopologues in @xmath189 , but only low-@xmath129 lines for @xmath0c@xmath1o in the @xmath188 band . given this disparate set of temperature ranges , our data for irs 43 are therefore not useful for measuring co isotopologue abundance ratios relative to @xmath0c@xmath1o . toward vv cra s ,
we derive a @xmath0c@xmath1o column density of @xmath190@xmath191 .
this is nearly three orders of magnitude higher than the value of @xmath192@xmath191 derived in @xcite .
our rotational temperature ( @xmath193k ) is also significantly lower than their value ( @xmath194k ) .
since @xcite used the @xmath48 spectra from the archive , these differences could be explained by strong saturation effects resulting from their use of the highly optically thick @xmath189 @xmath0c@xmath1o lines .
this difference reinforces the necessity of using @xmath188 data for deriving precise @xmath0c@xmath1o column densities .
@xcite proposed that the cold gas component observed toward vv cra s may be interpreted as either molecular cloud material in front of the vv cra binary system , or disk gas toward the primary component as suggested by @xcite .
we point out that the former is more likely because the presence of both warm and cold components are expected for a disk , while a foreground cloud will typically only show a single cold component , such as that seen towards vv cra n. taken together , the likely geometry for the vv cra system and the spectroscopic results render it unlikely that the vv cra s spectra represent outer disk .
vv cra n observations most likely represent warm and cold disk gas .
llccccc[ht ] doar24e s & single & @xmath195 & @xmath196 & : @xmath197 & &
+ doar24e n & single & @xmath198 & @xmath199 & : @xmath200 & & + vv cra n & high & @xmath201 & @xmath163 & @xmath202 & @xmath203 & @xmath204 + vv cra n & low & @xmath205 & @xmath206 & @xmath207 & @xmath208 & @xmath209 + vv cra s & single & @xmath210 & @xmath211 & : @xmath212 & : @xmath213 & @xmath214 + hl tau & high & @xmath215 & @xmath216 & @xmath217 & @xmath218 & @xmath219 + hl tau & low & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath224 + lln 19 & single & @xmath225 & @xmath226 & : @xmath227 & & + rno 91 & high & @xmath228 & @xmath229 & @xmath230 & @xmath231 & @xmath232 + rno 91 & low & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 + reipurth 50 & high & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 + reipurth 50 & low & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 + irs 63 & high & @xmath248 & @xmath249 & @xmath250 & @xmath251 & @xmath252 + irs 63 & low & @xmath253 & @xmath254 & & & + irs 43 & high & @xmath255 & : @xmath256 & : @xmath257 & : @xmath258 & @xmath259 + irs 43 & low & & : @xmath260 & : @xmath261 & : @xmath262 & : @xmath263 + irs 51 & single & @xmath264 & @xmath265 & & & + local ism & & & @xmath266 & @xmath267 & @xmath268 & @xmath269 [ ratio_table ] derived [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] ratios are plotted against galactocentric radius ( r@xmath270 ) in figure [ gc_co_plot ] .
the solar system and local ism values of @xmath271 @xcite and @xmath272 @xcite are plotted at 8 kpc , respectively . derived [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] toward our targets range from @xmath273 to 165 , values inconsistent with those toward most ysos and molecular clouds , or from gce @xcite and regression models @xcite .
the distribution of data from the literature approximates the general trend of increasing metallicity toward the galactic center , with the corresponding increase in @xmath0c/@xmath2c with greater r@xmath270 .
c/@xmath2c ratios vs. r@xmath270 ( kpc ) .
gray symbols show @xmath0c/@xmath2c values from the literature , derived from different carbon - bearing molecules , co , cn , ch@xmath14 and h@xmath15co .
data from our study are labeled for yso targets ( filled circles ) and foreground clouds ( unfilled circles ) .
colors connect these data by molecular cloud with those in figure [ delta_delta ] . for targets where there is both warm and cold gas along a line - of - site
, ratios are shown for the more robustly determined warm gas , unless otherwise noted .
local ism ( @xmath274 ; @xcite ) and solar system values ( @xmath275 ; @xcite ) are noted .
the black arrows indicate high @xmath0c/@xmath2c in diffuse regions of ophiuchus @xcite .
yso data for monr2 irs3 , lkh @xmath276101 , afgl 490 @xcite and hl tau ( lower value ; @xcite ) are marked ( dark purple).,width=321 ] previous studies have discussed the exchange between co solid- and gas - phase reservoirs in disks and envelopes ( e.g. * ? ? ?
* ; * ? ? ?
* ) . with a range of co ice column densities centered at @xmath277 m ( figures [ specm1a ] through [ specm2b ] ) ,
our data set provided an opportunity to investigate relationships between the gas - phase isotopologue ratios and the co ice fraction along single lines - of - sight .
co ice column densities were calculated from optical depth analyses in @xcite and @xcite . of the broad co ice profile ,
the pure , cold co component is most relevant with respect to interactions with co gas , as it is the solid state reservoir that would most likely directly exchange with the gas in cold , dense clouds .
we used optical depths for this component to derive total co ice column densities for each target . using the phenomenological decomposition analyses in @xcite
, the pure component corresponds to the middle optical depth ( @xmath278 ) profile measurement ; @xcite was used for lln 19 . for targets with no ( vv cra ,
doar24e ) or very weak ( hl tau ) spectral ice features , we used upper limits from the crires fundamental bands .
total column densities of pure co in the ice phase @xmath279 were computed using the relation from @xcite , @xmath280 where @xmath281 is the optical depth at line center of the pure co component , @xmath282 is the band strength of @xmath0co , and the numerical factor , 6.03cm@xmath27 , considers the effect of grain shape on the band strength . in this study
, we used the continuous distribution of ellipsoids ( cde ) model @xcite in computing solid co column densities .
optical depths , column densities , and total co ice fractions are listed in table [ ice_table ] , in ascending order of co ice optical depth . although it is assumed that cold co gas will more likely exchange with co ice , for targets where co gas was observed in both cold and warm environments , values for co ice fractions are shown separately for each environment .
the relation between the co ice fraction and [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] gas - phase ratios is shown in figure [ ice_vs_ratio ] , with co ice fractions plotted against derived [ @xmath0co]/[@xmath2co ] for cold ( @xmath284 to 30k ) and warm ( @xmath285 to 250k ) regimes , the latter only for targets where warm gas was observed .
we distinguish between targets where only cold co gas is observed , and targets where both warm and cold temperatures is found . in the latter cases ,
cold - gas measurements are likely more uncertain due to the contribution of the high - temperature on the low - temperature lines .
the optical depth for the @xmath2co ice toward the foreground cloud traced by irs 51 provides the only solid [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] ratio ( @xmath288 ) that can be compared to a gas - phase ratio along the same line - of - sight in this study .
the ice - phase ratio is significantly lower than our derived gas - phase ratio of @xmath289 .
the [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] value in ice for irs 51 is roughly consistent with other published [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] values for ice in yso targets : @xcite derive a [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] ratio of @xmath290 for the ice observed toward the massive yso ngc 7538 irs 9 , using keck - nirspec , and a lower ratio of @xmath291 was observed toward the crbr 2422.8 - 3423 edge - on disk using vlt - isaac @xcite .
llllll[ht ] doar24e s & single & @xmath292 & @xmath293 & @xmath294 & @xmath293 + doar24e n & single & @xmath292 & @xmath293 & @xmath295 & @xmath293 + vv cra n & low & @xmath296 & @xmath297 & @xmath298 & @xmath292 + vv cra n & high & @xmath296 & @xmath297 & @xmath299 & @xmath300 + vv cra s & single & @xmath296 & @xmath297 & @xmath301 & @xmath297 + hl tau & low & @xmath302 & @xmath296 & @xmath303 & @xmath304 + hl tau & high & @xmath302 & @xmath296 & @xmath305 & @xmath306 + lln 19 & single & @xmath307 & @xmath308 & @xmath309 & @xmath310 + rno 91 & low & @xmath311 & @xmath312 & @xmath313 & @xmath314 + rno 91 & high & @xmath311 & @xmath312 & @xmath315 & @xmath316 + reipurth 50 & low & @xmath317 & @xmath318 & @xmath319 & @xmath320 + reipurth 50 & high & @xmath317 & @xmath318 & @xmath321 & @xmath307 + irs 63 & low & @xmath322 & @xmath323 & @xmath324 & @xmath325 + irs 63 & high & @xmath322 & @xmath323 & @xmath326 & @xmath327 + irs 51 & single & @xmath328 & @xmath329 & @xmath330 & @xmath331 [ ice_table ] c@xmath1o ice fraction with the [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] gas - phase ratios for targets observed in this study . for targets where both warm and cold gas was observed , co ice fractions are plotted with red diamonds when warm gas was used in computing the fraction , and blue stars when cold gas was used . for targets where only cold gas was observed ,
co ice fractions are shown with over - plotted blue ovals .
errors are @xmath332 propagated from errors in ice and gas optical depths measurements . @xmath333
upper limits ( down - facing arrows ) are shown for doar24e n , s , vv cra n , s , and hl tau.,width=321 ]
we find significant heterogeneity in gas - phase [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] toward our targets ( figures [ gc_co_plot ] and [ ice_vs_ratio ] ) , with most values falling above local ism values of @xmath334 , as determined by sub - millimeter observations @xcite .
the large spread in [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] suggests either heterogeneity in the carbon isotopic composition in our local solar neighborhood ( e.g. r@xmath270@xmath335 kpc ) , and/or one or more new fractionation mechanisms for co operating in the dense envelopes of young stars . the distribution of co isotopologue ratios derived from our crires data do not match the general galactic trend , which suggests that the carbon reservoirs of young stellar envelopes do not necessarily isotopically represent their parent clouds , as is often assumed in gce models @xcite .
we are not the first to observe unusual [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] abundance ratios in ysos and molecular clouds ( figure [ gc_co_plot ] ) . in ultraviolet absorption ,
@xcite and @xcite found high [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] values toward diffuse regions in the ophiuchus cloud , with values ranging from @xmath336 to 167 .
these results have been attributed to selective co photodissociation by external ultraviolet radiation @xcite .
we discuss connections with co photodissociation in section [ photodissociation ] .
unusually high [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] values were also detected in rovibrational absorption using subaru ( @xmath337 toward the massive young stars lkh@xmath276101 and monr2 irs3 , with values of @xmath338 and 158 , respectively @xcite .
@xcite suggested that photon - trapping in the rotational lines of the most abundant isotope as a possible explanation for the apparently elevated carbon isotope ratios , although the potential effect was not quantified in their study .
a possible mechanism for the observed co fractionation is mass - dependent exchange between gas and ice .
compared to single - dish , beam - averaged millimeter measurements of co gas - phase isotopologue ratios in molecular clouds , rovibrational absorption through protostellar envelopes is likely to probe denser gas .
dense , cold gas is known to lead to efficient co freeze - out onto dust grains @xcite . if the adsorption / desorption rates of co are mass - dependent , isotopologue fractionation may occur .
further , if the heavier isotopologue ( @xmath2co ) is more tightly bound to the dust grains , we would expect a increase in the [ @xmath0co]/[@xmath2co ] ratio in the gas , which is consistent with the detection of large columns of co ice along the observed lines - of - sight . here
we discuss whether ice - gas partitioning is a possible explanation for the observed high co isotopologue ratios in the gas .
the relative fraction of co sequestered in ice , averaged along the line - of - sight , is plotted against the [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] abundance ratios in the gas ( figure [ ice_vs_ratio ] ) .
while we might expect to see increased values of the co abundance ratio for lines - of - sight with a high ice / gas fraction if @xmath2co is preferentially bound in the ice mantles , we instead find significant scatter , much larger than the individual errors on each data point , with no simple apparent relation between these quantities .
a correlation is particularly absent between cold - gas reservoirs and co ice fraction , although these two quantities are likely cospatial and trace the same total co reservoir . however , the line - of - sight with the largest ice fraction ( irs 51 ) also has the highest co isotopologue ratio in the gas
. therefore , while other factors are likely equally important , we can not rule out that ice - gas partitioning could play a significant role in the observed high gas - phase [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] .
with respect to other factors affecting the gas - phase ratios , we note that high [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] ratios are also observed in the warm gas in some lines - of - sight ( figure [ ice_vs_ratio ] , red diamonds ) , which is unexpected if ice - gas partitioning is the only fractionation mechanism , since environments warm enough to fully sublimate the co ice should reset the gas isotopologue fraction back to the canonical value of @xmath8 . in order to reconcile this with ice - gas partitioning
, there must be a mechanism to effectively remove co from the ice during the cold phase of envelope evolution .
chemical pathways removing co from the ice include the successive hydrogenation of co to formaldehyde ( h@xmath15co ) and methanol ( ch@xmath341oh ) @xcite , or reactions with oh or o to form co@xmath342 @xcite .
further , we would expect the column densities of ice to be comparable or larger than those of the gas in order for the ice to be a significant reservoir that could increase the gas - phase [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] . for irs
51 , the large discrepancy in [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] between the ice ( @xmath272 ) and the gas ( @xmath343 ) , along with a seemingly low co ice fraction ( @xmath344 ) , suggests instead that open - system chemistry may be occurring via co ice reactions . while the data are inconclusive with regard to carbon isotopic fractionation between reservoirs of co ice and gas , potential trends in cold- and warm - gas reservoirs can be further investigated with high - resolution spectroscopy for a larger sample of protostellar lines - of - sight .
further , models considering chemical evolution in protoplanetary environments including ysos could consider interactions between ice and gas reservoirs as a potential influence on isotope partitioning . to estimate whether there is an expectation of ice - gas partitioning for co , we briefly review the theoretical basis for a potential mass - dependence in the ice - gas balance . following a standard approach for ice - gas partitioning ( e.g. * ? ? ?
* ) , co separations between the solid- and gas - phase reservoirs can be investigated with respect to mass - dependent kinetics and the balance between the rates of @xmath0c@xmath1o and @xmath2c@xmath1o gas adsorption and desorption , as given by the relation , @xmath345 where @xmath346 is the number density of ice , @xmath347exp@xmath348@xmath349 @xmath350@xmath351 , and @xmath352 are the desorption and adsorption rates , respectively .
the factor , @xmath353 , is the binding energy of co. the factor @xmath45 accounts for co desorbing only from the top monolayer of the ice , @xmath354 is the grain radius , and @xmath355 is the sticking coefficient .
the mass of co @xmath356 is an obvious mass - dependency of this relation , and @xmath357 is the pre - exponential factor .
the number densities of the co gas and dust particles are given by @xmath358 and @xmath359 , respectively @xcite .
in addition to the mass itself , the binding energy and pre - exponential factors are likely both mass - dependent , although to unknown degrees .
co]/[@xmath2co ] ratio predicted by the simple model described by eq .
[ eq : rate ] , assuming a difference in binding energy of 10k between the two isotopologues , as allowed by the experimental uncertainty , and a pre - exponential that scales with @xmath360 .
the model is a snapshot at @xmath361yr and the individual curves are created by varying the gas density from @xmath362 to @xmath363 and temperature from 9.1 to 10.0k , as indicated.,width=321 ] some experimental work exists for the mass dependence of thermal desorption of co. in an ultra - high vacuum , temperature - programmed desorption ( tpd ) experiment , @xcite measured binding energies and pre - exponentials for desorption from pure @xmath0co and @xmath2co ices .
the parameters were derived using an empirical kinetic model . however , no difference between the @xmath0co and @xmath2co binding energies could be measured within the experimental uncertainties : @xmath364 k and @xmath365 k for @xmath0c@xmath1o and @xmath2c@xmath1o , respectively . to estimate the degree of ice - gas fractionation allowed by the boundaries imposed by the experimental evidence , we solved for the ice abundance in eq .
[ eq : rate ] , assuming a 10k difference in binding energy between the two co isotopologues and a pre - exponential that scales as @xmath360 @xcite . in figure [ ratio_model ]
it is seen that , under these assumptions , the [ @xmath0co]/[@xmath2c ] gas ratio may reach values as high as 170 , but only for a narrow range of temperatures close to those of co desorption ( above which no ice exists ) .
essentially , a dust temperature exists at which @xmath2co can efficiently remain bound to the dust surface , while @xmath0co can not .
nearly all of this effect is due to the assumed difference of only 10k in the binding energies .
compared to what is currently available , significantly more precise laboratory measurements of mass - dependent binding energies would be needed to rule out binding - energy differences as a significant cause of ice - gas fractionation . while such temperature - dependent fractionation may be consistent with the observed spread in isotopologue ratios , it remains unclear to which degree the physical conditions of real molecular clouds and envelopes must be fine - tuned to lead to significant ice - gas fractionation .
more realistic modeling of ice - gas fractionation of co is potentially complicated by several factors .
for example , investigations of the effects of co entrapment in the ice on the desorption and adsorption ( both potential mass - dependent processes ) in astrophysical environments have shown that diffusion of co into the porous h@xmath15o ice leads to competing co entrapment and diffusion effects on the desorption of co , the degree to which depends on the ice temperature @xcite .
further , a recent study has shown that three - phase modeling of h@xmath15o : co@xmath15 ice desorption depends on the ice thickness and mixing ratios , which may have significant effects on ice desorption @xcite .
isotope - selective photodissociation of co due to self - shielding refers to the differential shielding of co isotopologues from photodissociation by far ultraviolet ( fuv ) radiation in proportion to their column densities .
with regard to oxygen isotopes , this leads to spatial regions characterized by more rapid destruction of @xmath0c@xmath18o and @xmath0c@xmath17o compared to the much more abundant @xmath0c@xmath1o molecule .
oxygen liberated during this process is thought to eventually end up in h@xmath15o , providing an explanation for the overabundance of h@xmath366o and h@xmath367o relative to h@xmath368o in the early solar system .
isotope - selective photodissociation of co due to self - shielding has been shown to be a relevant process in the interstellar medium @xcite .
the solar system oxygen isotope anomaly is characterized by the mass - independent correlation between [ @xmath1o]/[@xmath17o ] and [ @xmath1o]/[@xmath18o ] found in meteoritic materials as compared to the mass - dependent distribution known in terrestrial rocks .
mass independence refers to relative differences in [ @xmath1o]/[@xmath17o ] and [ @xmath1o]/[@xmath18o ] among primitive rocky objects that are nearly identical , suggesting changes in [ @xmath1o ] relative to both [ @xmath18o ] and [ @xmath17o ] , rather than the expected mass - dependent trend in which relative changes in [ @xmath1o]/[@xmath18o ] are about half those in [ @xmath1o]/[@xmath17o ] on a log scale .
the solar system oxygen anomaly has defied conclusive explanation since its discovery @xcite , and has been investigated observationally by @xcite and @xcite , the latter study finding mass - independent signatures in oxygen isotopes observed in co toward the vv cra n disk , consistent with isotope - specific photodissociation by co self - shielding .
the compiled oxygen data for this study , which also includes refined analyses of vv cra n and reipurth 50 from @xcite , are shown in table [ ratio_table ] and figure [ delta_delta ] .
the mass - dependent lines in figure [ delta_delta ] are derived by estimating the oxygen isotope ratios for the ism as a function of distance from the galactic center .
these values can be estimated from the 4 kpc and 8 kpc values @xcite , such that for @xmath369/[^{18}o]$ ] : @xmath370/[^{18}o]}_{r_{\rm gc } } = ( 57.5 \pm 10)\ { r_{\rm gc } } + 97 , \label{eq:5}\ ] ] where the @xmath332 uncertainty in the slope reflects the uncertainty in the standard errors reported in table 4 of @xcite .
for @xmath369/[^{17}o]$ ] , equation [ eq:5 ] is combined with values for @xmath371/[^{17}o]$ ] in the ism , which vary from 3.5 @xcite to 4.1 @xcite ( figure [ delta_delta ] ) .
we find a mass - independent signature in the oxygen isotopes relative to solar system abundances , consistent with co self - shielding , toward the hl tau disk , and the same mass - independent signature is retrieved in the refined results for vv cra n ( as compared to @xcite ) .
these findings lend further support for models considering co self - shielding on disk surfaces @xcite .
it is particularly interesting to note that the recent suggestions of planet formation in the hl tau disk based on new alma imaging @xcite would imply that co self - shielding may be a process that occurs and/or persists later in the evolution of ysos than previously thought .
we would not expect uv photolysis to be a significant process in the foreground cloud traced by vv cra s , or in the relatively embedded objects , rno 91 , reipurth 50 , and irs 63 .
nevertheless , due to the errors in the rare isotopologue , @xmath0c@xmath18o , a mass - dependent effect can not be distinguished from mass - independence in oxygen fractionation for rno 91 and irs 63 .
additional high - quality co data for ysos would help in evaluating mass - independence in oxygen for objects in various stages of evolution . with regard to the rare co isotopolgues , with the exception of low - temperature gas in reipurth 50 , the [ @xmath0c@xmath17o]/[@xmath0c@xmath18o ]
ratios are within @xmath372 of molecular cloud radio measurements for local ism , further strengthening the claim that the solar system is anomalous in @xmath17o/@xmath18o with respect to the local solar neighborhood @xcite .
this consistency in our data is observational support for @xmath0c@xmath17o and @xmath0c@xmath18o being similarly affected in young stellar environments .
while critical evaluation of mass - independence requires three isotopes , we qualitatively considered if isotope - selective photodissociation by co self - shielding could be affecting the derived [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] toward our targets by comparing the oxygen and carbon isotope ratios where this was possible : vv cra n , vv cra s , hl tau , rno 91 , reipurth 50 , and irs 63 .
these are plotted in figure [ oxygen_carbon ] , with the qualitative relative trend for co self - shielding overplotted on the data .
the derived [ @xmath1o]/[@xmath18o ] changes by nearly a factor of two relative to [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] , while the [ @xmath0c@xmath1o]/[@xmath0c@xmath17o ] varies by roughly 1.5 times with [ @xmath0c]/[@xmath2c ] .
if co self - shielding were the dominant mechanism , we would expect at least positive , direct changes in the carbon with the oxygen isotopes along an integrated line - of - sight , as well as similar variability in [ @xmath0c@xmath17o ] and [ @xmath0c@xmath18o ] compared with [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] due to their similar rates of photodissociation ( e.g. * ? ? ?
* ; * ? ? ?
therefore , while we can not be certain that photodissociation of co has no affect on the [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] , we can not claim that it is a dominant process affecting this carbon reservoir .
o]/[@xmath18o ] and [ @xmath1o]/[@xmath17o ] from local ism values are shown in per mil using the linearized form of the delta notation . here , @xmath373 and @xmath374 = [ c@xmath17o]/[c@xmath1o ] .
ratios for the ysos are derived from the best - constrained ( high- or single - t ) values from the rotational analyses ( table [ ratio_table ] ) .
the solid and dashed lines show mass - independent fractionation ( slope @xmath375 ) lines with different assumptions for the @xmath17o/@xmath18o ratio , as indicated .
the mass - dependent line ( slope @xmath376 ) is also shown for comparison .
ellipses represent 65% confidence limits ( standard deviation ) derived from @xmath332 errors from the @xmath377 for each line .
targets are colored according to parent cloud ( see also figure [ gc_co_plot]).,width=321 ] c@xmath1o]/[@xmath0c@xmath18o ] ratio is shown in purple , and [ @xmath0c@xmath1o]/[@xmath0c@xmath17o ] in red .
the derived [ @xmath1o]/[@xmath18o ] changes by nearly a factor of two relative to [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] , while the [ @xmath1o]/[@xmath17o ] varies by roughly 1.5 times with [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] .
the arrows indicate the relative trends we would expect if isotopologue - specific photodissociation by co self - shielding were the predominant process affecting the oxygen and carbon isotope ratios.,width=321 ] models for low - mass ysos predict that [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] should not increase significantly due to photodissociation ( e.g. * ? ? ?
* ; * ? ? ?
atomic carbon liberated by co photodissociation is photoionized largely to c@xmath378 , and c@xmath378 reacts rapidly to exchange carbon with co by the reaction @xmath2c@xmath378 + @xmath379o @xmath380 @xmath0c@xmath378 + @xmath2co + 35 k @xcite , driving @xmath2c back into co and diminishing the signature of selective photodissociation .
however , very high [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] ratios derived for diffuse clouds have nevertheless been attributed to co self - shielding .
values of @xmath336 , @xmath381 @xcite , and @xmath382 @xcite have been observed toward diffuse ophiuchus clouds that are thought to be subjected to high - uv fields , and where co photodissociation is thought to dominate over the carbon exchange reaction that would otherwise erase a photochemical signature .
while the upper range in [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] in our sample is consistent with the ophiuchus data ( figure [ gc_co_plot ] ) , we do not find the expected mass - independent signatures in oxygen isotopes consistent with co self - shielding .
further , as compared to diffuse clouds , we would not expect the molecular gas deep inside embedded yso envelopes to be subjected to such high uv fields given their relatively high values of visual extinction ( @xmath383 ( e.g. * ? ? ? * ) ) .
heterogenous [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] gas - phase abundance ratios , with several unusually high values , have been found using infrared absorption spectroscopy at very high resolution toward a sample of ysos in nearby star - forming clouds . a possible partial explanation for this trend may be co isotopologue partitioning between the ice and gas reservoirs .
the degree of partitioning and the preservation of the relation in warm regions where the co ice has sublimated would require the action of a process that can remove co from the ice phase without returning it to the gas .
additional high - quality data could help substantiate these findings .
signatures consistent with isotopologue - selective photodissociation by co self - shielding in oxygen were observed in the oxygen isotopes toward the disks , hl tau and vv cra .
however , our results do not support co self - shielding as a dominant mechanism for explaining the trends in the gas - phase [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] heterogeneity .
we conclude that chemical evolution models for young stellar environments should consider the interaction between co ice and gas reservoirs as a potential fractionation mechanism affecting the carbon isotope budget and influencing carbon isotopic heterogeneity in the gas .
interactions between these reservoirs could also contribute to isotopic heterogeneity between solar - type ysos and their parent clouds .
finally , isotopologue - selective photodissociation by co self - shielding should be considered an important mechanism affecting oxygen isotopes in protoplanetary disks , potentially persisting in the later stages of planet formation .
coplen , t. b. , bhlke , j. k. , de bivre , p. , ding , t. , holden , n. e. , hopple , j. a. , krouse , h. r. , lamberty , a. , peiser , h. s. , rvsz , k. , rieder , s. e. , rosman , k. j. r. , roth , e. , taylor , p. d. p. , vocke , jr .
, r. d. , & xiao y.k .
2002 , pure appl .
, 74 ( 10 ) , 1987 hasegawa , t. i. , herbst , e. , & leung , c. m. 1992 , , 82 , 167 + herczeg , g. j. , brown , j. m. , van dishoeck , e. f. , & pontoppidan , k. m. 2011 , , 533 , a112 + hidaka , h. , watanabe , n. , shiraki , t. , nagaoka , a. , & kouchi , a. 2004 , , 614 , 1124 van dishoeck , e. f. , helmich , f. p. , de graauw , t. , black , j. h. , boogert , a. c. a. , ehrenfreund , p. , gerakines , p. a. , lacy , j. h. , millar , t. j. , schutte , w. a. , tielens , a. g. g. m. , whittet , d. c. b. , boxhoorn , d. r. , kester , d. j. m. , leech , k. , roelfsema , p. r. , salama a. , & vandenbussche , b. 1996 , , 315 , l349 | this study reports an unusual heterogeneity in [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] abundance ratios of carbon monoxide observed in the gas phase toward seven @xmath3 solar - mass ysos and three dense foreground clouds in the nearby star - forming regions , ophiuchus , corona australis , orion , vela and an isolated core , l43 .
robust isotope ratios were derived using infrared absorption spectroscopy of the 4.7@xmath4 m fundamental and 2.3@xmath4 m overtone rovibrational bands of co at very high resolution ( @xmath5/@xmath6@xmath7 ) , observed with the crires spectrograph on the very large telescope .
we find [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] values ranging from @xmath3 85 to 165 , significantly higher than those of the local interstellar medium ( @xmath8 to 69 ) .
these observations are evidence for isotopic heterogeneity in carbon reservoirs in solar - type yso environments , and encourage the need for refined galactic chemical evolution models to explain the @xmath0c/@xmath2c discrepancy between the solar system and local ism .
the oxygen isotope ratios are consistent with isotopologue - specific photodissociation by co self - shielding toward the disks , vv cra n and hl tau , further substantiating models predicting co self - shielding on disk surfaces .
however , we find that co self - shielding is an unlikely general explanation for the high [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] ratios observed in this study .
comparison of the solid co against gas - phase [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] suggests that interactions between co ice and gas reservoirs need to be further investigated as at least a partial explanation for the unusually high [ @xmath0c@xmath1o]/[@xmath2c@xmath1o ] observed . |
in the traditional approach to quantum field theory , one starts from classical fields and a lagrangean which describes the interaction .
these objects get quantized and @xmath0-matrix elements or greens functions are constructed with the help of the feynman rules .
a typical example which is used in this paper as a model theory is the massless @xmath1-theory , where the interaction hamiltonian density is given by the normal ordered third order monomial of a free uncharged massless scalar field and a coupling constant @xmath2 @xmath3 and @xmath4 is an element of minkowski space , denoted by @xmath5 in the following .
the perturbative @xmath0-matrix is then constructed according to the expansion @xmath6 where @xmath7 is the time - ordering operator .
it must be pointed out that the perturbation series eq .
( [ stoer1 ] ) is formal and it is difficult to make any statement about the convergence of this series .
furthermore , two problems arise in the expansion given above .
first , the time - ordered products @xmath8 are usually plagued by ultraviolet divergences .
however , these divergences can be removed by regularization , such that the operator - valued distributions @xmath9 can be viewed as well - defined , already regularized expressions .
second , infrared divergences are also present in eq .
( [ stoer1 ] ) .
this is not astonishing , since the @xmath9 s are operator - valued distributions , and therefore must be smeared out by test functions in @xmath10 , the schwartz space of functions of rapid decrease .
one may therefore introduce a test function @xmath11 which plays the role of an adiabatic switching and provides a cutoff in the long - range part of the interaction , which can be considered as a natural infrared regulator @xcite .
then the infrared regularized @xmath0-matrix is given by @xmath12 and an appropriate adiabatic limit @xmath13 must be performed at the end of actual calculations in the right quantities ( like cross sections ) where this limit exists .
this is not one of the standard strategies usually found in the literature , however , it is the most natural one in view of the mathematical framework used in perturbative quantum field theory .
the vertex diagram appears in @xmath1-theory at third order of perturbation theory via @xmath14 @xmath15 where @xmath16 is given by the scalar feynman propagator discussed below in further detail .
of course , the massless vertex functions plays also a prominent role in applied quantum field theories like qcd , where it emerges in the three - gluon vertex @xcite , or in perturbative models of quantum gravity . in this paper ,
the fourier transform of the massless vertex distribution ( also called vertex function in the sequel ) @xmath17 @xmath18[(q - k)^2+i0][k^2+i0 ] } \label{vertextrafo}\ ] ] will be calculated in a non - standard manner by making explicit use of causality as a fundamental property of local quantum field theories @xcite .
the presented dispersive method is based on considerations in configuration space and integrals in momentum space are not evaluated by the help of wick rotations .
furthermore , no mass terms are needed as regulators of infrared divergences , the vertex function is constructed directly without any limiting procedure .
it should be noted that the vertex function is translation invariant in the sense that @xmath19 , where @xmath20 is an arbitrary four - vector .
this symmetry property can be inferred directly from the decomposition of the vertex function into three feynman propagators given above . introducing new integration variables @xmath21 and @xmath22 , the fourier transform in eq .
( [ vertextrafo ] ) can also be written as ( @xmath23 ) @xmath24 @xmath25 which shows that @xmath26 does indeed not depend on @xmath27 .
the focus of this work is on the causal and analytic properties of the vertex function and its infrared - finite dispersive calculation in the massless case .
the more involved general - mass case of the vertex function has been treated by t hooft and veltman @xcite , such that results for the massless case could be obtained by considering the appropriate limit .
an explicit result for the massless case has been given in @xcite , and a discussion of the vertex function within a special framework based on a geometrical interpretation of kinematic invariants and other quantities of one - loop diagrams is contained in @xcite , where a short overview over different calculational methods for one - loop n - point functions can also be found . before investigating the causal structure of the scalar vertex function , some important analytic properties of the feynman propagator and other related tempered distributions will be recalled in the following section .
the feynman propagator of a free scalar and charge neutral quantum field @xmath28 fulfilling the wave equation @xmath29 is given in configuration space by the vacuum expectation value @xmath30 @xmath31 where @xmath32 denotes principal value regularization and @xmath33 is the one - dimensional dirac distribution depending on @xmath34 @xcite .
knowledge of the basic properties of the feynman propagator is sufficient for many applications in perturbative quantum field theory . for the forthcoming discussion
, it is instructive to review some of the useful properties of basic tempered distributions appearing in quantum field theory .
a scalar neutral field @xmath35 can be decomposed into a negative and positive frequency part according to @xmath36\ ] ] with @xmath37=\delta^{(3)}(\vec{k}-\vec{k } ' ) , \quad [ a(\vec{k}),a(\vec{k}')]=[a^\dagger(\vec{k}),a^\dagger(\vec{k}')]=0.\ ] ] the commutation relations for such a field are given by the positive and negative frequency jordan - pauli distributions @xmath38
= -i \langle 0 | [ \phi^{\mp}(x ) , \ , \phi^{\pm}(0 ) ] | 0 \rangle \ , , \label{commutator}\ ] ] which have the fourier transforms @xmath39 the fact that the commutator @xmath40=i\delta^+(x)+i\delta^-(x ) = : i \delta(x)\ ] ] with the fourier transform @xmath41 vanishes for spacelike arguments @xmath42 due to the requirement of microcausality , leads to the important property that the jordan - pauli distribution @xmath43 has _ causal support _ , i.e. it vanishes outside the _ closed _ forward and backward light cone such that @xmath44 in the sense of distributions .
a further crucial observation is the fact that one can introduce the retarded propagator @xmath45 which coincides with @xmath46 on @xmath47 , i.e. @xmath48 holds for all test functions in the schwartz space @xmath49 with support @xmath50 .
one may write in configuration space @xmath51 and transform this splitting formula into a convolution in momentum space could be replaced by @xmath52 with an arbitrary vector @xmath53 inside the open forward light cone . ]
@xmath54 the fourier transform of the heaviside distribution @xmath55 can be calculated easily @xmath56 for the special case where @xmath57 is in the forward light cone @xmath58 , one can go to a lorentz frame where @xmath59 such that eq .
( [ convo ] ) becomes @xmath60 hence , for arbitrary @xmath61 , @xmath62 would be given by the _ dispersion relation _
@xmath63 however , the integral in eq .
( [ disprel ] ) is undefined in the massless case considered here .
this illustrates the fact that it is not possible in general to convert the product of two distributions into a convolution via a fourier transform , even though expression ( [ welldef ] ) describes a well - defined distribution .
here , one can circumvent this problem , e.g. , by introducing a mass for the field @xmath35 as a regulator for the corresponding massive jordan - pauli distribution which is also defined as a field commutator ( for a different method , see , e.g. , @xcite ) .
for the massive jordan - pauli distribution @xmath64 one gets in momentum space @xmath65 and the following expression for the massive retarded distribution @xmath66 ( @xmath67 ) @xmath68 @xmath69}{2 \sqrt{k^2 } m ( 1-t+i0 ) } = \frac{1}{k^2-m^2 } \quad ( k^2 \ne m^2 ) \ , .\ ] ] the @xmath33-distribution in the fourier transformed jordan - pauli distribution eq .
( [ ftmjp ] ) contains a mass term which is not present in eq .
( [ ftjp ] ) , such that @xmath70 does not exhibit a singular behavior on the light cone in momentum space .
as a special case of the edge of the wedge theorem @xcite it is known that the fourier transform of the retarded distribution @xmath71 is the boundary value of an analytic function @xmath72 , regular in @xmath73 @xcite .
this way one obtains from @xmath74 , @xmath75 , and @xmath76 @xmath77 the analytic expression for the feynman propagator is then recovered , and one obtains @xmath78 it is helpful for the forthcoming to understand eq .
( [ causal1 ] ) .
this is best done in configuration space .
one obviously has @xmath79 @xmath80 and @xmath81 @xmath82 and therefore eq . ( [ composition ] )
is fulfilled @xmath83 @xmath84 the discussion presented here can be applied the same way to second order loop diagrams @xcite .
the problem of ultraviolet divergences can be solved in such cases by using subtracted dispersion relations @xcite .
the causal properties of the jordan - pauli distribution and the retarded propagator can be generalized in an analogous manner to the case of the vertex function . for this purpose
, one defines the causal operator - valued three - point functions @xmath85 , @xmath86 and related c - number valued contractions @xmath87 and @xmath88 , which are related to the vertex diagram via @xmath89 @xmath90 @xmath91 @xmath92 @xmath93 @xmath94 and @xmath95 @xmath96+[t_1(x_2 ) , t_2(x_1,x_3 ) ] -[t_1(x_3 ) , t_2(x_1,x_2)]\ ] ] @xmath97 + [ t_1(x_3 ) , t_1(x_2)t_1(x_1)]\ ] ] @xmath98 where @xmath86 and @xmath88 have causal support as distributions in the sense that @xmath99 where @xmath100 therefore , only when _ both _ arguments @xmath101 and @xmath102 are in the forward light cone or _ both _ arguments @xmath101 and @xmath102 are in the backward light cone with respect to @xmath27 , @xmath103 belongs to the support of @xmath104 . for the sake of brevity ,
only a short outline how one can prove the special support property of @xmath86 is given here by considering the case where , e.g. , @xmath105 and @xmath106 . in this case
, one can assume that one is in a lorentz system where @xmath107 , since the transformation properties of all @xmath9 s are fixed by the representation of the lorentz group on the fock space of massless scalar particles .
then one has , e.g. , @xmath108 , and accordingly for the full expression eq .
( [ defdelta3 ] ) @xmath109 @xmath110 @xmath111 @xmath112 @xmath113 @xmath114 the same way one may check for all other cases that @xmath86 vanishes whenever the causal condition ( [ causalcond ] ) is not fulfilled .
the general definition of causal distributions @xmath115 at higher orders can be found in @xcite .
one should point out that @xmath27 plays a distinguished role in the discussion above . whereas the time - ordered distribution @xmath116 is invariant under a permutation @xmath117 of the arguments according to its definition in eqns .
( [ stoer1],[toprod],[smootheds ] ) , @xmath118 is not .
however , the construction presented above would go through with @xmath101 or @xmath102 as distinguished arguments .
the crucial point is that the causal support property of @xmath86 and its c - number part @xmath88 will enable the calculation of a distribution which is retarded with respect to @xmath27 via a double dispersion relation in momentum space , as will be shown in the following .
this retarded distribution @xmath119 can be related directly to the vertex function in analogy to eq .
( [ causal1 ] ) .
the vertex function again possesses the full permutation symmetry in configuration space by construction .
the vertex function will be constructed below by performing the following calculational steps .
the fourier transform @xmath120 , which is independent of @xmath27 due to translation invariance ( see also chapter 3.1 in @xcite ) , is calculated from the expression eq .
( [ defdelta3 ] ) in configuration space .
then the corresponding retarded distribution @xmath121 is evaluated in momentum space by a dispersion relation analogously to eq .
( [ disprel ] ) , and finally the vertex function is obtained directly from analytic considerations .
it is also straightforward to check that on the operator level , the time - ordered product @xmath122 can be obtained from the retarded ( appropriately regularized ) distribution @xmath123 @xmath124 i.e. for the vertex distribution holds @xmath125 in close analogy to eq .
( [ causal1 ] ) .
starting from the first and second order tree diagram given by @xmath127 @xmath128 one readily obtains for @xmath88 in configuration space by exploiting the relevant wick contractions in eq .
( [ defdelta3 ] ) @xmath129 -0.8 cm @xmath130 note that a wick contraction _ without _ time - ordering leads to a @xmath131-distribution instead of a feynman propagator . making use of @xmath132 and introducing the advanced distribution @xmath133 such that @xmath134 and @xmath135 , eq .
( [ delta3e ] ) can be simplified to @xmath136 @xmath137 @xmath138 performing the fourier transform , one obtains @xmath139 @xmath140 @xmath141 @xmath142 - \ { p \leftrightarrow q\}. \label{delta3momentum}\ ] ] from the integral eq . ( [ delta3momentum ] )
one immediately derives the following symmetry properties of @xmath126 @xmath143 @xmath144 as an illustration , the full calculation of the first fourier integral appearing in eq .
( [ delta3momentum ] ) @xmath145\ ] ] @xmath146 is presented here .
introducing @xmath147 and using a new integration variable @xmath148 , one obtains @xmath149 the first @xmath150 and @xmath33 distributions can be written as @xmath151 ( @xmath152 is replaced by @xmath57 in the following ) . taking into account the second @xmath150 distribution
one easily derives @xmath153 , otherwise integral above would vanish .
for spacelike @xmath32 , @xmath154 , there exists a lorentz frame with @xmath155 such that @xmath156 also vanishes . for timelike @xmath32
, there exists a lorentz frame where @xmath157 .
the last @xmath33 distribution in eq .
( [ int1 ] ) then implies @xmath158 , and combined with @xmath159 from the @xmath33 distribution @xmath160 one can replace @xmath161 by @xmath162 in the feynman propagator appearing in eq .
( [ int1 ] ) after performing the trivial integral over @xmath161 . for timelike @xmath32 , there remains to calculate with @xmath163 replaced by @xmath164 @xmath165},\ ] ] with @xmath57 in the denominator above given by @xmath166 , such that one can write @xmath167 and one has @xmath168 .
hence ( @xmath169 ) , @xmath170 @xmath171 @xmath172 the distribution @xmath173 inserted in eq .
( [ igen ] ) does not affect the result in the rest frame of @xmath32 , but will account below for the fact that @xmath174 vanishes for spacelike @xmath32 when written in covariant form @xmath175 .
investigating the behavior of the expression @xmath176 under lorentz transformations one finds that the covariant form of @xmath176 is given by @xmath177 , where @xmath178 is the triangle function @xmath179 such that one arrives at @xmath180 @xmath181 it is left to the reader as a simple exercise to show that lightlike @xmath32 do not contribute to @xmath174 in a distributional sense . calculating the remaining two integrals along the same lines as above and subtracting the resulting expressions with @xmath182 and @xmath183 interchanged , one obtains @xmath184 @xmath185 note that all imaginary parts of the integrals appearing in eq .
( [ delta3momentum ] ) mutually cancel , since @xmath126 is real due to its antisymmetry properties in configuration space described above .
in close analogy to eq .
( [ disprel ] ) , @xmath187 is calculated in the following for @xmath188 and @xmath189 . the result for arbitrary @xmath182
, @xmath183 can be obtained by analytic continuation , since @xmath187 is the boundary value of an analytic function , regular in @xmath190 , @xmath191 ( see also @xcite ) .
the procedure is in two steps .
first , the integral @xmath192 is evaluated for @xmath188 .
this corresponds to the calculation of @xmath193 in configuration space in the lorentz system where @xmath194 .
then @xmath187 is obtained by calculating @xmath195 corresponding to the multiplication of @xmath193 with @xmath196 in configuration space .
note that the well - known splitting formula @xcite @xmath197 is infrared divergent in the case of a totally massless vertex , while the double dispersion relation eq .
( [ doublesplit ] ) is not .
for the actual calculations one may start by considering the first logarithm in eq .
( [ delta3result ] ) @xmath198 the real part of the integral ( [ i1p ] ) can be written down immediately ( @xmath147 ) @xmath199=-\frac{\pi^2}{4(2 \pi)^3 } \frac{\mbox{sgn}(p^0 ) \theta(p^2)}{\sqrt{n } } \log \biggl| \frac{pq+\sqrt{n}}{pq-\sqrt{n } } \biggr|.\ ] ] the lorentz system shall be chosen such that @xmath200 . since @xmath188 ( and @xmath183 arbitrary )
, one has to evaluate the cauchy principal value @xmath201=-\frac{i \pi}{4 ( 2 \pi)^3 } \log \biggl| \frac{pq+\sqrt{n}}{pq-\sqrt{n } } \biggr| \int \limits_{-\infty}^{+\infty } dt \frac{\mbox{sgn}(tp^0-q^0 ) \theta(((tp^0,\vec{0})-q)^2)}{t(1-t)}. \label{analogo}\ ] ] from the @xmath150 distribution , one obtains integral limits via @xmath202 and taking into account that @xmath203 , one obtains from @xmath204 @xmath205 the first splitting result @xmath206 as further step one may calculate @xmath207 the terms generated by the @xmath208-term can be written down immediately @xmath209 the remaining non - trivial integrals are @xmath210 and @xmath211 integral @xmath212 can easily be solved by the help of the identity @xmath213 and one obtains @xmath214 i.e. @xmath212 cancels the second term in eq.([izeros ] ) .
integral @xmath215 has a structure completely analogous to eq .
( [ analogo ] ) and is given by @xmath216 such the compact result ( @xmath188 , @xmath189 ) @xmath217 is obtained .
the integrals which appear in the calculation of the remaining parts of @xmath218 , namely @xmath219 and @xmath220 , can all be expressed by the help of the spence function ( or dilogarithm ) @xmath221 , which is defined by @xmath222 where the integral contour in eq .
( [ spence ] ) must be equivalent to a straight line from @xmath223 to @xmath224 avoiding @xmath225 .
this is due to the fact that all relevant integrals can be reduced directly to the form @xmath226 @xmath227 where @xmath228 is the real spence function , defined for real arguments @xmath229 by @xmath230 for the sake of brevity , the results are presented here @xmath231 @xmath232
@xmath233 @xmath234\ ] ] @xmath235 , \label{i2split}\ ] ] and @xmath220 ist given by @xmath236 .
note that the last three logarithmic terms in eq .
( [ i2split ] ) are superfluous .
they indeed appear in the actual calculations , however , they combine to zero , since @xmath237 it must also be pointed out that the spence function obeys very many identities .
e.g. , the real spence function satisfies @xmath238 @xmath239 @xmath240 looking for a compact expression for @xmath186 , one finds @xmath241 @xmath242 @xmath243 . \label{retarded}\ ] ] this expression can be transformed with the help of eq .
( [ identities ] ) and the observation that the arguments of the spence function can be written in many different ways @xmath244 @xmath245 note that @xmath187 is purely imaginary for @xmath246 . this insight can also be obtained directly from the naive splitting formula eq .
( [ naivesplitting ] ) , where only the imaginary part is infrared divergent . there
, one obtains for the real part of @xmath187 ( see also eq .
( [ i2split ] ) ) @xmath247=0.\ ] ] the retarded distribution @xmath187 can be obtained for _ arbitrary _ @xmath182 , @xmath183 from eq .
( [ retarded ] ) by analytic continuation , since @xmath187 is the boundary value of the corresponding function analytic for @xmath248 and @xmath249 .
furthermore , the full vertex function is finally obtained from eqns .
( [ vertexconfig ] , [ 3composition ] ) @xmath250 with @xmath251 @xmath252 the result for @xmath253 is given here with the remark that @xmath253 and @xmath254 can indeed be calculated in parallel .
a detailed calculation of @xmath255 and @xmath126 can also be found in @xcite , chapter 3.8 , where the vertex function appearing in qed with massive electrons is treated .
the result for the massless case can be easily obtained by setting the electron mass equal to zero , since no infrared problems are present in @xmath255 and @xmath126 . for @xmath188 and @xmath189
, @xmath256 vanishes and one can finally write down in a sloppy style the result that can be found in the literature @xcite @xmath257[(q - k)^2][k^2]}\ ] ] @xmath258 .
\label{vresult}\ ] ] note that in the result given in @xcite , a factor of two is missing .
( [ vresult ] ) is also equivalent to the forms derived in @xcite . | the massless one - loop vertex diagram is constructed by exploiting the causal structure of the diagram in configuration space , which can be translated directly into dispersive relations in momentum space . |
cygnus x-2 is a bright persistent low - mass x - ray binary ( lmxrb ) , whose x - ray spectrum has been studied from @xmath40.1 to several hundred kev .
together with some of the other bright persistent lmxrb sources it is classified as a `` _ z _ '' source ( hasinger & van der klis @xcite ) .
the changes in the x - ray spectral shape of _ z _ sources are subtle , but in an x - ray color - color diagram the sources trace out _ z_-like shaped patterns .
they move through the _ z _ in a smooth manner without jumping from branch to branch .
_ z _ sources are thought to be accreting material at near - eddington rates via an accretion disk onto a neutron star ( e.g. hasinger et al .
1990 ) .
the italian - dutch satellite is the first mission to simultaneously observe in the 0.1300kev energy range using a complementary payload of instruments ( boella et al .
the low - energy concentrator spectrometer ( lecs ) is sensitive in the energy range kev ( parmar et al .
its unique design utilizes a driftless gas scintillation proportional counter to make the lowest energies accessible with a good energy resolution while providing 16@xmath6s time resolution and moderate spatial resolution .
the lecs has a circular field of view of 37@xmath7 diameter and a 0.110kev background counting rate of 9.7@xmath110@xmath8arcmin@xmath9s@xmath10 .
the lecs energy resolution , @xmath11e / e , is 19% at 1 kev and varies as e@xmath12 . in this _ letter _
results from the lecs are presented for one of the science verification phase targets , .
we find that the 0.210kev spectrum can be described by an absorbed cut - off power - law model , with additional emission at energies near @xmath41kev .
llll & & + & cut - off power - law & cut - off power - law & units + & & + gaussian & + n@xmath13 & @xmath14 & @xmath15 & 10@xmath2hatoms@xmath3 + @xmath16 & @xmath17 & @xmath18 & + e@xmath19 & @xmath20 & @xmath21 & kev + k & @xmath22 & @xmath23 & photons@xmath3s@xmath10kev@xmath10 at 1kev + e@xmath24 & & @xmath25 & kev + fwhm & & @xmath26 & kev + ew & & @xmath27 & ev + @xmath28/dof & 1.45/925 & 1.20/922 & +
observed between 1996 july 23 00:19 and 18:18 utc .
good data were selected from intervals when the minimum elevation angle above the earth s limb was @xmath294@xmath30 and when the instrument configuration was nominal using the saxledas 1.4.0 data analysis package ( lammers @xcite ) .
the lecs was only operated during satellite night - time giving a total on - source exposure of 7.1ks .
a spectrum was extracted centered on the mean source position using the standard lecs extraction radius of 8@xmath7 and the appropriate response matrix was generated .
background subtraction was performed using a standard blank field 46ks exposure , but is not critical for such a bright source .
the spectrum was rebinned to have at least 20 counts per channel .
channels below 0.2kev were discarded since the source is absorbed below this energy .
the mean lecs count rate observed from is 67s@xmath10 , which corresponds to an observed 110kev flux of 7.4@xmath110@xmath31erg@xmath3s@xmath10 .
the ground calibration of the lecs is discussed in parmar et al .
the effective area @xmath322 kev is primarily limited by the entrance window transmission , and at higher energies by the loss of reflectivity of the mirror system .
great care was taken to determine the entrance window transmission using monochromatic x - ray sources ( bavdaz et al .
the lecs spectral response was updated following the first observation of the crab nebula on 1996 september 6 .
when the lecs spectrum is fit by an absorbed power - law model , the individual energy channel residuals are @xmath325%
. the calibration status of the instruments , derived from the above observation , is discussed in cusumano et al .
( 1997 ) .
the spectrum can be modeled by a cut - off power - law ( @xmath33 ) ) together with low energy absorption , n@xmath13 , which gives a @xmath34 of 1.45 for 925 degrees of freedom ( dof ) , see also table 1 .
examination of the residuals shows excess counts near 1kev which are unlikely to originate from calibration uncertainties or other instrumental effects .
these excess counts can be modeled by a blackbody of temperature 0.181@xmath35 kev giving a @xmath34 of 1.25 for 923 dof .
however , previous observations of have required much higher blackbody temperatures of 1.01.5 kev ( e.g. hasinger et al .
1990 ) , and it is therefore unlikely that this feature is the blackbody previously observed from the source . replacing the blackbody with a broad gaussian line feature with an energy , e@xmath24 , of @xmath25 kev , a full width at half - maximum , fwhm , of [email protected] and an equivalent width , ew , of ( 74@xmath5)ev in the fit gives a better fit with a @xmath34 of 1.20 for 922 dof .
f_-test indicates that this difference is significant , but only at the 1@xmath36 level .
higher energy resolution observations than that reported here ( e.g. using the _ einstein _ objective grating spectrometer ( ogs ) ; see vrtilek et al . 1988 ) , require the presence of multiple line features near 1 kev .
this supports the interpretation of the excess emission seen in the lecs as unresolved line emission .
table 1 gives the best - fit parameters for the gaussian line model .
1 shows the observed count spectrum and the best - fit model when the normalization of the emission feature is set to zero .
in addition to the feature at @xmath41kev , there is evidence for weak narrow - line features at [email protected] , [email protected] , and [email protected] in the lecs spectrum ( see fig . 1 ) .
including all three lines in the fit reduces the @xmath28 to 1.15 for 916 dof .
the energies of two of these lines are consistent with ly@xmath37 transitions of h - like si xiv and s xvi at 2.00 and 2.62kev , respectively ( see e.g. raymond 1993 ) .
however , we caution that these features have energies close to those of the mirror au edges , where the lecs calibration is more uncertain
. there is no evidence for gaussian line emission near 6.7kev , with a 3@xmath36 upper limit on the equivalent width of 62ev for a narrow line .
vrtilek et al . (
1988 ) have compiled the results of _ einstein _ ogs and medium - resolution solid state spectrometer ( sss ) observations of .
the ogs data indicate the presence of four narrow emission lines between 0.74 and 1.12kev ( see table 2 ) . if the lower resolution sss data is fit with the same set of line energies , the results are broadly consistent with the ogs results except that the ew of each of the lines is a factor @xmath42 larger ( vrtilek et al .
1988 ) . in order to investigate whether the excess emission seen by the lecs is consistent with the same blend of lines as seen by _
einstein _ ,
the broad gaussian emission feature was replaced by four narrow lines at the energies given in table 2 .
the resulting @xmath28 is 1.24 for 921 dof .
there is no evidence for the presence of the 0.74kev fe xvii and 0.77kev o viii / fe xviii features with 95% confidence upper limits on the ew of 14 and 16ev , respectively .
the ew of the 0.96kev fe xx / ni xx feature of 14@xmath03ev is similar to the values seen by the ogs and sss of 12@xmath05 and 30@xmath010ev , respectively .
however , the ew of the 1.12kev line of 24@xmath03ev is larger than measured by both the ogs ( 5@xmath02ev ) and sss ( 10@xmath03ev ) .
we caution that the narrow - line fit accounts for only about half the line flux observed , suggesting the presence of other emission features in this energy range . llccc line & line & + energy & identification & ogs & sss & lecs + ( kev ) & & ( ev ) & ( ev ) & ( ev ) + 0.74 & fe xvii & @xmath38 & @xmath39 & @xmath4014 + 0.77 & o viii / fe xviii & @xmath41 & @xmath39 & @xmath4016 + 0.96 & fe xx / ni xx & @xmath42 & @xmath43 & @xmath44 + 1.12 & fe xvii/ & @xmath45 & @xmath39 & @xmath46 + & fe xxiii xxiv & + + +
models for the continua of many bright lmxrb sources often require at least two components , e.g. a blackbody and a cut - off power - law model ( white et al .
a blackbody component is not required to fit the lecs spectrum .
the absence of the blackbody component when is on the upper left part of its horizontal branch of the _ z _ has been noted by kuulkers et al .
( @xcite ) .
the blackbody component appears to be present on all other parts of the _ z _ ( hasinger et al .
@xcite ; kuulkers et al .
@xcite ) .
although the timing properties can in principle be used to determine the position of in its _ z _ variability pattern , there are insufficient counts in the lecs data to uniquely determine it .
the intensity of is normally a factor of @xmath42 higher ( e.g. hasinger et al .
1990 ; kuulkers et al . 1995 ) than during the observation .
such low intensities are only reached during the so - called `` low intensity '' states of ( et al .
inspection of the _ rossi _ x - ray timing explorer all sky monitor ( asm ) light curves ( wijnands et al .
1996 ) indicates that the observation occured 510 days before entered a probable low intensity state , suggesting that the source was transitioning from a medium to a low intensity state .
we note that a similar transition may have been seen by vrtilek et al .
( 1986 ; their `` medium state b '' ) .
vrtilek et al .
( 1986 ) fit the 120kev _ einstein _ monitor proportional counter ( mpc ) spectra obtained during this state with a thermal bremsstrahlung model of kt @xmath47.08.7kev and n@xmath13 @xmath42.33.5@xmath47hatoms@xmath3 .
a comparison with the best - fit parameters obtained when the same model is fit to the 110kev lecs spectrum of [email protected] and @[email protected]@xmath49@xmath3 indicates that the spectral shape was similar on both occasions .
the presence of excess emission near 1.0kev was first suggested following a rocket flight in 1971 ( bleeker et al .
@xcite ) . as well as the ogs and sss results reported in vrtilek et al .
( 1986 , 1988 ; see also kallman et al .
@xcite and smale et al .
@xcite ) , branduardi - raymont et al .
( @xcite ) using the _
ariel v _ experiment c , chiappetti et al .
( @xcite ) using the exosat channel multiplier array , lum et al .
( @xcite ) using the _ einstein _ focal plane crystal spectrometer , smale et al .
( @xcite ) using the broad band x - ray telescope and smale et al .
( @xcite ) using the _ asca _ solid state imaging spectrometer , all report evidence for excess emission near 1kev .
these reports indicate that the strength of the excess emission varies from observation to observation .
no excess emission near 1kev was reported by pravdo ( @xcite ) , hirano et al .
( @xcite ) , and predehl & schmitt ( @xcite ) using lower energy resolution data obtained with the heao1a2 , _ hakucho _ proportional counter , and _ rosat _ position sensitive proportional counter instruments , respectively .
line emission at energies near 1kev is not unique to .
similar features have been observed in other lmxrb , e.g.in the _ z _ source ( see vrtilek et al.@xcite ) , and in the accreting pulsars ( mccray et al .
1982 ; oosterbroek et al .
1997 ) and ( angelini et al . 1995 ; owens et al .
the feature is most probably due to a combination of unresolved fe l - shell and ne k - shell line emission ( chiappetti et al . @xcite ; vrtilek et al .
@xcite ; lum et al .
@xcite ; smale et al .
@xcite , @xcite ; angelini et al .
vrtilek et al .
( 1988 ) modeled the @xmath41kev excess emission seen from with _
einstein _ using four narrow emission lines which they identify with fe xvii , o viii / fe xvii , fe xx / ni xx and fe xvii / fe xxii xxiv features ( see also kallman et al .
the lecs spectrum was fit with the same model , and we find that the fe xvii and o viii / fe xvii features are not present ( although the 95% confidence upper limits are consistent with the _ einstein _ detections ) , the ew of the fe xx / ni xx feature is similar to the values reported in vrtilek et al .
( 1988 ) , and the ew of the fe xvii / fe xxii xxiv feature is significantly greater ( see table 2 ) .
these differences imply that the shape of the feature , as well as its overall ew , varies from observation to observation .
it is notable that the narrow line model only accounts for about half the line flux of the excess emission as seen with the lecs , implying the presence of other features near 1kev .
the fe - l emission may be produced by photoionization of the surface of the accretion disk and the accretion disk corona by the strong x - ray continuum flux emanating from the central regions ( see e.g. kallman et al .
@xcite ; raymond @xcite ) .
recent calculations by kallman ( @xcite ) show that the ratio of the equivalent widths of the fe - l and fe - k lines should be close to unity for .
our measurements indicate that this ratio is @xmath321.2 ( 3@xmath36 ) , consistent with these calculations .
the strength of the low - energy feature is expected to be dependent on the temperature and density of the illuminated corona ( e.g. lidahl @xcite ; kallman @xcite ) , and probably indirectly on the mass accretion rate onto the neutron star which probably influences the x - ray continuum and shape of the inner accretion disk . since these parameters can vary from observation to observation
, it is not surprising that the strength and shape of the excess emission near 1kev is variable .
angelini l. , white n.e . , nagase f. , et al . , 1995 ,
apj 449 , l41 bavdaz m. , peacock a. , parmar a.n .
, et al . , 1994 .
inst . & meth .
res . a 345 , 549 bleeker j.a.m .
, deerenberg a.j.m . , yamashita k. , hayakawa s. , tanaka y. , 1972 , apj 178 , 377 boella g. , butler r.c .
, perola g.c .
, et al . , 1997 , a&as 122 , 299 branduardi - raymont g. , chiapetti l. , ercan e.n . , 1984 ,
a&a 130 , 175 chiappetti l. , treves a. , branduardi - raymont g. , et al . , 1990 ,
apj 361 , 596 cusumano g. , dal fiume d. , giarrusso s. , et al . , 1997 , a&a submitted hasinger g. , van der klis m. , 1989 , a&a 225 , 79 hasinger g. , van der klis m. , ebisawa k. , dotani t. , mitsuda k. , 1990 , a&a 235 , 131 hirano t. , hayakawa s. , kunieda h. , 1984 , pasj 36 , 769 kallman t.r . , 1995 ,
apj 455 , 603 kallman t.r . ,
vrtilek s.d . ,
kahn s.m . , 1989 , apj 345 , 498 kuulkers e. , van der klis m. , van paradijs j. , 1995 , apj 450 , 748 kuulkers e. , van der klis m. , vaughan b.a . , 1996 , a&a 311 , 197 lammers u. , 1997 , `` the sax / lecs data analysis system - software user manual '' , esa / ssd , sax / leda/0010 liedahl d.a .
, kahn s.m . ,
osterheld a.l . ,
goldstein w.h . , 1990 , apj 350 , l37 lum k.l . , canizares c.r . , clark g.w . , et al
, 1992 , apjs 78 , 423 mccray r.a , shull j.m .
, boynton p.e . , et al .
, 1982 apj 262 , 301 oosterbroek t. , parmar a.n .
, martin d.d.e .
, lammers u. , 1997 , a&a , submitted owens a. , oosterbroek t. , parmar a.n . , 1997 ,
a&a , submitted parmar a.n . , martin d.d.e . , bavdaz m. , et al .
, 1997 a&as 122 , 309 pravdo s.h . , 1983 ,
apj 270 , 239 predehl p. , schmitt j.h.m.m . , 1995 , a&a 293 , 889 raymond j.c . , 1993 ,
apj 412 , 267 smale a.p .
, done c. , mushotzky r.f . , et al . , 1993 ,
apj 410 , 796 smale a.p . ,
angelini l. , white n.e .
, mitsuda k. , dotani t. , 1994 , baas 185 , 1484 vrtilek s.d . ,
, grindlay j.e . ,
helfand d.j . , seward f.d . , 1986 apj 307 , 698 vrtilek s.d . ,
swank j.h . ,
kallman t.r . , 1988 ,
apj 326 , 186 vrtilek s.d .
, mcclintock j.e . , seward f.d . ,
kahn s.m . ,
wargelin b.j . , 1991 , apjs 76 , 1127 white n.e . , stella l. , parmar a.n . , 1988 , apj 324 , 363 wijnands r.a.d . , kuulkers e. , smale a.p . , 1996 ,
apj 473 , l45 | we present a 0.210kev spectrum of the low - mass x - ray binary cygnus x-2 obtained using the low energy concentrator spectrometer on - board .
the spectrum can be described by a cut - off power - law model with absorption of ( [email protected])@xmath110@xmath2atoms@xmath3 , a power - law index of [email protected] and a cut - off energy of [email protected] ( 68% confidence errors ) , except at energies near @xmath41kev where excess emission is present .
this can be modeled by a broad gaussian line feature with an energy of [email protected] , a full width half - maximum of [email protected] and an equivalent width of 74@xmath5ev .
this result confirms earlier reports of line emission near 1kev and shows the intensity and structure of the feature to be variable . |
we continue the study of semistable extremal neighborhoods pioneered by mori , kollr , and prokhorov @xcite . our motivation is to compute explicitly stable one - parameter degenerations of smooth canonically polarized surfaces and to describe their moduli space introduced by kollr and shepherd - barron @xcite .
we work throughout over @xmath3 .
extremal neighborhoods are fundamental building blocks of the minimal model program approach to @xmath0-dimensional birational geometry @xcite . the goal of this program is to construct a canonical representative in the birational class of a smooth ( or mildly singular ) proper @xmath0-fold by gradually forcing the canonical divisor to intersect curves non - negatively . on each step of the program
, one has to analyze the local picture of a @xmath0-fold around a curve which has negative intersection with the canonical divisor .
concretely , let @xmath4 be the germ of a @xmath0-fold along a proper reduced irreducible is allowed to be reducible , however we can always reduce to the irreducible case by factoring @xmath5 locally analytically over @xmath6 , see @xcite .
] curve @xmath7 such that @xmath4 has terminal singularities .
we say @xmath4 is an _ extremal neighborhood _ if there is a germ @xmath6 of a @xmath0-fold and a proper birational morphism @xmath8 such that @xmath9 , @xmath10 ( as sets ) and @xmath11 .
then @xmath12 .
the exceptional locus ( the union of positive - dimensional fibers ) of @xmath5 is either equal to @xmath7 or is a divisor with image a curve passing through @xmath13 .
we say @xmath5 is a _ flipping contraction _ in the first case and a _ divisorial contraction _ in the second .
if @xmath5 is a divisorial contraction then @xmath6 is a terminal singularity ( * ? ? ?
* th 1.10 ) .
if one is running the minimal model program , @xmath14 will ( locally ) give its next step .
if @xmath5 is a flipping contraction then one has to construct the flip .
the _ flip _ of @xmath5 is a germ @xmath15 of a @xmath0-fold along a proper reduced curve @xmath16 such that @xmath15 has terminal singularities , and a proper birational morphism @xmath17 such that @xmath18 , @xmath19 ( as sets ) , and @xmath20 .
the existence of the flip was proved by mori ( * ? ? ?
0.4.1 ) and uniqueness follows from @xmath21 cf .
6.4 ) . to classify extremal neighborhoods , following miles reid , we choose a general member @xmath22 .
let @xmath23 .
then @xmath24 and @xmath25 are normal surfaces with at worst du val singularities ( * ? ? ?
we say the extremal neighborhood @xmath26 is _ semistable _ if @xmath27 is a du val singularity of type @xmath28 @xcite .
semistable extremal neighborhoods are of two types @xmath2 and @xmath1 .
we say @xmath5 is of type @xmath2 or @xmath1 if the number of singular points of @xmath24 equals one or two respectively ( * ? ? ?
* , p.542 ) . in the case of a flipping contraction , @xmath5 is semistable iff the general hyperplane section through @xmath6 is a cyclic quotient singularity @xmath29 ( @xcite , corollary 3.4 and appendix ) , i.e. , is ( locally analytically ) isomorphic to the quotient of @xmath30 by @xmath31 acting by @xmath32 , where @xmath33 is a primitive root of unity .
we are primarily interested in semistable neighborhoods due to applications to compact moduli spaces of surfaces .
let @xmath34 be a general element .
consider the corresponding morphism @xmath35 and the induced morphism @xmath36 .
write @xmath37 thus we can view an extremal neighborhood as a total space of a flat family of surfaces .
let @xmath38 , @xmath39 , denote a nearby smooth fiber .
in this paper we assume that the second betti number @xmath40 our motivation for this assumption is as follows .
let @xmath41 be gieseker s moduli space of canonical surfaces of general type with given numerical invariants @xmath42 and @xmath43 .
let @xmath44 be its compactification , the moduli space of stable surfaces of kollr , shepherd - barron , and alexeev . given a map @xmath45 from a punctured smooth curve germ , i.e. a family @xmath46 of canonical surfaces , one is interested in computing its `` stable limit '' @xmath47 , perhaps after a finite base change .
in many important cases the family @xmath46 can be compactified by a flat family @xmath48 , where the special fiber @xmath49 is irreducible , normal , and has quotient singularities .
computing the stable limit then amounts to running the relative minimal model program for @xmath48 .
it turns out that we can always reduce to the case of flips and divisorial contractions of type @xmath2 and @xmath1 with @xmath50 ( in an analytic neighborhood of a contracted curve ) after a finite surjective base change by passing to a small partial resolution using simultaneous resolution of du val singularities ( see e.g. ( * ? ? ?
* th . 4.28 ) and @xcite ) .
see section [ s4 ] for more details . building on @xcite , we show in section [ s1 ] that if @xmath4 is an extremal neighborhood of type @xmath2 ( resp .
@xmath1 ) with @xmath50 then @xmath49 is normal and has one ( resp .
two ) cyclic quotient singularities along @xmath7 of the following special form @xmath51 for some @xmath52 , @xmath53 , a so - called _
wahl singularity_. in @xcite mori gave an explicit algorithm for computing flips of extremal neighborhoods of type @xmath1 .
we extend his result to neighborhoods of type @xmath2 .
the novelty of our approach is that instead of considering extremal neighborhoods one - by - one , we construct their universal family ( see section [ s2 ] ) , and in particular show that @xmath2 neighborhoods belong to the same deformation family as @xmath1 ( see proposition [ k1adegeneratestok2a ] , whose proof in given at the end of [ k1atok2a ] ) . for any extremal neighborhood of type @xmath2 or @xmath1 with @xmath50 , in section [ s2 ]
we define a four - tuple of integers @xmath54 and the following data which depends only on this four - tuple : a positive integer @xmath55 , a toric surface @xmath56 and a toric birational morphism @xmath57 which depend only on @xmath55 ( @xmath56 is only locally of finite type for @xmath58 ) , flat irreducible families of surfaces @xmath59 and morphisms @xmath60 such that the following holds .
let @xmath61 be an extremal neighborhood of type @xmath2 or @xmath1 with @xmath50 and with associated four - tuple @xmath54 .
then there is a morphism @xmath62 such that @xmath63 and @xmath64 is the pull - back of @xmath65 under @xmath66 . in other words
, @xmath67 is the universal family of extremal neigborhoods .
if @xmath5 is a flipping contraction then the flip @xmath68 is the pullback of @xmath69 under @xmath70 .
[ thmintro ] we give an explicit combinatorial algorithm to compute all @xmath2 and @xmath1 corresponding to this family ( see propositions [ k1ainvariants ] and [ k1adegeneratestok2a ] ) .
the construction of the universal family , which is the content of section [ s2 ] , follows very closely the division algorithm in @xcite , which one can interpret as a sequence of mutations in the cluster algebra of rank @xmath71 with general coefficients ( see remark [ cluster ] ) . if @xmath72 the morphism @xmath73 is the blowup of @xmath74 . if @xmath75 , @xmath56 is the toric variety associated to the fan @xmath76 with support @xmath77 where @xmath78 .
it has rays spanned by @xmath79 for @xmath80 given by @xmath81 , @xmath82 , @xmath83 for @xmath84 , and @xmath85 , @xmath86 , @xmath87 for @xmath84 . over torus orbits of dimension @xmath88 ( resp .
@xmath89 ) corresponding to rays of @xmath76 ( resp .
@xmath71-dimensional cones ) , fibers of @xmath90 are surfaces with one ( resp .
two ) wahl singularities . in the flipping case , the special fiber @xmath91 is a germ of a cyclic quotient singularity @xmath92 and the special fiber @xmath93 is its _ extremal p - resolution _ ,
i.e. , a partial resolution @xmath94 such that @xmath95 has only wahl singularities @xmath96 for @xmath97 , the exceptional curve @xmath98 , and @xmath99 is relatively ample .
let @xmath100 , and @xmath101 .
the wahl singularities @xmath102 and @xmath103 are minimally resolved by chains of smooth rational curves @xmath104 and @xmath105 respectively , such that @xmath106 , @xmath107 , @xmath108 , @xmath109 , and @xmath110 .
let @xmath111 be a surface germ around a @xmath98 such that @xmath95 has these two wahl singularities , @xmath16 passes through both of them , and in the minimal resolution of @xmath95 , the proper transform of @xmath16 is a @xmath112-curve which only intersects @xmath113 and @xmath114 , and transversally at one point . with this data
we build a @xmath115 as theorem [ thmintro ] .
we have @xmath116 . with this
we construct the universal antiflip @xmath117 , where the central fiber of @xmath118 is a surface germ with one singularity @xmath119 , and the corresponding birational toric morphism @xmath120 defined by the morphism between fans shown in figure [ f0 ] . in this figure ,
the rays @xmath121 , @xmath122 with @xmath123 in the fan @xmath76 are decorated by the data @xmath124 of the associated @xmath2 antiflip , which has as central fiber a surface with one singularity @xmath125 .
two consecutive @xmath126 or @xmath127 with @xmath128 represents a @xmath1 antiflip , which has two singularities , one from each of the @xmath2 of the rays .
these @xmath2 s are recovered from the @xmath1 by @xmath129-gorenstein smoothing up one of the singularities while preserving the other .
when @xmath130 tends to infinity , the rays @xmath131 and @xmath122 approach to two lines with irrational slopes @xmath132 and @xmath133 , where @xmath134 the rays @xmath135 and @xmath136 correspond to the wahl singularities @xmath103 and @xmath102 of @xmath95 respectively .
.,width=453 ] in section [ s2 ] , following @xcite , we show that an extremal p - resolution of a cyclic quotient singularity gives a unique family @xmath137 of ( terminal ) antiflips .
in addition we identify , in the versal deformation space of this cyclic quotient singularity @xmath95 , the locus of deformations such that the total space admits a terminal antiflip , see corollary [ existenceofterminalantiflips ] .
specifically , we show that a smoothing @xmath138 admits a terminal antiflip if and only if there is a divisor @xmath139 such that @xmath140 is the toric boundary of @xmath95 for some choice of toric structure and the axial multiplicities @xmath141 of the singularities of @xmath15 satisfy @xmath142 .
this raises the following question , which we find very interesting : classify smoothings of cyclic quotient singularities which admit an antiflip ( not necessarily terminal ) .
see also remark [ asflnsfgajf ] and example [ asdvasfgafg ] of a canonical antiflip . in [ 2extremalnbhd ] we classify extremal p - resolutions .
we show , by a combinatorial argument built on continued fractions representing zero @xcite ( see [ contfracpresol ] ) , that any cyclic quotient singularity admits at most two extremal p - resolutions ( and thus at most two families of antiflips ) .
moreover , we prove in [ delta ] that these two families must have the same parameter @xmath55 , i.e. , they live over the same toric surface @xmath56 . at the end of [ contfracpresol ] ,
we give a list with the first cyclic quotient singularities ( up to order @xmath143 ) admitting one or two extremal p - resolutions .
finally , in section [ s4 ] , we give an application to moduli of stable surfaces , which was the original motivation for this paper .
we start by constructing a stable surface which admits a @xmath144-gorenstein smoothing to a canonically polarized smooth surface with @xmath145 and @xmath146 , following ideas of lee and park @xcite .
the surface @xmath49 has two wahl singularities @xmath147 and @xmath148 , and no local - to - global obstructions to deform ( i.e. deformations of its singularities globalize to deformations of @xmath49 ) .
the kollr shepherd - barron alexeev moduli space of stable surfaces near @xmath49 is two - dimensional and has two boundary curves , which correspond to @xmath144-gorenstein deformations @xmath149 and @xmath150 which smoothen @xmath151 ( resp .
@xmath152 ) and preserve the remaining singularity .
we address the question of describing minimal models of @xmath153 and @xmath154 .
namely , we consider deformations @xmath155 and @xmath156 of surfaces obtained by simultaneously resolving @xmath151 ( resp .
@xmath152 ) and run our explicit mmp to find new limits @xmath157 of each family .
this allows us to conclude that @xmath158 is a rational surface and @xmath159 is a dolgachev surface of type @xmath160 , i.e. , an elliptic fibration over @xmath161 with two multiple fibers of multiplicity @xmath71 and @xmath0 , respectively .
more systematic study of the boundary of the moduli space of surfaces of geometric genus @xmath89 will appear elsewhere .
paul hacking was supported by nsf grants dms-0968824 and dms-1201439 .
jenia tevelev was supported by nsf grants dms-1001344 and dms-1303415 .
giancarlo urza was supported by the fondecyt inicio grant 11110047 funded by the chilean government .
we would like to thank i. dolgachev , j. kollr , s. mori , and m. reid for helpful discussions .
in particular , m. reid explained to us in detail his joint work with g. brown which describes explicitly the graded ring of a @xmath1 flip @xcite .
we recall that the minimal resolution of a cyclic quotient singularity of type @xmath29 has exceptional locus a nodal chain of smooth rational curves with self - intersection numbers @xmath162 where @xmath163= b_1 - \frac{1}{b_2-\frac{1}{\ddots -\frac{1}{b_r}}},\quad b_i \ge 2 \mbox { for all $ i$}\ ] ] is the expansion of @xmath164 as a hirzebruch
jung continued fraction .
see @xcite , p. 46 .
moreover , the minimal resolution can be described torically as follows .
identify the singularity @xmath165 with the germ @xmath166 of the affine toric surface corresponding to the cone @xmath167 in the lattice @xmath168 define @xmath169 for @xmath170 by @xmath171 and @xmath172 \mbox { for $ 2 \le i \le r$},\ ] ] @xmath173 \mbox { for $ 1 \le i \le r-1$}.\ ] ] define @xmath174 then each @xmath131 is a primitive vector in the lattice @xmath175 .
the minimal resolution @xmath176 is the toric proper birational morphism corresponding to the subdivision @xmath76 of the cone @xmath177 given by adding the rays @xmath178 generated by the @xmath131 for @xmath179 .
let @xmath180 denote the exceptional curve corresponding to the ray @xmath181 .
then @xmath180 is a smooth rational curve such that @xmath182 .
define @xmath183 let @xmath184 denote the strict transform of @xmath185 .
then the toric boundary of @xmath186 is the nodal chain of curves @xmath187 .
moreover , we have @xmath188 and @xmath189 ( indeed , since @xmath180 corresponds to the primitive vector @xmath190 , the orders of vanishing of @xmath191 and @xmath192 along @xmath180 equal @xmath193 and @xmath194 respectively . )
finally the discrepancies @xmath195 defined by @xmath196 are given by @xmath197 ( indeed , for any toric variety @xmath49 with boundary @xmath198 we have @xmath199 .
thus in our case @xmath200 and @xmath201 . now the formula for the discrepancies follows from the formulas for the pullback of @xmath202 and @xmath203 above . )
the wahl singularity is a cyclic quotient singularity of the form @xmath204 for some @xmath52 , @xmath53 .
the number @xmath205 is the index of the singularity .
we have an identification @xmath206 given by @xmath207 the deformation @xmath208 of @xmath165 is of the form @xmath209 for some @xmath210 and convergent power series @xmath211 , @xmath212 .
it follows that we have an isomorphism of germs @xmath213 the number @xmath210 is called the _ axial multiplicity _ of @xmath214 ( * ? ? ?
* def . 1a.5 , p. 140 ) .
in particular , the @xmath0-fold germ @xmath4 is analytically @xmath144-factorial by @xcite .
the following proposition is mostly a combination of results from @xcite and @xcite .
[ b2=1=>normal ] let @xmath61 be an extremal neighborhood of type @xmath2 or @xmath1 .
let @xmath215 be a general element and let @xmath216 be the corresponding morphism .
let @xmath217 denote the special fiber and @xmath38 , @xmath218 , a nearby fiber .
then @xmath50 iff @xmath49 is normal and has the following description .
if @xmath5 is of type @xmath1 then @xmath49 has two singular points @xmath219 , @xmath97 . for each @xmath97
there are numbers @xmath220 , @xmath221 , @xmath222 , and an isomorphism of germs @xmath223 if @xmath5 is of type @xmath2 then @xmath49 has one singular point @xmath224 .
there are numbers @xmath225 , @xmath226 , and an isomorphism @xmath227 for some @xmath228 such that @xmath229 the proper birational morphism @xmath230 can be described as follows .
if @xmath5 is of type @xmath1 then @xmath49 is normal and has two singular points @xmath219 , @xmath97 , of index greater than @xmath88 . for each @xmath97
there are numbers @xmath231 , @xmath221 , @xmath222 , and an isomorphism of germs @xmath232 see ( * ? ? ?
* rk . 2.3 ) .
if @xmath5 is of type @xmath2 then @xmath49 is not necessarily normal .
if @xmath49 is normal then @xmath49 has one singular point @xmath224 of index greater than @xmath88 and possibly an additional singular point @xmath233 of index @xmath88 .
there are numbers @xmath234 , @xmath226 , @xmath235 and an isomorphism @xmath236 for some @xmath228 such that @xmath237 if @xmath238 then @xmath239 is the unique singular point .
if @xmath240 we have an additional singular point @xmath241 and an isomorphism @xmath242 ( so in particular @xmath241 is a du val singularity of type @xmath243 . ) see ( * ? ?
1.8.1 ) .
if @xmath49 is normal then we have @xmath244 indeed , @xmath7 is irreducible by assumption , the link of a cyclic quotient singularity is a lens space , and the milnor fiber @xmath56 of a @xmath144-gorenstein smoothing of a singularity of type @xmath245 has milnor number @xmath246 . now the result follows from a mayer vietoris argument .
so @xmath50 iff @xmath247 as claimed .
it remains to show that if @xmath49 is not normal then @xmath248 .
we use the classification of the non - normal case given in @xcite , theorem 1.8.2 .
let @xmath249 denote the normalization of @xmath49 and @xmath250 the inverse image of @xmath7 .
then @xmath251 has two components @xmath252 meeting in a single point @xmath253 .
we have an isomorphism @xmath254 for some @xmath255 .
there are also points @xmath256 , @xmath257 and isomorphisms @xmath258 and @xmath259 for some @xmath260 . the non - normal surface @xmath49 is formed from @xmath261 by identifying the curves @xmath262 and @xmath263 .
the points @xmath264 and @xmath265 are identified to a point @xmath266 and we have an isomorphism @xmath267 moreover the image @xmath268 of @xmath253 is a degenerate cusp singularity ( cf .
* def . 4.20 ) ) and @xmath49 has normal crossing singularities along @xmath269 .
let @xmath270 be the minimal resolution .
since @xmath271 is a rational singularity ( by @xcite , corollary 3.4 ) , @xmath272 is a resolution of @xmath273 , and @xmath274 is a smoothing of @xmath273 , we have @xmath275 see @xcite , 4.1c .
we remark that @xmath42 is well - defined in our situation ( see e.g. @xcite ) , and @xmath276 is the second betti number as usual .
the morphism @xmath277 is a proper birational morphism of smooth surfaces , hence a composition of blowups , so @xmath278 also @xmath279 because @xmath216 is @xmath144-gorenstein , so we obtain @xmath280 we use this equation to show that @xmath248 if @xmath49 is non - normal .
write @xmath281 for the strict transform of @xmath282 .
the exceptional locus of @xmath283 is a disjoint union of @xmath0 chains of smooth rational curves @xmath284 , @xmath285 , and @xmath286 ( the exceptional loci over the singularities @xmath287 , @xmath288 , and @xmath253 ) such that the curves @xmath289 form a nodal chain .
we have @xmath290 for some @xmath291 .
write @xmath292 and @xmath293 . by lemma [ torick^2 ]
below , we have @xmath294 @xmath295 now since the cyclic quotient singularities @xmath287 and @xmath288 are conjugate , that is , of the form @xmath296 and @xmath297 for some @xmath298 , we have the identity @xmath299 ( proof : by induction using the following characterization of conjugate singularities .
we identify a cyclic quotient singularity @xmath296 with the associated continued fraction @xmath300 $ ] , @xmath301 .
then @xmath302,[2]$ ] is a conjugate pair , and if @xmath303 $ ] , @xmath304 $ ] is a conjugate pair , so is @xmath305 $ ] , @xmath306 $ ] . ) so we obtain @xmath307 thus @xmath308 we also compute @xmath309 @xmath310 @xmath311 where we have used the adjunction formula @xmath312 for @xmath313 . combining , we obtain @xmath314 by @xcite , theorem 1.8.2 , we have @xmath315 so @xmath316 now @xmath317 and @xmath318 , so combining we obtain @xmath319 now ( [ mpbound ] ) gives @xmath320 .
the first formula of the following lemma is well - known , see e.g. @xcite . [ torick^2 ]
let @xmath165 be a cyclic quotient singularity of type @xmath29 .
let @xmath176 be the minimal resolution of @xmath49 .
write @xmath321 $ ] , @xmath322 for all @xmath130 .
let @xmath323 denote the inverse of @xmath324 modulo @xmath325 .
then @xmath326 let @xmath327 be a curve such that @xmath328 let @xmath329 be the strict transform of @xmath330
. then @xmath331 write @xmath332 and let @xmath184 be the strict transform .
let @xmath284 be the exceptional curves of @xmath90 , a nodal chain of smooth rational curves with self - intersection numbers @xmath162 , such that @xmath333 meets @xmath113 and @xmath334 meets @xmath335 .
we have @xmath336 for some @xmath337 , @xmath338 .
the exceptional curves @xmath113 and @xmath335 correspond to the rays in the fan @xmath76 of @xmath186 generated by @xmath29 and @xmath339 , so @xmath340 and @xmath341 .
now we have @xmath342 @xmath343 where we have used the adjunction formula @xmath344 . in the second case , we have @xmath345 where @xmath346 in particular , @xmath347 .
so @xmath348 @xmath349 [ k1ainvariants ] let @xmath61 be an extremal neighborhood of type @xmath2 with @xmath50 as in proposition [ b2=1=>normal ] , in particular @xmath350 for some @xmath228 such that @xmath229 let @xmath176 denote the minimal resolution of @xmath49 .
thus the exceptional locus of @xmath90 is a nodal chain of smooth rational curves @xmath284 with self - intersection numbers @xmath351 , ... ,
@xmath352 where @xmath353.\ ] ] let @xmath354 denote the strict transform of @xmath7 .
then @xmath355 is a @xmath112-curve which intersects a single component @xmath180 of the exceptional locus of @xmath90 transversely in one point .
moreover , identifying @xmath356 with the germ of the affine toric variety given by the cone @xmath167 in the lattice @xmath357 the exceptional divisor @xmath180 is extracted by the toric proper birational morphism given by subdividing @xmath177 by the ray @xmath358 define @xmath359 , @xmath360 , @xmath361 for each @xmath362 , by @xmath363=m_0/(m_0-a_0)\ ] ] if @xmath364 , @xmath365 if @xmath366 , and @xmath367=m_2/(m_2-a_2)\ ] ] if @xmath368 , @xmath369 if @xmath370 .
the surface germ @xmath371 is isomorphic to a cyclic quotient singularity of type @xmath92 , where @xmath372= \delta/\omega,\ ] ] equivalently , @xmath373 we also have @xmath374 and @xmath375 the description of the minimal resolution together with the strict transform @xmath251 of @xmath7 is given by @xcite , theorem 1.8.1 .
contracting @xmath251 , we obtain a resolution of @xmath376 with exceptional locus a chain of smooth rational curves with self - intersection numbers @xmath377 hence @xmath376 is a cyclic quotient singularity of type @xmath92 where @xmath378.\ ] ] given @xmath379 with @xmath380 , @xmath381 , write @xmath163 , \quad b_i \ge 2 \mbox { for all $ i$}\ ] ] for the hirzebruch
jung continued fraction .
define @xmath382 by @xmath383 the continued fraction corresponds to the factorization in @xmath384 @xmath385 see @xcite , proposition 2.1 . now , assuming @xmath386 , the equalities @xmath363=m_0/(m_0-a_0)\ ] ] @xmath367=m_2/(m_2-a_2)\ ] ] @xmath387=m_1 ^ 2/(m_1a_1 - 1)\ ] ] give @xmath372=\delta/\omega,\ ] ] where @xmath388 @xmath389 so @xmath390 and @xmath391 as claimed .
the cases @xmath392 are proved similarly . from the discussion in
[ toric ] , the discrepancy @xmath393 in the formula @xmath394 is given by @xmath395 now @xmath396 note that @xmath397 is cartier so @xmath398 is an integer .
the matrix factorization @xmath399 @xmath400 implies that @xmath401 which gives @xmath402 and @xmath403 the equation implies that @xmath404 adding and gives @xmath405 finally , combining and gives .
[ k1adegeneratestok2a ] let @xmath406 be as in proposition [ k1ainvariants ] .
we retain the notations @xmath407 and @xmath408 , @xmath409 .
define proper birational morphisms @xmath410 for @xmath411 as follows .
we identify @xmath371 with the germ of the affine toric variety given by the cone @xmath167 in the lattice @xmath412 define @xmath413 let @xmath414 be the toric proper birational morphism corresponding to the subdivision @xmath415 of @xmath177 given by the ray @xmath416 , and @xmath417 the exceptional curve of @xmath414 .
let @xmath418 denote the torus fixed points corresponding to the maximal cones @xmath419 , @xmath420 of @xmath415 respectively .
then @xmath421 has cyclic quotient singularities of types @xmath422 moreover , @xmath423 the one parameter deformation of @xmath424 given by the trivial deformation of the @xmath425 singularity and the versal @xmath144-gorenstein smoothing of the remaining singularity of @xmath421 ( with the target @xmath371 being fixed ) has general fiber isomorphic to the morphism @xmath406 .
we postpone the proof until the end of
[ k1atok2a ] , since it is a consequence of the construction of the universal family .
we follow @xcite closely .
let @xmath426 be such that @xmath427 and @xmath222 for each @xmath97 , and 1 .
@xmath428 , 2 . @xmath429 .
write @xmath430 @xmath431 define an action of @xmath432 on @xmath433 by @xmath434 @xmath435 define @xmath436 and @xmath437 write @xmath438 then @xmath439 is a open covering of @xmath440 and we have identifications @xmath441 for each @xmath97 . here
the coordinates @xmath442 are the restrictions of the coordinates @xmath443 on @xmath433 to the @xmath444 cover of @xmath445 given by setting @xmath446 . similarly for @xmath447 .
the glueing @xmath448 of @xmath445 and @xmath449 is given by @xmath450 ( the expression for @xmath451 in @xmath449 is determined by the equation of @xmath445 . )
write @xmath452 and @xmath453 for each @xmath97
. then @xmath454 is the point @xmath455 for each @xmath97 , and @xmath7 is a proper smooth rational curve in @xmath440 with local equations @xmath456 for each @xmath97 .
write @xmath457 for each @xmath97 .
the morphism @xmath458 is a flat family of surfaces ( flatness is clear in the charts @xmath445 and @xmath449 ) .
write @xmath459 for the fiber over @xmath74 .
then @xmath49 is a toric surface .
write @xmath460 then @xmath461 and we have an identification @xmath462 given by @xmath463 the glueing @xmath464 is given by @xmath465 the surface @xmath49 is the toric variety associated to the fan @xmath76 in the lattice @xmath466 defined as follows .
let @xmath467 be vectors in @xmath175 .
let @xmath76 be the fan in @xmath175 with cones @xmath468 , @xmath469 , @xmath470 , @xmath471 , @xmath472 and @xmath473 .
the ray @xmath181 corresponds to the divisor @xmath474 for each @xmath97 and the ray @xmath475 corresponds to the curve @xmath476
. let @xmath477 @xmath478 denote the usual orientation of @xmath479 . note that @xmath480 by assumption .
so the support @xmath481 of the fan @xmath76 is the convex cone @xmath482 .
let @xmath273 denote the affine toric surface associated to the cone @xmath177 in @xmath175 and let @xmath483 denote the torus fixed point .
we compute @xmath484 where @xmath485 @xmath486 thus @xmath487 where @xmath488
we have a toric proper birational morphism @xmath489 with exceptional locus the curve @xmath476 .
let @xmath490 denote the cokernel of the homomorphism of abelian groups @xmath491 then @xmath492 and @xmath493 [ classgrouplemma ] the equality @xmath494 induces an identification @xmath495 such that @xmath496 \mapsto \bar{e}_i \in m_{\gamma}= { \mathbb z}^2 / { \mathbb z}(m_1,-m_2)\ ] ] for each @xmath97 . write @xmath497 then @xmath498 in particular @xmath499 is an irreducible divisor .
we have @xmath500 where @xmath501 is an irreducible divisor for each @xmath97 .
note that @xmath502 in particular @xmath503 and @xmath504 . now @xmath505 and @xmath506 for @xmath97 , so @xmath507 and @xmath508 write @xmath509 for the quotient map
. we can define a group homomorphism @xmath510 \mapsto \chi\ ] ] as follows .
write @xmath511 some @xmath512 . then , for each @xmath513 , we have @xmath514 for some @xmath515 . the action of the group @xmath432 on @xmath516 has finite stabilizers and is free in codimension @xmath88 .
it follows that the homomorphism ( [ classgroupmap ] ) is an isomorphism with inverse @xmath517 where we regard @xmath518 as a one dimensional representation of @xmath432 and we identify @xmath519 \in \operatorname{cl}({\mathbb x})$ ] with the isomorphism class of the rank one reflexive sheaf @xmath520 .
the quotient map @xmath521 is identified with the spectrum of the ( relative ) cox ring @xmath522 where cox@xmath523 \in \operatorname{cl}({\mathbb x } ) } { \mathcal o}_{\mathbb x}(d)$ ] , via @xmath524 strictly speaking , in order to define the cox ring we need to choose a representative @xmath525 of each class @xmath526 and compatible isomorphisms @xmath527 in our case we take representatives @xmath528 for @xmath529 a list of coset representatives for the quotient @xmath490 and isomorphisms derived from the isomorphism @xmath530 assume without loss of generality that @xmath531 .
define a sequence @xmath532 as follows .
define @xmath533 for @xmath84 , if @xmath534 is defined and @xmath535 , define @xmath536 [ termination ] there exists @xmath537 such that @xmath538 and @xmath539 .
let @xmath540 be the number defined by lemma [ termination ] .
define @xmath541 and @xmath542 if @xmath543 then @xmath544 and we have @xmath545 @xmath546 in this case we define @xmath534 for @xmath547 by requiring @xmath548 for each @xmath547 . then @xmath549 if @xmath58 then we define @xmath534 for @xmath550 and @xmath551 by requiring that @xmath552 if @xmath58 then @xmath553 for @xmath554 .
define a sequence @xmath555 as follows .
@xmath556 if @xmath543 define @xmath557 and define @xmath558 for @xmath547 by requiring @xmath559 for each @xmath547 . if @xmath560 define @xmath558 for @xmath547 by requiring @xmath561 if @xmath562 then @xmath563 for all @xmath564 .
we define a sequence @xmath565 for @xmath547 as follows .
define @xmath566 \in \operatorname{cl}({\mathbb x } ) , \quad l_2=[s_1 ] \in \operatorname{cl}({\mathbb x}).\ ] ] define @xmath567 for @xmath568 by @xmath569 define @xmath570 if @xmath571 then @xmath544 and @xmath566 , \quad l_2=[s_1 ] , \quad l_3= l_2 - l_1 , \quad l_4=-l_2 , \quad l_5= l_1-l_2.\ ] ] we define @xmath567 for @xmath547 by requiring @xmath572 for each @xmath547 .
then @xmath573 if @xmath58 define @xmath567 for @xmath574 and @xmath551 by requiring that @xmath575 we define a sequence @xmath576 for @xmath547 .
define @xmath577 we have @xmath576 for @xmath97 by the definitions , and for @xmath578 by the definition ( [ action ] ) of the @xmath432-action and lemma [ classgrouplemma ] .
note that the equations of @xmath433 can be rewritten @xmath579 @xmath580 define @xmath581 for @xmath582 by @xmath583 define @xmath584 by @xmath585 and @xmath586 by @xmath587 if @xmath571 then @xmath544 and one checks that @xmath588 .
we define @xmath581 for all @xmath547 by requiring @xmath589 for all @xmath547 .
if @xmath562 then we define @xmath581 for @xmath590 and @xmath551 by requiring @xmath591 we can rewrite the equations defining the @xmath581 as follows : @xmath592 where @xmath593 $ ] are defined as follows .
if @xmath72 then @xmath594 @xmath595 and @xmath596 , @xmath597 for each @xmath547 . if @xmath598 then @xmath599 and @xmath600 note in particular that @xmath601 are coprime monomials in @xmath602 $ ] for each @xmath130
moreover , we have @xmath603 in the terminology of @xcite the algebra @xmath604[\{f_i \ | \
i \in { \mathbb z}\}]\ ] ] is a _
cluster algebra _ of rank @xmath71 , with clusters @xmath605 and coefficient group @xmath606 the group of monomials in @xmath607 . in the notation of @xcite , example 2.5 , p. 502
, the exchange relations are given by @xmath608 and coefficients @xmath601 as defined above .
in particular the case @xmath543 is a cluster algebra of finite type associated to the root system @xmath609 .
[ cluster ] for each @xmath547 we have @xmath576 .
@xcite , theorems 3.10 , 3.12 , and 3.13 . ) from the definitions it is clear that @xmath581 is a rational section of @xmath567 for each @xmath547 .
we must show that it is a regular global section , or , equivalently , @xmath610 .
we say @xmath611 are _ coprime on @xmath440 _ if @xmath612 has codimension @xmath71 .
define statements @xmath613 for @xmath547 as follows .
* @xmath576 .
* @xmath581 and @xmath614 are coprime on @xmath440 .
* @xmath581 and @xmath615 are coprime on @xmath440 .
the following implications are immediate from the equations @xmath616 1 . for each @xmath547 ,
@xmath617 2 . for each @xmath618 , we have 1 .
@xmath619 2 .
@xmath620 indeed , for ( 1 ) we have @xmath621 similarly for @xmath622 . for ( 2a ) , we have @xmath623 using @xmath624 for @xmath625 .
case ( 2b ) is similar .
we show @xmath626 for each @xmath547 .
we have @xmath627 also , let @xmath628 denote congruence modulo @xmath581 in @xmath629 , then @xmath630 @xmath631 @xmath632 @xmath633 so @xmath634 .
now @xmath581 and @xmath635 are coprime on @xmath440 , so by we have @xmath636 is regular in codimension @xmath88 on @xmath516 .
hence @xmath636 is regular on @xmath516 because @xmath433 is normal .
similarly @xmath637 is regular on @xmath516 .
we verify @xmath638 for @xmath639 and @xmath640 for @xmath409 by direct computation .
now by induction using the implications above we prove @xmath641 for @xmath642 , @xmath643 for @xmath644 and @xmath640 for @xmath645 .
we claim that @xmath646 and @xmath647 hold .
( note that the implication ( 2b ) was not established for @xmath648 so can not be used here . )
observe that the loci @xmath649 are irreducible divisors .
so it suffices to prove that @xmath650 and @xmath651 do not contain any of these divisors .
but by lemma [ restrict_to_h ] below the intersection @xmath652 of these divisors is not contained in @xmath650 or @xmath651 .
this establishes the claim .
finally we deduce that @xmath613 hold for all @xmath547 by induction .
[ restrict_to_h ] define a sequence @xmath653 by @xmath654 , @xmath655 , and @xmath656 for @xmath657 .
( then @xmath658 , and @xmath659 unless @xmath370 or @xmath543 and @xmath660 . )
then @xmath661 in particular , set - theoretically we have @xmath662 and @xmath663 with equality for @xmath664 .
@xcite , ( 3.2 ) , p. 172 . ) the description of @xmath665 and @xmath666 for @xmath667 are verified by direct computation .
we establish the description of @xmath666 for @xmath668 by induction using the equations @xmath669 the final statements follow using @xmath670 .
define a sequence @xmath671 as follows .
define @xmath672 define @xmath673 for @xmath568 by @xmath674 we also define @xmath675 if @xmath543 then @xmath544 so @xmath676 @xmath677 we define @xmath673 for @xmath547 by requiring @xmath678 . if @xmath58 we define @xmath673 for @xmath547 by requiring @xmath679 ( compare with the definition of the sequence @xmath534 , @xmath547 . )
define @xmath680 by @xmath681 and if @xmath682 define @xmath683 by @xmath684 define @xmath685 and @xmath686 define an action of @xmath687 on @xmath688 by @xmath689 @xmath690 define @xmath691 and @xmath692 write @xmath693 then @xmath694 , @xmath695 for each @xmath97 , and the glueing is given by @xmath696 @xmath697 write @xmath698 @xmath699 and @xmath700 for each @xmath97 .
then @xmath701 is a toric surface , @xmath702 is a proper smooth rational curve , and @xmath703 are the torus fixed points of @xmath701 .
we also write @xmath704 for each @xmath97 .
define @xmath705 note that @xmath706 has codimension @xmath71 , and @xmath688 is normal .
thus the coordinate ring of the affine variety @xmath118 is given by @xmath707 where @xmath708\ ] ] and @xmath709 we have a natural morphism @xmath710 define a birational toric morphism @xmath711 via @xmath712 define a morphism @xmath713 over @xmath714 using the following diagram @xmath715^{\pi^ * } \ar@{^{(}->}[r ] & \gamma(w',\o_{w ' } ) \ar[d ] & \ar[l ] \ar[dl]^{\phi } r ' \\ \gamma({\mathbb x},\o_{{\mathbb x}})=\text{cox}({\mathbb x})^{\gamma } \ar@{^{(}->}[r ] & \text{cox}({\mathbb x } ) & } \ ] ] by @xmath716 note that by construction the equations of @xmath688 correspond to the relations @xmath717 and @xmath718 between the @xmath719 .
also , the group @xmath720 is identified with @xmath721 via @xmath722 , [ s_2]\ ] ] and @xmath723 here we are using the case @xmath724 in the second formula of : for @xmath725 we have @xmath726 and an identification @xmath727 cf .
@xcite , propositions 3.14 and 3.15 , p. 174 .
it follows that @xmath90 is well defined .
[ contractinggmaction ] there exists a @xmath728-action on @xmath440 , @xmath714 , @xmath729 , @xmath118 , and @xmath730 such that 1 .
the morphisms @xmath731 , @xmath732 , @xmath733 , @xmath90 , and @xmath734 are @xmath728-equivariant , and 2 .
@xmath728 fixes the points @xmath735 , @xmath736 , and @xmath737 and in each case acts with positive weights on the corresponding maximal ideal in the affine coordinate ring .
we define the @xmath728 action on @xmath731 by @xmath738 recall the equalities @xmath739 , @xmath740 , @xmath741 , and @xmath742 . using the equations @xmath592
we find that @xmath581 is a @xmath728-eigenfunction for each @xmath547 .
( the equations are seen to be homogeneous for the @xmath728-action by induction on @xmath130 using the equalities @xmath743 . )
moreover , the weight of @xmath581 for @xmath744 is given by @xmath745 @xmath746 and @xmath747 cf .
lemma [ restrict_to_h ] .
recall that @xmath748 , @xmath749 , @xmath750 , @xmath751 , and @xmath752 , @xmath753 .
so we obtain a compatible @xmath728-action on @xmath754 , and an induced action on @xmath732 .
we observe that @xmath728 acts with positive weights on @xmath714 , @xmath730 , and @xmath118 as claimed .
( note that if @xmath544 then @xmath728 acts with weight @xmath89 on @xmath755 , but no power of @xmath756 is @xmath687-invariant . ) assume @xmath757 .
then @xmath758 , @xmath759 , and we have an identification @xmath760 ( cf .
@xcite , theorem 4.5 , p. 178 . )
mori s proof works verbatim . in particular , in the case @xmath757 the family @xmath732
is independent of @xmath761 , that is , it is the pullback of a family @xmath762 under the projection @xmath763 . for the moment we will keep the redundant variable @xmath761 in order to treat both cases @xmath757 and @xmath764 simultaneously .
[ universalflip ] assume @xmath764 .
the morphism @xmath765 is a proper birational morphism with exceptional locus @xmath766 .
moreover @xmath767 or , equivalently , @xmath768 , where @xmath769 is the self - intersection of the proper transform of @xmath251 in the minimal resolution of @xmath701 .
@xcite , theorem 4.7 , p. 178 . )
the morphism @xmath770 is clearly birational .
we have @xmath771 for some @xmath772 , for each @xmath97 and @xmath773 . recalling that @xmath774 we deduce @xmath775 .
it follows that @xmath734 is proper over a neighbourhood of @xmath737 , and hence over @xmath118 using the existence of a contracting @xmath728-action on @xmath776 , see lemma [ contractinggmaction ] .
we compute @xmath777 using the toric description of the section @xmath701 .
observe that @xmath778 and @xmath251 is the only proper curve contained in @xmath779 .
( indeed the fibers of @xmath780 are obtained from the special fiber @xmath781 ( the toric boundary of @xmath701 ) by smoothing a subset of the two nodes . )
now since @xmath782 is relatively ample over @xmath118 ( using @xmath783 , @xmath784 , and the contracting @xmath728-action ) it follows that the exceptional locus of @xmath734 equals @xmath251 . 1 .
assume @xmath764 .
the morphism @xmath90 is a proper birational morphism with exceptional locus the codimension 2 set @xmath785 we have @xmath786 or explicitly @xmath787 2 .
assume @xmath757 .
the morphism @xmath90 is a proper birational morphism with exceptional locus the irreducible divisor @xmath788 we have @xmath789 ( cf .
@xcite , theorem 4.3 , p. 177 . )
it is easy to see that @xmath90 is birational .
now consider the base change @xmath790 and the morphism @xmath791 .
let @xmath792 denote the origin .
if @xmath764 then the element @xmath793 is @xmath687-invariant for some @xmath794 , for each @xmath97 and @xmath773 . if @xmath757 then the elements @xmath795 are @xmath687 invariant for some @xmath796 .
it follows that @xmath797 @xmath798 now @xmath799 and @xmath800 by lemma [ restrict_to_h ] ( using @xmath801 , @xmath802 , @xmath803 , @xmath804 ) .
thus @xmath805 .
it follows that there is an open neighbourhood @xmath806 such that @xmath90 is proper over @xmath807 .
moreover , by lemma [ contractinggmaction ] we have a contracting @xmath728-action on @xmath808 .
it follows that @xmath90 is proper .
suppose @xmath764 .
then @xmath809 and @xmath810 by @xcite , proposition 3.14(2 ) , p. 174 .
thus @xmath811 and @xmath812 are relatively ample over some open neighbourhood @xmath806 ( because @xmath813 ) . using the contracting @xmath728-action again we deduce that @xmath814 and @xmath812 are relatively ample over @xmath815 .
now @xmath816 where @xmath817 , @xmath818 .
so the exceptional locus of @xmath90 is contained in @xmath819 .
we have @xmath820 by lemma [ restrict_to_h ] .
also , @xmath819 is preserved by the contracting @xmath728-action ( because @xmath821 and @xmath822 are eigenfunctions for the @xmath728-action ) . moreover ,
each fiber of @xmath819 over @xmath714 is either empty or irreducible of dimension @xmath88 .
thus every fiber of @xmath823 is algebraically equivalent to a multiple of @xmath7 .
it follows that the exceptional locus of @xmath90 is equal to @xmath819 .
similarly , if @xmath757 , then @xmath812 is relatively ample over @xmath815 , and @xmath824 satisfies @xmath825 by lemma [ restrict_to_h ] ( using @xmath826 ) .
also , the locus @xmath819 is preserved by the contracting @xmath728-action .
it follows that each fiber of @xmath823 is a proper curve , such that each component is algebraically equivalent to a multiple of @xmath7 .
so the exceptional locus of @xmath90 equals @xmath819 .
finally , since the milnor number of a @xmath144-gorenstein smoothing of a singularity of type @xmath125 equals @xmath89 , we have @xmath827 for all @xmath828 .
it follows that the fibers of @xmath823 are irreducible , and hence that @xmath819 is irreducible .
see @xcite , theorem 4.5 , p. 178
for the identification of @xmath829 .
we identify the exceptional locus @xmath819 explicitly in the case @xmath764 of flipping contractions .
[ exceptional_locus_local_eq ] assume @xmath764 .
the locus @xmath830 is given in the charts @xmath831 as follows .
we have @xmath832 for @xmath544 and @xmath833 for @xmath834 .
write @xmath835 for @xmath97
. then @xmath836 @xmath837 and similarly , for @xmath834 , @xmath838 @xmath839 in particular , the nonempty fibers of @xmath823 are irreducible of dimension @xmath88 .
first observe that @xmath840 implies @xmath841 using the equations @xmath717 and @xmath718 .
hence @xmath832 if @xmath544 and @xmath833 if @xmath834 .
recall that @xmath842 by lemma [ restrict_to_h ] . in proposition [ glueingfamilies ]
we show that , for @xmath834 , the restriction of the morphism @xmath843 to the open subset @xmath844 is identified with the restriction of the morphism @xmath845 associated to the data @xmath846 to the open subset @xmath847 thus it suffices to describe @xmath848 .
let @xmath849 .
consider the fibers @xmath850 @xmath851 define closed irreducible curves @xmath852 as follows .
@xmath853 @xmath854 @xmath855 @xmath856
@xmath857 @xmath858 now , similarly to lemma [ restrict_to_h ] , one checks @xmath859 thus , noting that @xmath860 , we obtain @xmath861 let @xmath863 be such that @xmath864 .
explicitly , if @xmath58 then we require @xmath865 if @xmath764 and @xmath866 if @xmath757 . if @xmath543 we require @xmath867 if @xmath764 and @xmath868 if @xmath757 .
we will call the allowed values of @xmath869 _ admissible_. in what follows the superscripts @xmath869 are understood modulo @xmath870 in the case @xmath543 . define @xmath871 thus @xmath872 and @xmath873 for each @xmath97 .
we have @xmath874 and @xmath875 for each @xmath97 and all admissible @xmath869 .
moreover , @xmath876 and @xmath877 we apply the above constructions to the data @xmath878 , and denote the result by the same notations decorated by a superscript @xmath869 .
[ glueingfamilies ] for each admissible @xmath869 the family @xmath879 is identified with @xmath754 by identifying variables with the same name ( forgetting the superscript ) , that is , @xmath880 in particular we have an induced identification @xmath881 .
let @xmath882 be admissible .
then we have an identification @xmath883 given by @xmath884 @xmath885 and @xmath886 where the subscript @xmath130 is understood modulo @xmath870 if @xmath543 . using the above identifications we obtain the following : 1 . a toric surface @xmath56 ( only locally of finite type for @xmath562 ) together with a toric birational morphism @xmath887 such that @xmath56 is the union of the toric open affine sets @xmath888 for @xmath869 admissible and the restriction of @xmath73 to @xmath888 is the morphism @xmath889 2 . a proper birational morphism @xmath890 such that @xmath891 and the restriction of @xmath892 to @xmath893 is the morphism @xmath894 the toric birational morphism @xmath895 can be described explicitly as follows
we identify @xmath730 as the toric variety associated to the cone @xmath167 in the lattice @xmath466 .
we define a fan @xmath76 in @xmath175 with support @xmath481 contained in @xmath177 such that the toric birational morphism corresponds to the map of fans @xmath896 .
if @xmath543 , define @xmath897 , @xmath898 , and @xmath899 .
let @xmath76 be the fan with cones @xmath468 , @xmath900 , @xmath901 , and @xmath902 , @xmath97 .
( so in the case @xmath72 the morphism @xmath73 is the blowup of @xmath736 . )
if @xmath560 , define primitive vectors @xmath903 for @xmath80 by @xmath904 @xmath905 and @xmath906 @xmath907 ( so @xmath908 where @xmath909 . )
let @xmath76 be the fan with cones @xmath468 , @xmath910 for @xmath80 , @xmath911 for @xmath912 , and @xmath913 for @xmath912 . for @xmath560
define @xmath914 that is , @xmath132 is the larger root of the quadratic equation @xmath915 equivalently , @xmath132 is given by the infinite hirzebruch - jung continued fraction @xmath916 then the support @xmath481 of the fan @xmath76 is the region @xmath917 or , equivalently , @xmath918 let @xmath919 denote the exceptional locus of @xmath892 ( thus @xmath920 ) .
write @xmath921 and @xmath922 .
[ universality ] let @xmath61 be an extremal neighborhood of type @xmath2 or @xmath1 .
let @xmath923 be the morphism determined by a general element @xmath215 .
assume that the general fiber @xmath38 of @xmath216 satisfies @xmath50 .
then there is a universal family @xmath924 as in proposition [ glueingfamilies ] and a morphism @xmath925 such that @xmath926 and the diagram @xmath927 is isomorphic to the pullback of the diagram @xmath928 under @xmath66 .
if @xmath5 is a flipping contraction and @xmath17 is the flip of @xmath5 then the diagram @xmath929 is identified with the pullback of the diagram @xmath930 under @xmath70 .
furthermore , if @xmath931 is a general element , then we may assume that @xmath330 is the pullback of @xmath932 .
( then , if @xmath5 is a flipping contraction , the strict transform @xmath933 is the pullback of @xmath934 . )
we note that the cone @xmath935 corresponds to an extremal neighborhood of type @xmath1 with singularities @xmath936 if it is of flipping type , then @xmath937 . in this case
one of the singularities of the surface @xmath701 , which is the fiber over @xmath938 of @xmath754 , is @xmath939 ( see [ contraction ] ) .
similarly for the cone @xmath940 .
in this way , both singularities of @xmath701 can be read from these initial " @xmath1 .
( it is possible that @xmath941 . in that case
the corresponding point in @xmath701 is smooth . )
[ initialk2a ] [ locallyversal ] the flat family of surfaces @xmath942 induces a versal @xmath144-gorenstein deformation of each fiber .
the flat families of pairs @xmath943 and @xmath944 induce a versal @xmath144-gorenstein deformation of each fiber . by openness of versality and the existence of the contracting @xmath728-actions ( lemma [ contractinggmaction ] ) it suffices to consider the toric fibers .
we will establish the statement for pairs first .
we prove it for @xmath945 , the other case being identical .
let @xmath946 be a toric chart and @xmath947 the toric fiber over @xmath735 .
the logarithmic tangent sheaf @xmath948 is the sheaf of infinitesimal automorphisms of the pair @xmath947 .
since @xmath947 is toric we have an identification @xmath949 where @xmath175 is the lattice of @xmath88-parameter subgroups of the torus acting on @xmath49 .
thus @xmath950 for @xmath951 .
it follows that the local - to - global map from the versal @xmath144-gorenstein deformation space of the pair @xmath947 to the product of the @xmath144-gorenstein versal deformation spaces of the singularities @xmath952 , @xmath97 of the pair is an isomorphism .
recall also that a singularity @xmath953 of type @xmath125 has versal @xmath144-gorenstein deformation space a smooth curve germ @xmath954 , with versal family @xmath955 moreover , writing @xmath956 for the toric boundary , the forgetful map @xmath957 from the versal @xmath144-gorenstein deformation space of the pair @xmath958 to the versal @xmath144-gorenstein deformation space of @xmath953 is an isomorphism .
( indeed @xmath198 is cartier on the index one cover @xmath959 given by @xmath960 and the @xmath144-gorenstein deformation is induced by an equivariant deformation of the index one cover .
so there are no obstructions to lifting @xmath198 to the deformation . moreover , since @xmath961 , it s easy to see that every infinitesimal embedded deformation of @xmath198 is induced by an infinitesimal automorphism of @xmath962 . combining
it follows that the forgetful map ( [ forgetdivisor ] ) is an isomorphism . )
now it is clear from the explicit charts @xmath445 , @xmath449 for the restriction of the family @xmath942 to @xmath946 that @xmath945 induces a versal @xmath144-gorenstein deformation of the pair @xmath947 .
finally , we prove versality for @xmath942 . by the above , it suffices to show that @xmath963 for each toric fiber @xmath49 , so that the local - to - global map from the versal @xmath144-gorenstein deformation space of @xmath49 to the product of the versal @xmath144-gorenstein deformation spaces of the singularities of @xmath49 is an isomorphism .
let @xmath964 be the orbifold stack associated to @xmath49 , that is , @xmath964 is the deligne
mumford stack with coarse moduli space @xmath49 determined by the local smooth covers @xmath965 at each quotient singularity of @xmath49 .
write @xmath966 and @xmath967 , @xmath968 for the associated stacks
. then we have an exact sequence of sheaves on @xmath964 @xmath969 ( where @xmath970 denotes the normal bundle of a closed embedding @xmath499 ) . pushing forward via the natural morphism @xmath971
, we obtain an exact sequence @xmath972 where @xmath973 .
so , since @xmath974 for @xmath951 ( see above ) and @xmath975 are affine , we have @xmath976 and @xmath977 .
finally , we compute that the degree of the line bundle @xmath978 on @xmath12 equals @xmath979 , so @xmath980 as required . in general ,
if @xmath981 is a line bundle on a proper orbifold curve @xmath982 , with coarse moduli space @xmath983 , and orbifold points @xmath984),\ ] ] where @xmath985 and the action of @xmath986 on the fiber @xmath987 is via the character @xmath988 , @xmath989 , then @xmath990 recall that we have isomorphisms @xmath223 hence the degree of the line bundle @xmath978 on @xmath12 is given by @xmath991 using the formula @xmath992 ( see @xcite , proof of proposition 2.6 ) , we find @xmath993 .
thus @xmath980 as required .
[ k1atok2a ] the computation of the singularities of @xmath421 and the intersection number @xmath994 are straightforward toric calculations .
( note that it suffices by symmetry to treat the case @xmath370 .
now the formulas for @xmath995 and @xmath996 agree with those in
[ k2adeformation ] and in particular we see that the vector @xmath997 lies in @xmath175 and is primitive . ) now by the description of the versal deformation @xmath713 of @xmath998 over @xmath714 given in [ contraction ] and the identification of the fibers over @xmath999 ( in particular the description of the exceptional curve in proposition [ exceptional_locus_local_eq ] ) we find that the general fiber over the appropriate component of the boundary @xmath999 is isomorphic to the given contraction @xmath1000 .
indeed , it suffices to check that the strict transform of the exceptional curve for the deformation of @xmath414 intersects the same component of the exceptional locus of the minimal resolution of the @xmath425 singularity as in the original @xmath2 surface contraction @xmath1001 . by the description of the local equation of the exceptional curve in proposition [ exceptional_locus_local_eq ] this component corresponds to the ray @xmath1002 in the toric description of the minimal resolution of the singularity .
( to see this , note first that @xmath1003 by definition ( see proposition [ k1ainvariants ] ) , and that the number @xmath55 coincides with the invariant of the same name for the @xmath1 surface @xmath421 defined in 2 , by the computation of @xmath994 .
now observe that the local equation of @xmath251 is homogeneous with respect to the grading determined by the primitive generator of @xmath475 .
so the strict transform meets the interior @xmath1004 of the associated exceptional divisor . )
this agrees with the description of the original @xmath2 surface @xmath49 given in proposition [ k1ainvariants ] . by construction
( and the classification of @xmath1 extremal neighborhoods ) , the hyperplane section @xmath1005 of a @xmath1 extremal neighbourhood occurs as a fiber of a universal family @xmath942 over a torus fixed point of @xmath56 . see @xcite , remark 2.3 , p. 159 . by proposition [ k1adegeneratestok2a ] ,
the hyperplane section of a @xmath2 extremal neighborhood occurs as a fiber of a universal family over an interior point of a @xmath88-dimensional toric stratum of @xmath56 .
moreover , in each case , the deformation of the fiber given by the family @xmath942 is a versal @xmath144-gorenstein deformation .
this establishes the existence of @xmath66 . in the case of a flipping contraction
it is immediate that the pullback @xmath1006 of @xmath1007 is a proper birational morphism with exceptional locus @xmath251 such that @xmath1008 is @xmath144-cartier and @xmath1009 .
( note here that the proper birational morphism @xmath1007 has exceptional locus @xmath1010 , see proposition [ universalflip ] , and @xmath1011 . )
so it is the flip of @xmath5 . for the final statement ,
note first that @xmath1012 is surjective by kawamata
viehweg vanishing and cohomology and base change ( using @xmath1013 relatively ample ) , so the restriction of a general element @xmath330 of @xmath1014 is a general element of @xmath1015 .
we observe that every pair @xmath947 of a @xmath2 or @xmath1 surface @xmath49 together with a general element @xmath1016 occurs as a fiber of @xmath945 .
( see lemma [ restrict_to_h ] and the proof of proposition [ exceptional_locus_local_eq ] for the the description of the fibers of @xmath921 in the @xmath1 and @xmath2 cases respectively . )
the family of pairs @xmath1017 gives a versal @xmath144-gorenstein deformation of each fiber by lemma [ locallyversal ] .
so we may assume @xmath330 is the pullback of @xmath932 .
[ existenceofterminalantiflips ] let @xmath94 be a partial resolution of a cyclic quotient singularity such that @xmath95 has t - singularities and @xmath99 is relatively ample , and @xmath1018 a one parameter @xmath144-gorenstein smoothing of the germ @xmath111 .
then the morphism @xmath1019 extends to a proper birational morphism @xmath17 over @xmath954 .
assume that the general fiber @xmath1020 , @xmath1021 satisfies @xmath1022 .
then @xmath16 is irreducible and @xmath95 has only cyclic quotient singularities of type @xmath125 for some @xmath52 with @xmath53 .
the @xmath0-fold @xmath15 has @xmath144-factorial terminal singularities .
the morphism @xmath1019 is toric .
choose a toric structure and let @xmath1023 denote the toric boundary .
let @xmath1024 , @xmath97 be the torus fixed points , @xmath1025 the index of @xmath454 , and define @xmath1026 by @xmath1027 .
assume that the exceptional locus of @xmath1028 equals @xmath16 .
then there exists an antiflip @xmath8 of @xmath1028 such that @xmath4 has terminal singularities if and only if the following conditions are satisfied . 1 . for some choice of toric structure
, there is a divisor @xmath139 such that @xmath1029 , the toric boundary of @xmath95 .
if @xmath560 , define @xmath1030 . then the axial multiplicities @xmath141 of the singularities of @xmath15 satisfy @xmath1031 or @xmath1032 .
equivalently , we have @xmath1033 in the case that @xmath15 has fewer than two singularities , we define the axial multiplicities @xmath141 as follows .
let @xmath330 be a general element of @xmath1034 . by condition ( 1 ) the restriction
@xmath1029 is a choice of toric boundary for @xmath95 .
let @xmath1035 be the singular points of @xmath1036 .
if @xmath454 is a smooth point of @xmath15 , then the local deformation @xmath1037 of @xmath1038 is of the form @xmath1039 for some @xmath210 and convergent power series @xmath211 , @xmath212 .
equivalently , @xmath1040 is a du val singularity of type @xmath1041 .
we define @xmath1042 .
if @xmath454 is a singular point of @xmath15 then the definition of the axial multiplicity is the usual one .
the morphism @xmath1043 extends to a morphism @xmath1044 because @xmath1045 ( see @xcite , proposition 11.4 ) .
the morphism @xmath1043 is a @xmath1046-resolution of @xmath273 @xcite , definition 3.8 .
it is toric by @xcite , lemma 3.14 .
the singularities of @xmath95 are cyclic quotient singularities of type @xmath1047 for some @xmath1048 with @xmath1049 ( by the existence of the @xmath144-gorenstein smoothing , see @xcite , proposition 3.10 ) .
the milnor fiber @xmath56 of a @xmath144-gorenstein smoothing of such a singularity has milnor number @xmath246 .
thus , if @xmath1050 is the number of irreducible components of @xmath16 and the singularities @xmath1024 have invariants @xmath181 , a mayer vietoris argument gives @xmath1051 hence @xmath1022 iff @xmath16 is irreducible and @xmath1052 for each @xmath130 .
the @xmath0-fold @xmath15 has singularities of the form @xmath1053 the number @xmath1054 is the index of the singularity and the number @xmath210 is the _ axial multiplicity _ of the singularity @xcite , definition
corollary 1a.5 , p. 140 .
as already noted , these singularities are terminal , and analytically @xmath144-factorial by @xcite , 2.2.7 .
thus @xmath15 is terminal and @xmath144-factorial .
( note : in general , a one parameter @xmath144-gorenstein smoothing of a surface quotient singularity is terminal @xcite , corollary 3.6 . the modern proof of this fact uses inversion of adjunction " , see e.g. @xcite , theorem 5.50(1 ) , p. 174 . )
since @xmath1055 is a cyclic quotient singularity , it follows from the classification of non - semistable flips @xcite , appendix that if @xmath26 is a terminal antiflip of @xmath1028 then @xmath5 is semistable , that is , of type @xmath2 or @xmath1 .
moreover @xmath1056 is normal and @xmath247 by proposition [ b2=1=>normal ] because @xmath1057 for @xmath1021 .
let @xmath1058 be a general element .
then , by theorem [ universality ] , @xmath1059 is the pullback of a universal family @xmath1060 via a morphism @xmath62 , such that @xmath1011 .
let @xmath933 denote the strict transform of @xmath330 .
then @xmath1061 is the pull back of @xmath1062 via the morphism @xmath70 , and @xmath1063 is a general element such that @xmath1064 , so ( 1 ) holds . the morphism @xmath1065 is given by @xmath1066 for some @xmath1067 , @xmath1068 ( where @xmath1069 denotes higher order terms in @xmath1070 ) . by the description of the family @xmath754 , we see that @xmath141 are the axial multiplicities of the singularities of @xmath15 .
identify the fan @xmath76 of @xmath56 with its image in the cone @xmath1071 of @xmath730 under the map of fans corresponding to @xmath73 .
the point @xmath1072 lies in the toric boundary stratum of @xmath56 corresponding to the smallest cone @xmath1073 of the fan of @xmath56 containing the point @xmath1074 .
in particular @xmath1074 lies in the support of the fan of @xmath56 .
now the explicit description of the fan of @xmath56 above gives @xmath1075 .
conversely , suppose the conditions ( 1 ) and ( 2 ) are satisfied .
let @xmath1076 be general , so @xmath1029 is a choice of toric boundary for @xmath95 by ( 1 ) .
let @xmath1077 be the deformation of @xmath1078 constructed in [ contraction ] .
by lemma [ locallyversal ] , @xmath1079 is the pullback of @xmath1077 via a map @xmath1080 such that @xmath1081 .
the numerical data to build @xmath67 from @xmath729 is the following .
let @xmath1082 be the singularities of @xmath1083 such that @xmath1084 $ ] , @xmath1085 $ ] , and @xmath1086,\ ] ] where @xmath769 is the self - intersection of the proper transform of @xmath16 in the minimal resolution of @xmath95 .
if a singularity ( or both ) is ( are ) actually smooth , then we set @xmath1087 .
define @xmath1088 define @xmath1089 , @xmath1090 if @xmath1091 , or @xmath1092 else , and @xmath1093 , @xmath1094 .
one can check that @xmath1095 and gcd@xmath1096 .
if @xmath1097 $ ] , then one can check @xmath1098= \frac{\delta}{\omega}.\ ] ] one can verify that this gives an initial " @xmath1 for a mori sequence whose flipping " surface is @xmath95 ( see remark [ initialk2a ] ) . for the other initial @xmath1 ,
we define @xmath1099 , @xmath1100 if @xmath1101 , or @xmath1102 else , and @xmath1103 , @xmath1104 .
notice that these are the only two possibilities for initial @xmath1 corresponding to @xmath729 . by assumption
( 2 ) and the description of the toric morphism @xmath895 , the morphism @xmath1105 admits a lift @xmath62 such that @xmath1106 . indeed , as above , ( 2 ) is equivalent to requiring that the vector @xmath1074 given by the vanishing orders of the components of @xmath1105 lies in the image of the fan of @xmath56 in the cone @xmath167 corresponding to @xmath714 . this in turn is equivalent to the existence of the lift @xmath66 . now
the pull back of @xmath1107 by @xmath66 is the desired terminal antiflip @xmath1108 of @xmath1028 .
[ asflnsfgajf ] if we only impose condition ( 1 ) in corollary [ existenceofterminalantiflips ] then there is an antiflip @xmath26 such that @xmath4 has _
canonical _ singularities .
this is an application of @xcite , theorem 3.1 , p. 561 . indeed , by ( 1 ) , for a general divisor @xmath1076 the restriction @xmath140 is a choice @xmath1036 of toric boundary for @xmath1109 .
then , since @xmath140 is a nodal curve , @xmath1076 is a normal surface with at worst du val singularities of type @xmath28 ( at the nodes of @xmath1110 ) .
moreover the contraction @xmath1111 has at worst a du val singularity of type @xmath28 .
( more precisely , for @xmath330 general , we have du val singularities of type @xmath1112 at @xmath1040 for @xmath97 , where @xmath1025 is the index and @xmath1113 the axial multiplicity of @xmath1114 , and a du val singularity of type @xmath1115 at @xmath1116 . ) now by @xcite , theorem 3.1 , the @xmath1117-algebra @xmath1118 is finitely generated , and @xmath1119 has canonical singularities .
the morphism @xmath26 is the antiflip of @xmath1044 .
[ asdvasfgafg ] this is an explicit example of a @xmath129-gorenstein smoothing of a @xmath95 as in corollary [ existenceofterminalantiflips ] which has a nonterminal but canonical antiflip .
let @xmath49 be a surface with one singularity given by the @xmath1120-quotient of a simple elliptic singularity ( cf .
@xcite , theorem 9.6(3 ) ) .
the exceptional locus of its minimal resolution @xmath1121 is a union of @xmath1122 smooth rational curves @xmath113 , @xmath1123 , @xmath1124 , and @xmath1125 so that the @xmath180 are disjoint , each meets the curve @xmath1125 transversally at a single point , and @xmath1126 , @xmath1127 , @xmath1128 , and @xmath1129 .
assume @xmath49 has a smooth rational curve @xmath7 such that its proper transform in @xmath1121 is a @xmath112-curve intersecting only @xmath113 and transversally at one point .
notice that @xmath1130 and @xmath1131 . the surface @xmath49 has a @xmath129-gorenstein smoothing @xmath1132 ( cf .
the corresponding singularity in @xmath4 is canonical by @xcite , theorem 5.1 .
let @xmath1133 be the contraction of @xmath1134 ( the blowing down deformation ; cf . ) .
then @xmath371 is the cyclic quotient singularity @xmath1135 .
by @xcite , proposition 3.1(4 ) , the extremal nbhd @xmath1136 is @xmath129-factorial .
therefore , if it is of divisorial type , then @xmath14 would be @xmath129-gorenstein but @xmath1137 is not a t - singularity .
so , it is flipping , and the flip is given by the p - resolution @xmath1138 whose exceptional curve is a @xmath1139 passing through two singularities of type @xmath1140 .
[ canexample ]
fix integers @xmath1141 with gcd@xmath1142 .
let @xmath371 be the cyclic quotient singularity @xmath92 . an _
extremal p - resolution _ of @xmath371 is a partial resolution @xmath94 such that @xmath95 has only _ wahl singularities _ ( @xmath1143 cyclic of the type @xmath125 with gcd@xmath1144 ) , the exceptional curve @xmath98 , and @xmath99 is relatively ample .
they appear when we perform flips with the extremal neighborhoods of type @xmath2 or @xmath1 of the previous sections ( see corollary [ existenceofterminalantiflips ] ) .
the surface @xmath95 has at most two wahl singularities @xmath1145 ( @xmath97 ) such that @xmath1146 $ ] , @xmath1147 $ ] , and @xmath1086,\ ] ] where @xmath769 is the self - intersection of the proper transform of @xmath16 in the minimal resolution of @xmath95 .
if a singularity ( or both ) is ( are ) actually smooth , then we set @xmath1148 .
we define @xmath1149 then @xmath1150 , @xmath1151 by @xcite , theorem 3.9 , there is a natural bijection between p - resolutions of @xmath371 and irreducible components of the formal deformation space @xmath1152 . in @xcite , christophersen and stevens prove that p - resolutions of @xmath371 are in bijection with certain continued fractions representing zero . for extremal p - resolutions we have the following .
let @xmath1153 $ ] .
then the extremal p - resolutions of @xmath371 are in bijection with pairs @xmath1154 such that @xmath1155.\ ] ] we label these p - resolutions by @xmath1156 $ ] .
we now recall how to describe torically the p - resolution from the zero continued fraction .
we follow @xcite .
let us define in @xmath1157 the vectors @xmath1158 , @xmath1159 , and @xmath1160 for @xmath1161 .
if @xmath1162 , then @xmath1163 $ ] for @xmath1164 . note that @xmath1165
the fan @xmath76 of the toric surface @xmath95 , where the toric birational morphism @xmath94 is the extremal p - resolution , has two cones @xmath1166 and @xmath1167 where @xmath1168 , @xmath1169 , and @xmath1170 ; see @xcite pp.78 .
the @xmath1171 which correspond to @xmath1172 respectively in @xcite ( giving the roofs " of the cones ) are @xmath1173 one can verify that @xmath1174 .
[ contfracpres ] the extremal p - resolution @xmath1156 $ ] has @xmath1175 , \ \ \
\frac{m_1}{a_1}=[c_s,\ldots , c_{\beta+1}],\ ] ] ( if @xmath1176 and/or @xmath1177 , the corresponding points are smooth ) and either @xmath1178 $ ] , if @xmath1179 , for some @xmath1180 , or @xmath1181 . since @xmath1182
, we have @xmath1183 $ ] .
hence @xmath1184 $ ] . by symmetry
, we have @xmath1185 $ ] . we know that @xmath1186 .
then use lemma [ partfrac ] below with @xmath1187 $ ] and @xmath1163=:[e_1,\ldots , e_i]=\frac{p_{i+1}}{q_{i+1}}$ ] , setting @xmath1188 and @xmath1189 for the second equality in that lemma .
[ partfrac ] let @xmath1190 $ ] be a continued fraction with @xmath1191 and @xmath1192>0 $ ] for all @xmath1193 .
define @xmath1194 , @xmath1195 , and @xmath1196
. then @xmath1197\ ] ] for all @xmath1198 , @xmath1199 .
we have @xmath1200 and @xmath1201 for all @xmath1198 .
we now do induction on @xmath130 for the first equality , and induction on @xmath1202 for the second .
the following are all the extremal p - resolutions for cyclic quotient singularities @xmath92 with @xmath1203 , apart from the type @xmath1204 $ ] ( which corresponds to @xmath1205 ) .
the underlined notation represents another extremal p - resolution for the same singularity .
@xmath1206 \ \ \frac{11}{8}=[2,\overline 2,3 , \overline 2 ] \ \
\frac{13}{10}=[2,2,\overline 2,\overline 4 ] \ \
\frac{15}{11}=[\overline 2,\underline 2,3,\overline 2 , \underline 2]\ ] ] @xmath1207 \ \
\frac{19}{15}=[\overline 2,2,2,3,\overline 2,2 ] \
\ \frac{19}{12}=[2,\overline 3 , \overline 2 , 3]\ ] ] @xmath1208 \ \
\frac{22}{17}=[2,2,\overline 2,4,\overline 2 ] \ \
\frac{23}{19}=[\overline 2,2,2,2,3,\overline 2,2]\ ] ] @xmath1209 \ \
\frac{25}{16}=[\overline 2,3,2,\overline 2,3 ] \
\ \frac{27}{23}=[\overline 2,2,2,2,2,3 , \overline 2,2]\ ] ] @xmath1210 \ \
\frac{31}{27}=[\overline 2,2,2,2,2,2,3,\overline 2,2 ] \ \
\frac{31}{26}=[2,2,2,2,\overline 2,\overline 6]\ ] ] @xmath1211 \ \
\frac{34}{25}=[\overline 2,2,3,2,\overline 2,3 ] \ \
\frac{35}{31}=[\overline 2,2,2,2,2,2,2,3,\overline 2,2]\ ] ] @xmath1212 \ \
\frac{37}{30}=[2,2,2,\overline 2,5,\overline 2 ] \ \
\frac{37}{27}=[2,2,\overline 3,\overline 2,4]\ ] ] @xmath1213 \ \ \frac{39}{29}=[2,\overline 2,3,2,2,2,2,2,2,2,\overline 2]\ ] ] @xmath1214 \ \
\frac{40}{29}=[2,\overline 2,3,3,\overline 2,2 ] \ \
\frac{41}{18}=[3,2,\overline 2,3,\overline 3]\ ] ] @xmath1215 \ \
\frac{43}{37}=[2,2,2,2,2,\overline 2,\overline 7 ] \ \
\frac{43}{34}=[\overline 2,2,2,3,2,\overline 2,3]\ ] ] @xmath1216 \ \
\frac{45}{29}=[\overline 2,3,2,2,\overline 2,4 ] \ \
\frac{45}{19}=[3 , \overline 2,3,\overline 2,3]\ ] ] the purpose of the next two subsections is to prove [ atmost2 ] a cyclic quotient singularity @xmath92 can admit at most two distinct extremal p - resolutions .
[ samedelta ] if a cyclic quotient singularity @xmath92 admits two extremal p - resolutions , then the @xmath55 s are equal .
theorem [ atmost2 ] says that there are at most two universal antiflips over a given@xmath92 , and theorem [ samedelta ] says that the @xmath55 of a universal antiflip is an invariant of @xmath92 .
the proofs for both theorems are going to be completely combinatorial , using the continued fraction of an extremal p - resolution and the formula for @xmath55 in proposition [ contfracpres ] . according to @xcite ,
@xmath1217 $ ] is a zero - continued fraction if and only if there exists a triangulation of the @xmath1218-gon with vertices indexed by @xmath1219 such that the number of triangles meeting at the @xmath130-th vertex is equal to @xmath131 .
we define @xmath1220 as the number of triangles meeting at the @xmath89-th vertex .
we obviously have @xmath1221 we say that a sequence of integers @xmath1222 is _ of ww type _ if there exists @xmath1223 such that @xmath1224\ ] ] is a zero - continued fraction . to prove theorem [ atmost2 ] , we proceed by induction on @xmath1050 , and along the way we more or less classify sequences of @xmath1225 type .
a sequence @xmath1226 of @xmath1225 type gives rise to a triangulation that corresponds to .
we say that this triangulation is _ compatible _ with @xmath1226 .
the claim is that there are at most two triangulations compatible with @xmath1226 . for any triangulation , we call @xmath1227 the vertex sequence . by , all triangulations compatible with @xmath1226 have the same @xmath1220 .
we define @xmath1228 .
notice that @xmath1229 can be equal to one .
it is easy to see ( e.g. by induction ) that the vertex sequence always contains at least two @xmath88 s .
so we have the following possibilities : 1 .
@xmath1230 , all vertex sequences compatible with @xmath1226 have two @xmath88 s , @xmath1226 is obtained by substituting these @xmath88 s with @xmath71 s .
2 . @xmath1231
. we can handle case ( b ) by induction .
indeed , in this case all triangulations compatible with @xmath1226 contain a triangle @xmath1232 formed by vertices @xmath1050 , @xmath89 , and @xmath88 .
therefore they define triangulations of the @xmath1050-gon obtained by chopping off the triangle @xmath1232 .
these triangulations are compatible with the sequence @xmath1233 if @xmath1234 contains two @xmath88 s then @xmath1235 .
this implies that @xmath1236 and we are triangulating a square .
this case is obvious .
if @xmath1234 contains at most one @xmath88 then we are done by induction .
now we deal with case ( a ) .
first we inductively classify all triangulations with exactly two @xmath88 s in the vertex sequence .
if @xmath1237 then we can chop off a triangle with vertices @xmath1238 .
unless @xmath1236 , a new triangulation will have exactly two @xmath88 s , one of them at @xmath1239 or @xmath1240 .
we chop off the next vertex with @xmath88 , and proceed inductively . in other words , we can get any sequence @xmath1226 of this type by the following procedure
. start with a sequence @xmath1241 . then do a recursion analogous to enumeration of all @xmath1232-singularities : add a new @xmath71 on one end and increase another and by @xmath88 .
so , for example , we can do @xmath1242\to [ 3,\bar 2,2,2]\to [ 2,3,\bar 2,2,3]\to [ 2,2,3,\bar 2,2,4]\to \ldots\ ] ] after doing several steps , fold a sequence around a circle and add a new @xmath1243 between the ends . now choose @xmath1220 anywhere except at overlined @xmath71 s ( which should be decreased to ones in the vertex sequence ) .
we are going to ignore @xmath1220 from now on .
in particular , we will sometimes rotate indices if this will be convenient .
start with our sequence @xmath1244 of type ( a ) and let @xmath1245 be the number of @xmath130 s such that @xmath1246 .
now take any triangulation compatible with @xmath1226 .
let @xmath1247 be the indices such that @xmath1248 . by the discussion above , unless we are triangulating a square , there exist unique indices @xmath1249 such that the @xmath73-th vertex is adjacent to the vertex @xmath1250 , and similarly for @xmath1251 .
we rotate , and if necessary reflect the indices ( in the dihedral group ) so that , for the first triangulation , @xmath1252 and @xmath1253 . we create two new sequences : * @xmath1254 is a subsequence of @xmath1226 of terms greater than @xmath71 . *
@xmath1255 for @xmath1256 is the number of @xmath71 s in @xmath1226 between the terms that correspond to @xmath1113 and @xmath1257 increased by @xmath88 .
the indexing set here is @xmath1258 .
then the inductive procedure above allows us to compute @xmath1113 s in terms of @xmath1255 s .
there are two cases
. if @xmath1259 is even , we have @xmath1260 if @xmath1261 is odd , we have @xmath1262 now consider the second triangulation .
we change notation , and let @xmath1263 denote the data of the second triangulation .
the sequence @xmath1264 has @xmath1265 , @xmath1266 , and @xmath1267 .
the overlined and the underlined @xmath71 s should be decreased by @xmath88 to get the two vertex sequences .
the first triangulation has @xmath1252 , @xmath1268 ; the second has @xmath1269 , @xmath1270 . in this way , @xmath1271 , @xmath1272 , @xmath1273 , @xmath1274 , and @xmath1275 , and so @xmath1276 , @xmath1277 , @xmath1278 , @xmath1279 , and @xmath1280 .
the sequence @xmath1281 has @xmath1282 , @xmath1283 , and @xmath1284 .
we have @xmath1285 , @xmath1272 , @xmath1286 , @xmath1287 , @xmath1288 , @xmath1289 , @xmath1290 , and @xmath1291 , and so @xmath1292 , @xmath1277 , @xmath446 , @xmath1293 , @xmath1294 , @xmath1295 , @xmath1296 , and @xmath1297 .
_ we first consider the case of odd @xmath1261 .
_ after interchanging @xmath73 and @xmath1298 if necessary , we can assume that the vertex @xmath73 is located clock - wise from the vertex @xmath1299 .
we define @xmath1300 through @xmath1301 .
notice that @xmath1302 .
then the sequence @xmath1303 satisfies another system of equations : @xmath1304 ( as a reminder , the indexing set is @xmath1258 ) .
we claim that @xmath1305 is uniquely determined by the sequence @xmath1306 .
this will show that the second triangulation compatible with @xmath1226 is unique ( if exists ) .
we have the following system of equations when @xmath1307 or @xmath1308 @xmath1309 let @xmath1310 then the system of equations above shows that @xmath1255 is constant on @xmath432-cosets in @xmath1258 unless a coset contains @xmath89 , @xmath1311 , @xmath1202 , or @xmath1312 .
next we compute a gauss sum @xmath1313 where @xmath283 is a primitive @xmath1245-th root of unity .
clearly only one or two @xmath432 cosets contribute non - trivially to the gauss sum .
namely , if @xmath1314 then there will be two non - trivial cosets , and we have @xmath1315 where @xmath1316 is the minimal positive integer such that @xmath1317 . geometric series and some manipulations give @xmath1318 we can clearly recover @xmath1319 , and hence @xmath1305 , by applying a fractional - linear transformation . if @xmath1320 then in fact @xmath1321 , so there is only one coset .
we have @xmath1322 , and we choose a minimal @xmath1070 such that @xmath1323 .
there are two further subcases , either @xmath1324 and we have @xmath1325 and so @xmath1326 , or @xmath1327 , in which case a similar calculation gives @xmath1328 . for the cases
@xmath1329 and @xmath1330 we obtain same @xmath1331 as well .
therefore @xmath1305 is uniquely determined by the @xmath1306 .
_ now let s consider the case of even @xmath1259 . _
after interchanging @xmath73 and @xmath1298 if necessary , we can assume that @xmath1299 satisfies @xmath1332 with @xmath1333 .
there are two further subcases : in the _ clockwise _ ( resp .
_ counter - clockwise _ ) case the vertex @xmath73 is located clockwise ( resp .
_ counter - clockwise _ ) from the vertex @xmath1299 .
notice that in the clockwise case @xmath1302 .
the sequence @xmath1303 satisfies a new system of equations : @xmath1334 in the clockwise case and @xmath1335 in the counter - clockwise case .
this gives a system of equations on @xmath1255 : @xmath1336 in the clockwise case and @xmath1337 in the counter - clockwise case .
we have to recover @xmath1305 ( and whether we have a clockwise or a counter - clockwise case from the sequence @xmath1255 ) . to simplify this problem , consider a new sequence @xmath1338 defined as @xmath1339 we think about @xmath1338 as indexed by the cyclic group @xmath1340 .
it satisfies the following system of equations : @xmath1341 in the clockwise case and @xmath1342 in the counter - clockwise case .
we define a subgroup @xmath1343 in the clockwise ( resp .
counter - clockwise ) case .
we also define a gauss sum @xmath1344 where @xmath283 is a primitive @xmath1202-th root of unity .
the system of equations above shows that @xmath1338 is constant on @xmath432-cosets in @xmath1340 unless a coset contains @xmath89 or @xmath1311 .
moreover , in the latter case the coset must contain both @xmath89 and @xmath1311 , as otherwise we will have a non - constant periodic monotonous function .
only this coset will contribute non - trivially to @xmath1331 . in the clockwise case we have @xmath1345 @xmath1346 , and
@xmath1347 therefore @xmath1348 obviously @xmath1349 and @xmath1331 determines @xmath1305 . in the counter - clockwise case
we have @xmath1350 hence @xmath1351 where @xmath1352 .
so we obtain @xmath1353 it follows that @xmath1354 and so @xmath1355 indeed , as we connect a fixed vertex @xmath88 of an @xmath1202-gon to other vertices by diagonals , two consecutive diagonals never have equal length , unless @xmath1356 , but this would contradict .
in particular , we can not have both a clockwise ( in which case @xmath1357 ) and a counter - clockwise case .
it follows from that we can find @xmath1305 by solving the quadratic equation @xmath1358 by , this equation can not have two solutions .
we now prove theorem [ samedelta ] . as for theorem [ atmost2 ]
, we only need to consider case ( a ) .
let @xmath1244 be a sequence of type ( a ) .
we adopt the same conventions as before , and define @xmath1359 , and the @xmath1360 , where @xmath1361 and @xmath1362 are the indices defining the two zero continued fractions corresponding to @xmath1226 .
we define for @xmath773 @xmath1363,\ ] ] for some @xmath1364 ( see proposition [ contfracpres ] ) .
one can verify that @xmath1365\ ] ] for some @xmath1366 with gcd@xmath1367 ; this is a property of the dual of continued fractions from wahl singularities ( see @xcite , 3.5 ) .
the following is @xcite , lemma 3.4 . given @xmath1368=\frac{m^2}{ma-1}$ ] with @xmath1369 and gcd@xmath1144 , we define as in [ toric ] the sequences @xmath1370 $ ] for @xmath1371 , and @xmath1372 $ ] for @xmath1373 .
define @xmath1374 and @xmath1375 .
let @xmath1376 for @xmath1377 .
let @xmath1378 , @xmath1379 , @xmath1380 be the numbers for @xmath1381 $ ] .
then @xmath1382 for @xmath1383 , and @xmath1384 .
this lemma gives a way to compute @xmath260 for continued fractions of wahl singularities from the inductive procedure ( cf .
@xcite , proposition 3.11 ) @xmath1190 \to [ e_1 + 1,\ldots , e_r,2]$ ] or @xmath1385 $ ] , starting with @xmath1386 $ ] . for @xmath1386
$ ] one has @xmath1387 ; here @xmath1388 , @xmath1389 .
thus for @xmath1390 $ ] we have @xmath1391 , and so on .
therefore , the @xmath1392 s can be used to compute @xmath1393 from the continued fraction of @xmath1394 .
start with @xmath1395 $ ] and assign @xmath88 for the first @xmath71 and @xmath88 for the last @xmath71 .
for the new continued fraction we assign to the new @xmath71 the addition of the extremal previous assignations . to obtain @xmath205 from @xmath1190 $ ] , we just add the numbers assigned to @xmath1396 and to @xmath1397 . for @xmath1395
$ ] we have @xmath1398 . for @xmath1399 $ ]
we have @xmath1400 , for @xmath1401 $ ] @xmath1402 , for @xmath1403 $ ] @xmath1404 , etc .
the continued fraction @xmath1405 where @xmath1406 for all @xmath130 , will be denoted by @xmath1407 .
we have @xmath1408 consider a sequence @xmath1409 of type ( a ) as above .
we recall that for @xmath1410 we have indices @xmath1411 and an integer @xmath1412 .
then @xmath1413 @xmath1414 [ deltalemma ] the proof uses the inductive way to compute @xmath1415 for the continued fraction of @xmath1416 , and the numbers @xmath1417 defined before .
the role of @xmath1418 is taken by @xmath1419 ( if @xmath1420 ) or @xmath1421 ( if @xmath1422 ) to compute @xmath1423 .
we start with @xmath1424 $ ] and use @xmath1418 to obtain @xmath1425 $ ] .
then use @xmath1426 to obtain @xmath1427 $ ] , and continue . during this process
we keep track of the assigned multiplicities explained above .
this defines the recursion @xmath1428 , @xmath1429 , @xmath1430 for @xmath1431 , and @xmath1432 . in this way , @xmath1433 we are going to compare the @xmath1392 s , but before another lemma . [ magic ]
let @xmath1434 be a matrix such that @xmath1435 .
define matrices @xmath1436 consider an arbitrary sequence @xmath1437 , where @xmath1438 , and the matrix @xmath1439 ( which ends with @xmath49 or @xmath1440 depending on parity of @xmath1298 )
. then 1 .
@xmath1441 , where @xmath1442 is a polynomial which depends only on @xmath324 .
2 . if @xmath1298 is even , the left - upper corner of @xmath1443 is equal to the left - upper corner of @xmath1444 , where @xmath1445 is a sequence given by @xmath1446 for any @xmath1316 .
the first statement follows by induction on @xmath1298 , where we can show , more precisely , that @xmath1447 if @xmath1298 is odd and @xmath1448 if @xmath1298 is even , where @xmath1449 and @xmath1450 are polynomials in @xmath324 only . the second statement of the lemma follows from the first .
indeed , in order to compute @xmath1442 we can assume that @xmath1451 is a symmetric matrix of determinant @xmath88 .
then @xmath1452 ( and in particular their upper - left entries are the same ) , where @xmath1453 is a sequence obtained from @xmath1454 by reversing the order . on the other hand , we can also assume that @xmath1455 .
then @xmath1442 is equal ( up to sign ) to the numerator of the hirzebruch
jung continued fraction with entries @xmath1456 . this numerator does not change if we reverse the order .
we will show @xmath1457 case by case .
some notation : a sequence @xmath1458 is a sequence of @xmath1459 , @xmath1460 is a sequence of integers , and @xmath1461 is the same sequence of integers but in reversed order . *
case i. * assume @xmath1259 and @xmath1422 , i.e. the counter - clockwise case .
here we have @xmath1462 .
one can check that for all @xmath1463 , @xmath1464 . by the formulas of lemma [ deltalemma ] , we have @xmath1457
. * case ii . *
assume @xmath1259 and @xmath1420 , i.e. the clockwise case .
we know that either @xmath1311 or @xmath1465 belongs to @xmath432 .
assume that @xmath1466 .
we also redefine @xmath1467 and @xmath1468 .
then , the transposed formulas of lemma [ deltalemma ] are @xmath1469 and @xmath1470 for some @xmath1471 and @xmath1472 .
but we know that @xmath1392 s are computed via the upper - left corner of the multiplication of the corresponding @xmath1473 matrices , and if @xmath1474 we can use lemma [ magic ] to prove @xmath1457 .
we notice that @xmath1475 * case iii . * assume @xmath1259 , @xmath1420 , and that @xmath1476 .
we also redefine @xmath1467 and @xmath1468 .
then , the transposed formulas of lemma [ deltalemma ] are @xmath1477 and @xmath1478 for some @xmath1471 and @xmath1472 .
so we do the same as in case ii , noticing that by lemma [ magic ] the use of @xmath49 or @xmath1440 does not change the @xmath1392 .
* case iv . *
assume @xmath1261 , @xmath1479 or @xmath1480 , and that @xmath1481 .
then @xmath1482 . in this case @xmath1420 .
one can verify that @xmath1483 for all @xmath1463 , and so by lemma [ deltalemma ] if @xmath1484 then @xmath1485 .
this gives @xmath1457 . * case v. * assume @xmath1261 , @xmath1479 or @xmath1480 , and that @xmath1202 is not in @xmath432
. here @xmath1420 .
this case can be treated as case ii .
* case vi . *
assume @xmath1261 , and that @xmath1329 .
then via lemma [ deltalemma ] we have @xmath1486 and @xmath1487 . in the case @xmath1330
, we would have @xmath1488 and @xmath1489 .
therefore , in any case , @xmath1457 .
let @xmath41 be the gieseker s moduli space of canonical surfaces of general type with given numerical invariants @xmath42 and @xmath43 .
let @xmath44 be its compactification , the moduli space of stable surfaces of kollr , shepherd - barron , and alexeev . given a map @xmath45 from a punctured smooth curve germ , i.e. a family @xmath46 of canonical surfaces , one is interested in computing its `` stable limit '' @xmath47 , perhaps after a finite base change .
we would like to make more explicit a well - known algorithm for computing a stable limit under an extra assumption that the family @xmath46 can be compactified by a flat family @xmath48 , where the special fiber @xmath49 is irreducible , normal , and has quotient ( = log terminal ) singularities .
we will show that the stable limit can be found using only flips and divisorial contractions from the extremal neighborhoods of type @xmath2 and @xmath1 of the previous sections .
we are going to perform various operations on families ( e.g. simultaneous partial resolutions ) which make it necessary to assume that the total space of the family is an algebraic space ( or an analytic space ) rather than a scheme , even if the total space of the original family is a scheme .
however , our main applications are to moduli of surfaces of geometric genus @xmath89 , in which case any flat family is projective , and therefore its total space is a scheme . so the reader can ignore this technicality .
the next lemma shows that we can reduce our analysis to the case when @xmath48 is a @xmath129-gorenstein smoothing and the special fiber has wahl singularities only .
this is a global version of the m - resolution @xcite of behnke and christophersen .
let @xmath1490 be a flat family of irreducible reduced projective surfaces such that the special fiber has quotient ( = log - terminal ) singularities .
there exists a finite morphism @xmath1491 , an analytic space @xmath1492 , and a proper birational map @xmath1493 such that 1 . the special fiber @xmath701 is irreducible and has wahl singularities .
@xmath1492 has terminal singularities .
3 . @xmath1125 induces a minimal resolution of general fibers @xmath1494 .
@xmath1495 is @xmath1125-nef . by @xcite ,
theorem 3.5 , we can choose @xmath1496 such that all properties above are satisfied except that @xmath701 is allowed to have arbitrary t - singularities ( i.e. @xmath129-gorenstein smoothable quotient singularities ) . notice that locally on @xmath1497 this is nothing but a p - resolution .
by @xcite , any singular point of @xmath1496 has an analytic neighborhood @xmath807 and a proper crepant map @xmath1498 ( a crepant m - resolution ) which has the properties required in the lemma , in particular the special fiber has wahl singularities only .
gluing these resolutions gives the required analytic space .
[ t3 ] assume we have a @xmath129-gorenstein smoothing @xmath1499 of a projective surface @xmath49 with only wahl singularities .
suppose some positive multiple of @xmath1500 is effective .
then , after a finite sequence of flips and divisorial contractions from extremal neighborhoods of type @xmath2 and @xmath1 ( with @xmath50 , as in proposition [ b2=1=>normal ] ) , we obtain a @xmath129-gorenstein smoothing @xmath1501 with @xmath1008 @xmath1125-nef .
notice that @xmath1502 is nef if and only if @xmath1503 is @xmath1125-nef . indeed , suppose @xmath1504 for some curve in a general fiber . without loss of generality
we can assume that @xmath1505 is a @xmath112-curve , and therefore deforms over @xmath1506 .
since @xmath4 is @xmath129-gorenstein , @xmath1500 will intersect negatively one of the irreducible components of the central fiber of the closure of this family of @xmath112-curves . to avoid issues with the fact that @xmath4 in general is not a variety
, we will run a mmp on @xmath4 guided by a mmp on the special fiber .
suppose @xmath1500 is not nef .
then we have an extremal contraction @xmath1507 of log terminal surfaces , where the exceptional locus @xmath7 is an irreducible rational curve .
singularities of @xmath273 are log terminal , and therefore rational . by blowing down deformations @xcite ,
we get an extremal neighborhood @xmath64 of either divisorial or flipping type , and in the latter case the flip exists in the category of analytic spaces by @xcite .
it remains to show that we only encounter flips ( or divisorial contractions ) of types @xmath2 and @xmath1 ( with @xmath50 ) .
since @xmath273 has quotient singularities , we just have to rule out the possibility of having non cyclic quotient singularities .
however , this is known by @xcite , pp .
157159 , the proof that subcase ( 3c ) of page 154 does not occur .
below we construct a stable surface @xmath49 which belongs to @xmath44 with @xmath145 and @xmath146 ( so @xmath1508 and @xmath1509 ) .
the surface @xmath49 has two wahl singularities @xmath147 and @xmath148 , and no local - to - global obstructions to deform ( i.e. deformations of its singularities globalize to deformations of @xmath49 ) .
we will show that a @xmath129-gorenstein smoothing of @xmath49 is a simply connected smooth projective surface of general type with the above invariants .
this is the first example in the literature with those invariants and two wahl singularities , which is the maximum possible under no local - to - global obstructions .
examples with one wahl singularity appeared in @xcite .
we consider a @xmath129-gorenstein deformation of @xmath49 which preserves the singularity @xmath454 and smooths @xmath1510 . since this deformation is trivial around @xmath454 ,
we resolve simultaneously @xmath454 to obtain a @xmath129-gorenstein smoothing of a surface with a single wahl singularity at @xmath1510 . on this deformation
we explicitly run the mmp of theorem [ t3 ] .
we do it for each @xmath97 .
let us consider the cubic pencil @xmath1511 in @xmath1512 with @xmath1513 \in \p_{t_1,t_2}^1 $ ] .
let @xmath1514 , @xmath1515 , @xmath1516 , @xmath1517 , and @xmath1518 . the curves @xmath198 and @xmath28 are tangent at @xmath1519 $ ] .
this pencil defines an elliptic fibration @xmath1520 with singular fibers @xmath1521 ( kodaira s notation ; cf .
@xcite , p. 201 ) , after we blow - up the nine base points .
choose a nodal @xmath1522 fiber and denote it by @xmath1125 , and consider the line @xmath56 passing through the point @xmath1523 and the node of @xmath1125 .
the line @xmath56 becomes a double section of @xmath66 .
we blow up @xmath1524 times @xmath962 to obtain @xmath1525 .
write @xmath1526 for the composition of these blow - ups .
the curves @xmath1527 in figure [ f3 ] are the exceptional curves of @xmath1073 .
the order of the blow - ups is indicated by the subindices of the @xmath1527 . in figure [ f3 ]
we label all relevant curves for the construction .
the self - intersections of all of them are in figure [ f4 ] .
let @xmath1528 be the contraction of the configurations @xmath1529 and @xmath1530 .
let @xmath454 be the image of @xmath1531 , so @xmath1532 and @xmath1533 .
the surface @xmath49 has no local - to - global obstructions to deform .
a @xmath129-gorenstein smoothing of @xmath49 is a simply connected projective surface of general type with @xmath145 and @xmath146 .
the surface @xmath49 is a stable surface .
[ p4 ] according to @xcite , section 2 , if @xmath1534 , then @xmath49 has no local - to - global obstructions to deform .
let @xmath1535 be the composition of blow - ups corresponding to @xmath1536 , @xmath1537 , and @xmath1538 .
let @xmath7 be the general fiber of @xmath1539 , and let @xmath1540 .
then @xmath1541 and so @xmath1542 . we can follow the strategy in @xcite to show @xmath1543 .
relevant intersections are @xmath1549 , @xmath1550 , @xmath1551 , @xmath1552 , @xmath1553 , @xmath1554 , and @xmath1555 .
this says that @xmath1500 is nef .
in addition , note that the support of @xmath1544 above contains @xmath1125 , @xmath1556 , @xmath1536 , and @xmath1557 .
this is the support of a fiber , and so @xmath1544 zero curves should be contained in fibers .
there are no such curves .
then by nakai - moishezon criterion we have @xmath1502 is ample , hence @xmath49 is stable .
we clearly have @xmath1558 and @xmath1559 , and so for the general fiber of the smoothing . to compute the fundamental group of the general fiber of a @xmath129-gorenstein smoothing of @xmath49
, we follow the same strategy as in @xcite , theorem 3 .
let us write @xmath1560 for two equivalent loops @xmath1561 in @xmath1562 .
we refer to @xcite ( see pages 12 and 20 ) for the relations we will use among loops . in particular notice that for any loop @xmath1202 around @xmath1563 we have @xmath1564 , and for any loop @xmath1565 around @xmath1566 we have @xmath1567 .
the relevant loops for us are : @xmath202 around @xmath1537 , @xmath203 around @xmath1538 , @xmath1568 around @xmath1569 , @xmath1570 around @xmath198 , @xmath1571 around @xmath56 , and @xmath1572 around @xmath1573 . because of @xmath1574 , we have @xmath1575 . in this way ,
@xmath1576 in @xmath1562 . by @xcite
, we have @xmath1577 , and so @xmath1578 . by @xcite
, we also have @xmath1579 , and so @xmath1580 .
because of @xmath1581 , we have the relation @xmath1582 , and so @xmath1583 .
but , again by @xcite , we have @xmath1584 .
hence , since @xmath1585 , we obtain @xmath1586 . therefore , the relation @xmath1579 implies @xmath1587 .
but now we use @xmath1588 to have @xmath1589 , and so @xmath1590 .
this together with @xmath1591 gives @xmath1592 .
then @xmath1593 and @xmath1594 , and so @xmath1595 .
the fact that @xmath1596 in @xmath1562 is enough to use directly the method in @xcite , to conclude that a @xmath129-gorenstein smoothing of @xmath49 is simply connected . consider a @xmath129-gorenstein deformation of @xmath49 which preserves the singularity @xmath151 and smooths @xmath152 .
let @xmath153 be the minimal resolution of @xmath239 .
since this deformation is trivial around @xmath151 , we resolve minimally simultaneously @xmath151 to obtain a @xmath129-gorenstein smoothing @xmath1597 of @xmath153 .
we are going to run theorem [ t3 ] on @xmath1597 . in general , we use the following notation to save writing when running the mmp of theorem [ t3 ] . [ dualgraphs ] locally , let @xmath1598 be an extremal neighborhood of type @xmath2 or @xmath1 ( with @xmath50 ) .
let @xmath1121 be the minimal resolution of @xmath49 .
we remark that the proper transform @xmath1599 of @xmath7 is a @xmath112-curve .
we have the following dual graph for the exceptional divisors and @xmath1600 .
we draw @xmath1601 for the exceptional curves and a @xmath1602 for @xmath1600 , edges are the transversal intersections between curves . if @xmath1125 is divisorial , then we draw @xmath1601 for exceptional curves in the dual graph of the minimal resolution of @xmath273 .
if @xmath1125 is flipping , we draw the dual graph of the exceptional curves in the minimal resolution of @xmath95 ( the surface corresponding to the extremal p - resolution ) and the proper transform @xmath1603 of @xmath16 , @xmath1601 for the exceptional and a @xmath1604 for @xmath1603 . numbers are minus the self - intersections of the curves , no number means @xmath112-curve .
globally , let @xmath49 be a projective surface with wahl singularities , which has a @xmath129-gorenstein smoothing @xmath1605 .
let @xmath1606 be its minimal resolution .
we consider the dual graph of all exceptional divisors and some curves which may be used when we perform flips and divisorial contractions of @xmath4 . at the beginning
we draw @xmath1601 for the exceptional curves and @xmath1607 for the rest .
one of them is a @xmath1602 which defines the first birational operation ( locally as above ) .
the operation is indicated with an arrow .
the resulting dual graph is the minimal resolution of the new central fiber ( either @xmath273 or @xmath95 ) . when flipping a @xmath1604
will appear , indicating the proper transform of the flipping curve .
also , a new @xmath1602 will indicate the next birational operation .
of course the new dotted graph depends on the computations of the previous sections , and they will be omitted .
we start with the curve @xmath7 as in figure [ f5 ] .
this curve defines a flipping extremal neighborhood of type @xmath2 @xmath1608 .
( we use same letter for a curve and its proper transform . )
we perform the flip and obtain @xmath1609 .
this is our new central fiber . in figure [ f5 ]
we show the action of the flip on the minimal resolutions of @xmath153 and @xmath95 .
figure [ f6 ] shows the dual graphs ( as explained above ) for a sequence of flips and divisorial contractions starting with @xmath153 .
the first arrow represents the situation in figure [ f5 ] . in order
we have : two flips , two divisorial contractions and two flips .
the three concurrent edges of the last graph represent a ( simple ) triple point .
the general fiber of @xmath1610 is a rational surface . in this way
the singularity @xmath239 labels a boundary curve in @xmath44 which generically consists of rational stable surfaces with only @xmath1611 as singularity .
[ p5 ] we now do it for @xmath241 .
as before , consider a @xmath129-gorenstein deformation of @xmath49 which preserves the singularity @xmath152 and smooths @xmath151 .
let @xmath154 be the minimal resolution of @xmath241 .
since this deformation is trivial around @xmath152 , we resolve minimally simultaneously @xmath152 to obtain a @xmath129-gorenstein smoothing @xmath1612 of @xmath154 .
we now perform @xmath1613 flips , the first @xmath1614 are of type @xmath2 and the last of type @xmath1 .
the corresponding sequence of dual graphs is shown in figures [ f7 ] and [ f8 ] .
( we clarify that in figures [ f7 ] and [ f8 ] there is one darker edge connecting a @xmath1615-curve and a @xmath1616-curve which are tangent .
this is after the 9th flip . )
we will prove that the canonical class of the last special fiber ( see lower right corner of figure [ f8 ] ) is nef . to save notation , we first contract the configuration @xmath1617 into @xmath1618 .
let @xmath1619 be the resulting surface ( last special fiber ) , and let @xmath1620 be its minimal resolution . in figure [ f9 ] we show relevant curves in the blow - up of @xmath1620 at one point ( the corresponding @xmath112-curve is @xmath1621 ) .
notice that the singular @xmath1619 has two t - singularities given by the configurations @xmath1622 and @xmath198 .
the @xmath581 s in figure [ f9 ] are the exceptional curves to obtain the minimal initial elliptic fibration ( with singular fibers @xmath1521 ) from @xmath1623 .
the @xmath180 s are the exceptional curves for the four further blow - ups needed to get @xmath1620 .
notice we do not have the @xmath1624th blow - up associated to @xmath1621 .
let @xmath1625 be the minimal resolution of both t - singularities , and let @xmath1626 be the composition of the @xmath1627 blow - ups . in that way , if @xmath1628 is the class of a line in @xmath1623 , then the canonical class of @xmath1620 is @xmath1629 we know that the @xmath129-gorenstein smoothing of @xmath1619 is simply connected with @xmath146 , and it has a @xmath1637 and a @xmath1616 smooth rational curves @xmath1638 and @xmath1639 respectively , coming from the singularity at @xmath152 . if @xmath38 is the general fiber of the @xmath129-gorenstein smoothing of @xmath1619 , then @xmath1640 , where @xmath1331 is the general fiber of the elliptic fibration of @xmath38 , and @xmath1641 are the multiplicities of the two multiple fibers . in this way , we find @xmath1642 and @xmath1643 using the equations @xmath1644 and @xmath1645 . therefore @xmath38 is a dolgachev surface of type @xmath1646 @xcite , p.383 .
the general fiber of @xmath1612 is a dolgachev surface of type @xmath1646 which has a chain of four @xmath1139 s with consecutive self - intersections @xmath1637 , @xmath1616 , @xmath1615 , @xmath1615 . in this way
the singularity @xmath241 labels a boundary curve in @xmath44 which generically parametrizes dolgachev surfaces of type @xmath1646 containing the exceptional divisor of @xmath1647 .
[ p5 ] w. p. barth , k. hulek , c. a. m. peters and a. van de ven , _ compact complex surfaces _ , ergebnisse der mathematik und ihrer grenzgebiete .
3 . folge . , second edition , vol .
4 , springer - verlag , berlin , 2004 .
j. a. christophersen , _ on the components and discriminant of the versal base space of cyclic quotient singularities _ , singularity theory and its applications , part i(coventry , 1988/1989 ) , lecture notes in math .
1462 , springer , berlin ( 1991 ) , 8192 .
j. stevens , _ on the versal deformation of cyclic quotient singularities _ , singularity theory and its applications , part i(coventry , 1988/1989 ) , lecture notes in math .
1462 , springer , berlin ( 1991 ) , 302319 . | we study semistable extremal @xmath0-fold neighborhoods , which are fundamental building blocks of birational geometry , following earlier work of mori , kollr , and prokhorov .
we classify possible flips and extend mori s algorithm for computing flips of extremal neighborhoods of type @xmath1 to more general @xmath2 neighborhoods .
the novelty of our approach is to show that @xmath2 belong to the same deformation family as @xmath1 , in fact we explicitly construct the universal family of extremal neighborhoods .
this construction follows very closely mori s division algorithm , which can be interpreted as a sequence of mutations in the cluster algebra .
we identify , in the versal deformation space of a cyclic quotient singularity , the locus of deformations such that the total space admits a ( terminal ) antiflip .
we show that these deformations come from at most two irreducible components of the versal deformation space . as an application
, we give an algorithm for computing stable one - parameter degenerations of smooth projective surfaces ( under some conditions ) and describe several components of the kollr shepherd - barron alexeev boundary of the moduli space of smooth canonically polarized surfaces of geometric genus zero . |
in the chromium spinels @xmath2cr@xmath3 ( @xmath2 = zn , cd , hg ; @xmath4 = o , s , se ) competing antiferromagnetic ( afm ) and ferromagnetic ( fm ) interactions @xcite establish a fascinating phase diagram with complex ground states .
@xcite in these compounds the cr@xmath5 ions exhibit a half - filled @xmath6 shell constituting a spin - only state with @xmath7 and vanishing spin - orbit coupling . the oxides , with small lattice constants ,
are dominated by direct afm cr - cr exchange . as the chromium ions form a corner - sharing tetrahedral network , which is a prototypical example of a pyrochlore lattice , these compounds are geometrically frustrated @xcite and undergo antiferromagnetic spin order at temperatures @xmath8 , where @xmath9 is the curie - weiss temperature .
the magnetic transitions are accompanied by structural distortions which have been explained in terms of a spin - driven jahn - teller ( jt ) effect .
@xcite in most of these afm spinels , the phonon eigenmodes reveal significant splittings below the afm ordering temperature .
@xcite the phonon splittings are driven by magnetic exchange interactions via strong spin - phonon coupling , despite the fact that cr@xmath5 in an octahedral crystalline electric field reveals a spherical charge distribution with a @xmath10-value close to @xmath11 , synonymous with the absence of conventional spin - orbit coupling . at large lattice constants , in some of the sulfides and selenides ,
@xcite 90@xmath12 cr-@xmath4-cr exchange becomes dominant , leading to fm semiconductors with relatively high magnetic transition temperatures .
recently relaxor ferroelectricity has been reported for cdcr@xmath0s@xmath1 ( ref . ) and hgcr@xmath0s@xmath1 .
@xcite this seems to be a further experimental evidence concerning the long - standing question of local polar distortions in spinel compounds .
@xcite however , the origin of these polar distortions is unclear @xcite and in ab initio phonon calculations @xcite no indications for a structural instability due to soft phonon modes have been detected .
recently , evidence accumulates that the large magneto - capacitive ( mc ) effects in the chromium spinels strongly depend on thermal treatment and/or doping . in pure ceramics , for example ,
mc effects are absent or strongly reduced , but they are recovered in indium doped samples . on the contrary ,
annealing in vacuum eliminates mc effects in cl - doped single crystals , but enables them in cl - free single crystals .
@xcite the observation that marginal doping with charge carriers induces dipolar relaxation phenomena reminds of similar effects in transition - metal monoxides doped on a subpercent level with li ( ref . ) or on metal deficient semiconducting nio .
@xcite specifically for mno it has been suggested that li doping leads to mn@xmath13 ions with a strongly jahn - teller active @xmath14 electron configuration , inducing polaron - like bound charge carriers . in a similar way ,
cdcr@xmath0s@xmath1 with sulphur deficiency or doped with cl@xmath15 will exhibit cr@xmath16 with a jt active @xmath14 configuration and possible dipolar relaxation of bound polarons .
additional experiments will be necessary to unravel if the observed relaxor phenomena can be explained by doping only or by intrinsic lattice instabilities driven by off - centering of the cr ions as has been proposed by recent raman scattering experiments .
@xcite it is known since long time that in cdcr@xmath0s@xmath1 strong spin - phonon coupling is active , @xcite leading to an unusual temperature dependence of eigenfrequencies and dampings of the infrared ( ir ) active phonon modes . in addition , recent raman scattering experiments provide experimental evidence for polar distortions well above the magnetic ordering temperature .
@xcite in this work we communicate the results of a detailed ir study .
we report on the temperature dependence of eigenfrequencies , dampings and effective plasma frequencies of all ir active phonons in hgcd@xmath0s@xmath1 and cdcr@xmath0s@xmath1 .
our results show , that while there are indeed no indications of strong phonon softening , we observe strong effects in the temperature dependence of the plasma frequencies of specific modes , indicative for significant changes of the character of bonds and that the effective charges undergo large variations as a function of temperature .
in addition it seems worthwhile to search for phonon splitting in afm hgcr@xmath0s@xmath1 , as has been observed in the related cr spinels zncr@xmath0s@xmath1 ( ref . ) and zncr@xmath0se@xmath1 ( ref . ) below the antiferromagnetic phase - transition temperature .
hgcr@xmath0s@xmath1 is close to fm order and moderate magnetic fields well below 1 t induce a polarized ferromagnetic state which is shifted by almost 60 k in external fields of 5 t. @xcite it thus seems particularly important to study the spin - phonon coupling also as function of an external magnetic field .
cdcr@xmath0s@xmath1 and hgcr@xmath0s@xmath1 crystallize in the normal cubic spinel structure ( @xmath17 ) with lattice constants @xmath18 nm and @xmath19 nm and sulphur fractional coordinates @xmath20 and @xmath21 for the cadmium and mercury compound , respectively .
@xcite both compounds are dominated by strong fm exchange with curie - weiss temperatures of the order of 150 k. but while cdcr@xmath0s@xmath1 reveals fm order at @xmath22 k , @xcite hgcr@xmath0s@xmath1 exhibits spiral - like afm spin order below 22 k. @xcite the ir experiments on cdcr@xmath0s@xmath1 were performed on as - grown single crystals obtained by chemical transport reaction method using cl as a transport agent .
the measurements were repeated on high - purity ceramic samples to check for possible effects on the phonon properties by doping with chlorine . in the single crystals , a cl concentration @xmath23mol% and a concomitant charge carrier density
may be present , but from x - ray studies with synchrotron radiation we certainly can exclude any clustering of cl ions . from detailed wave length sensitive microprobe analysis we are sure that single crystals and ceramics investigated in these studies are very close to ideal stoichiometry .
our ir results in cdcr@xmath0s@xmath1 are in good agreement with those reported by wakamura and arai @xcite on ceramic samples .
cdcr@xmath0s@xmath1 reveals relaxor - type ferroelectric behavior below approximately 130 k. this glass - like transition is dynamic and strongly depends on the measuring frequency .
@xcite soft ferromagnetism with a fully developed magnetic moment of the chromium spins appears below 84.5 k. the measurements on hgcr@xmath0s@xmath1 , documented in this work , were made on ceramic pellets with polished surfaces .
hgcr@xmath0s@xmath1 is dominated by strong ferromagnetic fluctuations , @xcite but undergoes afm spin order below 22 k. @xcite ferroelectric hysteresis loops have been reported close to 70 k , well above the afm phase transition .
@xcite the reflectivity measurements were carried out in the far - infrared range using the bruker fourier - transform spectrometer ifs 113v equipped with a he - bath cryostat and with a split - coil magnet for measurements in external magnetic fields up to 7 t. with this setup we were able to measure the frequency range from 50 to 700 @xmath24 with high precision .
the reflectivity measurements were performed on polished surfaces of the samples . to get an independent estimate of the background dielectric constant ,
the static dielectric constant @xmath25 has been measured for both compounds between 150500 ghz using millimeter spectroscopy .
@xcite in these measurements the real and imaginary part of the dielectric constant can be measured directly via transmission and phase shift , without relying on any kramers - kronig - type analysis for the calculation of the complex permittivity .
room temperature far - infrared spectra of hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 were published by lutz and coworkers @xcite and by wakamura _
et al . _
@xcite incomplete spectra of cdcr@xmath0s@xmath1 have also been published by lee .
@xcite some preliminary far - infrared results on hgcr@xmath0s@xmath1 have been published in ref . .
the temperature dependence of the phonon eigenfrequencies and dampings in cdcr@xmath0s@xmath1 has been reported by wakamura and arai .
@xcite the present experiments were undertaken with special attention to search for fingerprints in the phonon spectra for a possible onset of ferroelectricity in both compounds and to search for phonon splittings in afm hgcr@xmath0s@xmath1 , which could be expected on symmetry arguments alone .
in addition , we also performed far - infrared measurements as function of external magnetic fields to search for magneto - dielectric effects at far - infrared frequencies .
the phonon contribution of the complex dielectric function @xmath26 of an insulating crystal is obtained by calculating the factorized function @xmath27 here @xmath28 , @xmath29 , @xmath30 and @xmath31 correspond to longitudinal ( l ) and transversal ( t ) eigenfrequencies ( @xmath32 ) and dampings ( @xmath33 ) of mode @xmath34 , respectively . @xmath35 results from high - frequency electronic absorption processes and can be experimentally determined from the reflectivity or the index of refraction at frequencies larger than the phonon eigenfrequencies . at normal incidence @xmath36
is related to the reflectivity @xmath37 via @xmath38 using eqs .
( [ equ1 ] ) and ( [ equ2 ] ) the reflectivity spectra can be unambiguously described using a four - parameter fit routine , which has been recently developed by a. kuzmenko .
@xcite in most cases , when the modes are well separated , this fit allows a precise determination of @xmath35 , longitudinal and transversal eigenfrequencies ( @xmath39 ) and dampings ( @xmath40 ) . using this formalism , the dielectric strength @xmath41
can then be calculated via @xmath42 where @xmath25 corresponds to the _ static _ @xmath43 dielectric constant and @xmath44 to the strength of mode @xmath34 , which in the case of non - overlapping modes can be explicitely derived : @xmath45 assuming that the damping of transversal and longitudinal optical phonons is identical , eq .
( [ equ1 ] ) reduces to @xmath46 this is the well known lorentzian lineshape with three independent parameters .
@xmath32 equals @xmath29 of the four - parameter model and @xmath35 is an additional variable , which , of course , has to be considered in both models .
the mode strength @xmath41 can be related to an effective `` ionic '' plasma frequency @xmath47 with @xmath48 the effective plasma frequency of an ir active mode @xmath34 , @xmath49 can then be related to the eigenfrequencies via @xmath50 where @xmath51 is the appropriate high - frequency dielectric constant for mode @xmath34 , including all ionic contributions of modes @xmath52 with @xmath53 ( see eq .
( [ equ4 ] ) ) : @xmath54 finally , the dielectric strength of all polar modes can be related to a sum over all effective plasma frequencies @xmath49 with @xmath55 in favorable cases , using eqs .
( [ equ5])([equ7 ] ) , the ionicity or covalency of bonds involved in polar modes can be calculated .
this is possible on the basis of eq .
( [ equ6 ] ) if the pattern of displacements of a specific phonon mode is known from an eigenvector analysis .
but even if this is not the case , one can arrive at definit conclusions about the effective charge of an ion .
it has been pointed out by scott @xcite and has later been applied by wakamura _ et al . _
@xcite to spinel compounds that for a multi mode system one can assume @xmath56 here @xmath57 is the unit - cell volume , @xmath58 the dielectric permittivity of free space , and @xmath59 is the effective charge of the @xmath60-th ion contributing to a specific polar mode with mass @xmath61 .
the @xmath60 sum is over all atoms in the unit cell .
this equation has to be combined with the expression for lattice charge neutrality @xmath62 in ternary compounds eqs .
( [ equ10 ] ) and ( [ equ11 ] ) do not allow to determine all charges of the ions unambiguously .
in these cases specific assumptions concerning the effective charges of at least one ion have to be made .
@xcite in the present investigation we are also interested in possible soft - mode phenomena accompanying ferroelectric phase transitions . hence we will focus on the generalized lyddane - sachs - teller ( lst ) relation , which can be written as @xmath63 in the case of a soft transverse optic mode the lst relation gives a diverging static dielectric constant , which indeed is observed in many proper ferroelectrics . to get an estimate of the influence of the magnetic exchange interactions on the phonon properties we tried to describe the purely anharmonic temperature dependence of eigenfrequencies and dampings by a simple model , assuming @xmath64 for the temperature dependence of the transverse eigenfrequencies and @xmath65 for the temperature dependence of damping due to canonical anharmonic effects .
@xmath29 and @xmath31 are eigenfrequency and damping of the transversal optical mode @xmath34 and @xmath66 is the debye temperature , which has been determined from an average of the four ir active phonon frequencies .
for the analysis of the anharmonic contributions , debye temperatures of 361 k ( hgcr@xmath0s@xmath1 ) and 372 k ( cdcr@xmath0s@xmath1 ) have been determined . @xmath67 and @xmath68 are the values of phonon frequency and damping of the transverse mode @xmath34 at 0 k. @xmath69 and @xmath70 are mere fitting parameters determining the strength of the anharmonic contributions .
detailed calculations of temperature and frequency dependent anharmonic contributions to eigenfrequencies and dampings can be found in ref . .
( color online ) reflectivity vs wave number in hgcr@xmath0s@xmath1 at 5 k ( open red squares ) and 295 k ( open black circles ) .
the inset shows the phonon response close to 240 @xmath24 as measured in zero magnetic field ( red open squares ) and in a magnetic field of 7 t ( blue open triangles ) . in external magnetic fields the excitation frequency is shifted towards higher energies .
the solid lines are the results of fits as described in the text . ]
the spin - phonon coupling in magnetic semiconductors has been described by baltensperger and helman , @xcite baltensperger , @xcite bresch and dambrogio , @xcite lockwood and cottam @xcite and wesselinowa and apostolov .
@xcite these authors have shown that the frequency shift of a given phonon mode as function of temperature is determined by a spin - correlation function @xmath71 here @xmath72 is the renormalized phonon frequency , @xmath73 denotes the eigenfrequency in the absence of spin - phonon coupling and @xmath74 the spin - phonon coupling constant . on the basis of these theories
wakamura and arai @xcite attempted to describe the experimentally observed frequency shifts @xmath75 in cdcr@xmath0s@xmath1 by taking into account a sum of two terms , namely due to fm and afm exchange , respectively : @xmath76 here @xmath77 and @xmath78 are spin dependent force constants of the lattice vibrations deduced as the squared derivatives of the exchange integrals with respect to the phonon displacements .
@xcite @xmath77 describes the nearest - neighbor ( nn ) fm , @xmath78 the afm next - nearest - neighbor ( nnn ) exchange . using this formalism , negative and positive frequency shifts are depending on the strength of fm or afm exchange interactions , respectively . in cdcr@xmath0s@xmath1 and hgcr@xmath0s@xmath1
we identify the negative contributions to the eigenfrequencies with nn fm cr - s - cr exchange and positive contributions with nnn afm cr - s - cd / hg - s - cr exchange .
hence , vibrations involving cr - s bonds will predominantly be influenced by fm exchange , while eigenmodes of cd - s or hg - s ions will be sensitive to afm exchange .
the temperature dependence of the phonon damping due to spin - phonon interactions has been calculated by wesselinova and apostolov .
@xcite these authors show that in the magnetically ordered phase an additional contribution to the phonon damping arises due to spin - phonon interactions , which vanishes in the paramagnetic phase .
( color online ) temperature dependencies of eigenfrequencies ( transverse optical modes ) of hgcr@xmath0s@xmath1 ( left frames ( a)(d ) ) and cdcr@xmath0s@xmath1 ( right frames ( e)(h ) ) .
the four phonon modes are labeled with increasing wave number .
eigenfrequencies in zero external field are shown as closed red circles , in magnetic fields of 7 t as closed blue squares .
eigenfrequencies in cdcr@xmath0s@xmath1 as determined in zero magnetic field in a ceramic sample are shown by open green circles .
the red solid lines are fits to the eigenfrequencies in zero field assuming a purely anharmonic temperature dependence of the eigenfrequencies utilizing a fixed debye temperature ( eq . ( [ equ13 ] ) , see text ) . ]
figure [ fig1 ] shows the reflectivity @xmath79 of hgcr@xmath0s@xmath1 as measured on ceramic samples at 295 k ( black circles ) and at 5 k ( red squares ) .
already this figure provides clear evidence of a significant temperature dependence of the eigenfrequencies and an enormously strong temperature dependence of damping and dielectric strength of the modes . in a normal anharmonic solid
one expects an almost temperature independent dielectric strength , on increasing temperature eigenfrequencies should moderately decrease , whereas damping increases .
the solid lines through the measured reflectivity correspond to the four - parameter fit using eqs .
( [ equ1 ] ) and ( [ equ2 ] ) .
this fit provides a reasonable description of the experimental results .
deviations appear only at the maximum of mode 3 close to 330 @xmath24 , which probably arise due to imperfections of the sample surface or due to multi - phonon processes .
the main results of the analysis of the reflectivity spectra of ceramic hgcr@xmath0s@xmath1 and single - crystalline cdcr@xmath0s@xmath1 as a function of temperature are shown in figs .
[ fig2 ] and [ fig3 ] .
figure [ fig2 ] shows the eigenfrequencies for hgcr@xmath0s@xmath1 ( fig .
[ fig2](a)(d ) ) and cdcr@xmath0s@xmath1 ( fig .
[ fig2](e)(h ) ) and fig .
[ fig3](a)(h ) the corresponding damping of all modes .
the modes @xmath34 are labeled with the numbers 1 to 4 with increasing wave number .
only parameters of the transverse optical modes are shown . .[tab1]transversal
@xmath80 [ @xmath24 ] and longitudinal @xmath82 [ @xmath24 ] optical eigenfrequencies of cdcr@xmath0s@xmath1 compared with theoretical values predicted by lsda+u ( ref . ) .
the extrapolated eigenfrequencies at 0 k @xmath67 [ @xmath24 ] in the absence of spin - phonon coupling are also listed . [ cols="^,^,^,^,^,^ " , ] in the right frames of figs .
[ fig2 ] and [ fig3 ] we also included eigenfrequencies and phonon damping for cdcr@xmath0s@xmath1 as obtained on high - purity ceramic samples .
ceramic cdcr@xmath0s@xmath1 has been investigated to determine a possible influence of the doping with cl which can not be fully avoided during the single crystal growth process . with respect to the eigenfrequencies ,
the agreement between single crystalline and ceramic results is reasonable .
only mode 3 shows slight deviations of the order of 1 @xmath24 .
however , we have to keep in mind that the fit of mode 3 is not perfect ( see fig . [ fig1 ] ) and
the apparent differences probably can be attributed to the fitting procedure .
more significant deviations appear in the temperature dependence of the damping constants . while the general temperature dependence is rather similar , specifically mode 2 and 4 reveal rather different absolute values .
nevertheless , we conclude from these results that while a small amount of doping significantly influences the dielectric properties at low frequencies , the phonon properties seen via the reflectivity are only barely influenced . in what follows
we only will discuss the results obtained on ceramic hgcr@xmath0s@xmath1 and single - crystalline cdcr@xmath0s@xmath1 .
( color online ) phonon damping of the transverse optical modes versus temperature for hgcr@xmath0s@xmath1 ( left frames ( a)(d ) ) and cdcr@xmath0s@xmath1 ( right frames ( e)(h ) ) . damping constants in zero external field are shown as closed red circles , in magnetic fields of 7 t as closed blue squares .
damping constants in cdcr@xmath0s@xmath1 as determined in a ceramic sample are shown by open green circles .
the solid lines are fits assuming a purely anharmonic temperature dependence of the damping utilizing a fixed debye temperature ( see text ) . ]
the eigenfrequencies for hgcr@xmath0s@xmath1 ( fig .
[ fig2](a)(d ) ) span a range from 70377 @xmath24 , for cdcr@xmath0s@xmath1 ( fig .
[ fig2](e)(h ) ) from 95 to 381 @xmath24 .
this observation indicates that the high - frequency modes in both compounds exhibit almost identical eigenfrequencies and cr and s ions can be involved only in mode 4 .
the frequencies of phonon modes 1 scale with the square root of the inverse masses of mercury and cadmium , pointing towards the fact that in the low - frequency modes mainly the closed - shell metals of the zinc group are involved .
striking similarities between antiferromagnetic hgcr@xmath0s@xmath1 and ferromagnetic cdcr@xmath0s@xmath1 can also be recognized for the temperature dependence of the damping constants . at low temperatures the damping constants of both compounds are small for modes 1 , 2 and 4 , but rather large for mode 3 .
mode 2 reveals a strikingly strong temperature dependence for hgcr@xmath0s@xmath1 as well as for cdcr@xmath0s@xmath1 . the solid lines in figs .
[ fig2 ] and [ fig3 ] correspond to the expected anharmonic temperature dependence of frequencies and dampings as calculated according to eqs .
( [ equ13 ] ) and ( [ equ14 ] ) .
these curves have been derived by fitting the high - temperature ( @xmath83 k ) values of the eigenfrequencies and dampings .
this procedure seems to be straight forward and correct for the eigenfrequencies of all modes in both compounds ( fig .
it also properly describes the temperature dependence of the phonon dampings of modes 3 and 4 for hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 .
the temperature dependence of the damping constants of modes 1 and 2 of both compounds can not be described in this way .
for these two modes a continuous increase of the damping is enhanced by a bump - like feature close to 150 k , which could result from structural instabilities which appear definitely far above the magnetic ordering temperatures .
@xcite we would like to recall that according to theory @xcite one would expect additional contributions to the damping from phonon - magnon scattering in the magnetically ordered phases , which vanish in the paramagnetic states .
this definitely is not observed .
these strong cusplike features in the damping constants of both compounds are rather extended and definitely do neither exhibit a maximum close to @xmath84 in cdcr@xmath0s@xmath1 ( @xmath85 k ) nor close to 60 k in hgcr@xmath0s@xmath1 , where in zero external field the strongest ferromagnetic fluctuations are observed .
@xcite hence we think that an explanation of the damping anomalies of mode 2 of both compounds in terms of ferromagnetic fluctuations is rather unlikely . in recent raman scattering experiments
@xcite pronounced anomalies in intensity and frequency of raman - active modes have been observed , arising at an onset temperature @xmath86 k , which coincides with striking anomalies in the ir properties . already at first sight
it becomes clear that afm hgcr@xmath0s@xmath1 and fm cdcr@xmath0s@xmath1 behave similar : the eigenfrequencies of modes 1 and 2 in both compounds reveal clear positive shifts as compared to normal anharmonic behavior , while modes 3 and 4 exhibit negative shifts when entering the magnetic phase .
this again documents that in modes 3 and 4 the eigenfrequencies are determined mainly from force constants between cr and s ions , which are coupled ferromagnetically . on the other hand ,
cd / hg - s ions are involved in the vibrations of modes 1 and 2 , which obviously are strongly influenced by the afm cr - s - cd / hg - s - cr exchange .
it is important to note that the temperature dependence of the eigenfrequencies looks very similar for both compounds , despite the fact that hgcr@xmath0s@xmath1 reveals antiferromagnetic order below 22 k , while cdcr@xmath0s@xmath1 is ferromagnetic below 84.5 k. no indications of phonon splittings in antiferromagnetic hgcr@xmath0s@xmath1 are observable .
however , both compounds show strong ferromagnetic fluctuations already at rather elevated temperatures and it is probably this fact which explains the similarity in the temperature dependence of the phonons .
strong anomalous contributions are also observed in the temperature dependence of the damping .
modes 1 and 2 exhibit a significant decrease of the damping below 150 k for both compounds .
this unusual behavior could be attributed to the onset of local polar order . on the contrary
, the dampings of modes 3 slightly increase below 100 k , while modes 4 of hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 are very close to the normal temperature dependence of the damping in anharmonic solids .
magnetic field dependence of the eigenfrequencies of hgcr@xmath0s@xmath1 for modes 1 ( a ) to 4 ( d ) at 6 k. the lines are drawn to guide the eye . ] the eigenfrequencies at 5 k are listed in tab .
[ tab1 ] and compared to the results from lsda+u calculations of fennie and rabe .
@xcite here we also included extrapolated eigenfrequencies @xmath67 ( see eq .
( [ equ13 ] ) ) , which correspond to the eigenfrequencies at 0 k in the absence of spin - phonon coupling . with the exception of mode 3 we find good agreement . from tab .
[ tab1 ] we can also infer that the frequency shift at 0 k in cdcr@xmath0s@xmath1 due to spin - phonon coupling @xmath87 is positive and of the order of 5 @xmath24 for modes 1 and 2 , but negative and significantly smaller for modes 3 and 4 . the very same is true for the spin - phonon coupling in hgcr@xmath0s@xmath1 ( fig . [ fig2](a)(d ) ) .
this signals , as outlined above , that the modes 3 and 4 exhibiting negative shifts are only weakly influenced by fm 90@xmath88 cr - s - cr exchange , while modes 1 and 2 are strongly influenced by afm nnn cr - s - cd / hg - s - cr exchange .
this correlates nicely with calculations showing that in the vibrational pattern of modes 3 and 4 only cr and s ions are involved , while a more complex pattern of motion , including all three types of ions , characterizes modes 1 and 2 .
@xcite on hgcr@xmath0s@xmath1 ceramics and on cdcr@xmath0s@xmath1 single crystals additional reflectivity measurements have been performed in external magnetic fields up to 7 t. the inset of fig .
[ fig1 ] shows mode 2 of hgcr@xmath0s@xmath1 at 5 k in zero magnetic field compared to measurements in a field of 7 t. in an external field of 7 t this mode is shifted to higher frequencies by an amount of approximately 1 @xmath24 .
the temperature dependencies of eigenfrequencies and damping constants as obtained in both compounds for all modes at 7 t are included in figs .
[ fig2 ] and [ fig3 ] ( blue squares ) .
as expected for a ferromagnet with a high ordering temperature , the eigenfrequencies remain unchanged within experimental uncertainties for cdcr@xmath0s@xmath1 ( right frames of fig .
[ fig2 ] ) . for antiferromagnetic hgcr@xmath0s@xmath1 , which shows a metamagnetic transition below 1 t ,
@xcite shifts of the order of 1 @xmath24 are observed for @xmath89 k. the shifts are positive for modes 1 and 2 ( see fig .
[ fig2](a ) and ( b ) ) , but negative for modes 3 and 4 ( see fig .
[ fig2](c ) and ( d ) ) . in mode 3
[ fig2](c ) ) the magnetic - field dependent shifts also show up at room temperature .
again we have to remind that the fits of mode 3 have some uncertainty due to a double - peak structure and in this case the differences of the eigenfrequencies as a function of field may signal the experimental uncertainties .
( color online ) temperature dependence of the lyddane - sachs - teller relation for all modes ( lower four frames ) and the generalized lst relation ( upper two frames ) for hgcr@xmath0s@xmath1 ( left frames ( a)(c ) ) and cdcr@xmath0s@xmath1 ( right frames ( d)(f ) ) at zero magnetic fields ( closed red circles ) and in an external magnetic field of 7 t ( closed blue squares ) . the lines are drawn to guide the eye . ]
the most precise measurements of the magnetic - field dependence of the phonon eigenfrequencies can be made at a constant temperature and fixed spectrometer setting .
as representative example , fig .
[ fig4 ] shows the field dependence of all eigenmodes for hgcr@xmath0s@xmath1 at 6 k. as discussed above , the eigenfrequencies of modes 1 and 2 ( fig .
[ fig4](a ) , ( b ) ) increase , while the frequencies of modes 3 and 4 ( fig .
[ fig4](c ) , ( d ) ) decrease with increasing magnetic field .
the frequency shifts are of the order of 1 @xmath24 for mode 2 , but smaller for the other modes .
the strongest gradients for all modes appear well below 1 t , where afm hgcr@xmath0s@xmath1 reveals a transition into a polarized ferromagnetic state .
@xcite much stronger effects as a function of magnetic field are observed for the damping constants ( see fig . [ fig3 ] ) . for all modes in both compounds we can state that , while the overall temperature dependence remains similar with the same characteristic trends , we observe rather large effects in the absolute values .
the reasion of these significant field dependencies of the phonon damping is unclear at present and further investigations are necessary to clarify this question .
( color online ) temperature dependencies of @xmath25 ( closed blue circles ) and @xmath35 ( closed red squares ) for hgcr@xmath0s@xmath1 ( left frame ) and cdcr@xmath0s@xmath1 ( right frame ) .
room - temperature values of @xmath35 as obtained by thz spectroscopy are included ( closed green triangles ) . ]
with a more elaborated analysis we try to get insight into a possible polar short - range - order phase transition above the magnetic phase transition temperatures in both compounds . to do
so we calculated effective plasma frequencies and lst relations .
the lower frames of fig .
[ fig5 ] show the lyddane - sachs - teller relation for modes 1 to 4 for both compounds , the upper two frames of fig .
[ fig5 ] display the generalized lst relation for the static dielectric constant as calculated via eq .
( [ equ12 ] ) . within 2%
these relations remain constant for all modes .
hence , no significant softening of to modes becomes apparent , and it is obvious that a phonon softening of the transverse acoustic modes , as observed in conventional ferroelectrics , plays no role in hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 . for the latter compound this has already been concluded from first principle calculations .
@xcite however , it seems worthwhile to note that beyond experimental uncertainties the lst relation of mode 4 of both compounds shows a slight increase , signaling a slight hardening of lo or softening of to modes obviously related to spin - phonon coupling . in fig .
[ fig5 ] we also included the results of measurements in an external field of 7 t. within experimental uncertainty the results agree with those obtained at zero field .
figure [ fig6 ] shows the temperature dependence of @xmath25 and @xmath35 for hgcr@xmath0s@xmath1 ( left frame ) and cdcr@xmath0s@xmath1 ( right frame ) .
here we also included the results of @xmath25 obtained by thz spectroscopy at room temperature .
both dielectric constants reveal only a moderate temperature dependence .
it is clear that the strong effects in the temperature dependence of the dielectric constants as observed in radio - frequency measurements @xcite are not mirrored in these far - infrared results .
those strong effects must be of relaxational origin and obviously are related to slow polarization fluctuations . at room temperature ,
@xmath35 is approximately 7.0 for hgcr@xmath0s@xmath1 and for cdcr@xmath0s@xmath1 .
for hgcr@xmath0s@xmath1 , @xmath35 has also been determined by lutz _
et al . _
@xcite as 7.4 , compared to 8.5 as obtained by wakamura .
@xcite for cdcr@xmath0s@xmath1 , room temperature values of @xmath90 ( ref . ) , 7.6 ( ref . ) and 7.84 ( ref . ) have been reported . in insulating solids
the dielectric constant at frequencies beyond the ionic vibrations is determined by the electronic polarizability and the band gap .
the electronic polarizability is expected to be somewhat larger in the mercury compound and despite minor effects due to thermal expansion practically temperature independent . but
@xmath35 also scales with the inverse squared band gap .
the band gap is of the order of 1.4 ev for hgcr@xmath0s@xmath1 with a smooth red shift on decreasing temperature .
@xcite in cdcr@xmath0s@xmath1 at room temperature the band gap has been estimated as 1.57 ev exhibiting an unusual blue shift on decreasing temperatures .
@xcite these band gaps could not explain the observed temperature dependence of @xmath35 of the two compounds under consideration which are very similar and suggest a red shift of the band edges from room temperature down to 100 k and a subsequent blue shift in the ferromagnetic phase ( cdcr@xmath0s@xmath1 ) or in the state dominated by ferromagnetic fluctuations ( hgcr@xmath0s@xmath1 ) . and
indeed , it has been suggested , that these reports on the band gaps in the chromium spinels are heavily influenced by the cr@xmath5 crystal field excitations and do not represent the true electronic band edges .
@xcite ( color online ) temperature dependence of the effective plasma frequencies for modes 1 to 4 ( lower four frames ) and for the sum over all plasma frequencies ( upper two frames ) of hgcr@xmath0s@xmath1 ( left frames ) and cdcr@xmath0s@xmath1 ( right frames ) . the lines are drawn to guide the eye . ] in a second step , using eqs .
( [ equ1 ] ) and ( [ equ7 ] ) we calculated the effective `` ionic '' plasma frequencies which , aside from thermal expansion corrections , only depend on the effective charges of the ions under consideration .
the lower four frames of figure [ fig7 ] show the temperature dependence of the plasma frequencies for hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 for all modes , the upper two frames indicate the sum over all plasma frequencies in zero external magnetic field .
the plasma frequencies of both compounds for all modes except for mode 2 show a moderate increase on decreasing temperature , some saturate or go through a slight maximum close to 100 k. these effects are of the order of 10% .
the temperature dependence of mode 2 of both compounds is significantly different and strongly decreases below 150 k. in cdcr@xmath0s@xmath1 the overall effect of this mode is of the order of 25% indicating strongly decreasing effective charges towards low temperatures .
this can only be explained assuming strong charge transfer and/or that the character of the bonds changes considerably .
it could be a first experimental hint that charge ordering phenomena are responsible for the appearance of local polar order .
local charge order is thought to be responsible for the local polar state in la : srmno@xmath91 ( ref . ) and charge - stripe order is the origin of electronic ferroelectricity in lufe@xmath0o@xmath1 .
@xcite the sum over all effective ionic plasma frequencies is dominated by the strongest modes 3 and 4 and reveals the over all behavior with a shallow maximum close to 100 k. the larger ( @xmath92% ) plasma frequency of cdcr@xmath0s@xmath1 as compared to the mercury compound indicates a stronger ionicity for the former compound , a fact that also follows from general considerations . according to pauling s values of electronegativity , both cr - s and hg - s form rather mixed bonds , however , with a significant increase of covalency for the latter .
( color online ) left frame ( a ) : magnetic field dependence of the magnetization in hgcr@xmath0s@xmath1 at 5 k and at 40 k , which is well below and above the antiferromagnetic ordering temperature at 22 k. right frame ( b ) : magnetic field dependence of the frequency shift of the transverse optical phonon mode 2 at 6 k and 40 k. the lines are drawn to guide the eye . ]
the plasma frequencies of cdcr@xmath0s@xmath1 , as derived from these experiments at 5 k , can be compared to those calculated by fennie and rabe @xcite in their lsda+u approach .
good agreement is found for modes 1 - 3 , however for mode 4 at low temperatures we experimentally determined @xmath93 @xmath24 , while theoretically 202 @xmath24 has been calculated .
in addition , from our experiments we determined @xmath41 ( eq .
( [ equ3 ] ) ) of the order of 2.9 at 5 k compared to the theoretical value of 2.3 .
our value is in good agreement with the result by wakamura _
et al . _ , @xcite who found @xmath94 .
corresponding to the smaller plasma frequencies and to the higher covalency , @xmath41 for hgcr@xmath0s@xmath1 is significantly smaller when compared to cdcr@xmath0s@xmath1 and is of the order of 2 for all temperatures ( see fig .
[ fig6](a ) ) . with the experimentally determined effective plasma frequencies as documented in fig .
[ fig7 ] , we tried to get some insight into the bonding of hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 using eqs .
( [ equ10 ] ) and ( [ equ11 ] ) .
utilizing eq .
( [ equ10 ] ) , we can directly calculate the effective plasma frequencies of both compounds assuming the nominal valencies @xmath95 with hg@xmath16/cd@xmath16 , cr@xmath5 and s@xmath96 . for 300 k and with an averaged @xmath35 we calculated theoretical ionic plasma frequencies of 1568 @xmath24 for hgcr@xmath0s@xmath1 and of 1517 @xmath24 for cdcr@xmath0s@xmath1 .
these values have to be compared with the results shown in figs . [ fig7](c ) and
[ fig7](f ) , where we find an average value of the effective plasma frequency of 415 @xmath24 for the mercury and of 515 @xmath24 for the cadmium compound .
hence , we can state that the overall ionicity , i.e. the normalized effective charge , @xmath97 amounts 26% in hgcr@xmath0s@xmath1 and 34% in cdcr@xmath0s@xmath1 .
proceeding one step further one may calculate the effective charges of the different ions in both compounds combining eqs .
( [ equ10 ] ) and ( [ equ11 ] ) .
however , due to the fact that three unknown quantities , namely @xmath98 for each ion , correspond to two equations only , we have to make another ad hoc assumption , namely that the electronegativity of the closed shell transition metal equals that of cr . in other words ,
the relation of the effective charges has to be @xmath99 , or equivalently @xmath100 .
this produces an average reduction of the effective charges of all ions according to the values determined from the overall theoretical and experimental plasma frequencies .
we achieve the effective charges for ( hg / cd , cr , s ) as ( 0.52 , 0.78 , -0.52 ) for the mercury and ( 0.66 , 0.99 , -0.66 ) for the cadmium compound . the effective plasma frequencies as obtained in an external magnetic field of 7 t are not shown .
the magnetic field dependence of the far - infrared intensities does not scale significantly enough with the external magnetic field to compare results on different runs with different adjustments of the spectrometer .
however , as outlined above , the sensitivity is good enough to search for changes of eigenfrequencies , dampings and intensities at a given temperature as a function of external magnetic field without changing the instrument setting .
a representative example of such a measurement is given in the inset of fig .
[ fig1 ] . with
a given setting and at constant temperature we measured the magnetic field dependence of phonon mode 2 in hgcr@xmath0s@xmath1 at 6 k in the afm state ( @xmath101 ) and at 40 k in the paramagnetic state ( @xmath102 ) .
the results are shown in the right frame of fig .
[ fig8 ] and compared to the magnetization at similar temperatures ( fig .
[ fig8](a ) ) .
the frequency shift of the phonon mode is strongly correlated with the magnetization , even revealing a strong meta - magnetic transition in the afm state .
thus , mutual correspondence of magnetization and phonon shift documents a significant magneto - electric coupling even at optical frequencies .
our far - infrared study of hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 was in part motivated to search for experimental evidence for a polar phase transition close to 100 k , above the onset of magnetic order .
the onset of relaxor ferroelectricity has been brought up in refs . and and has reinforced old speculations of polar distortions in spinel compounds at elevated temperatures .
@xcite for cdcr@xmath0s@xmath1 negative thermal expansion below 150 k , @xcite the broadening of bragg reflections in x - ray diffraction experiments at temperatures above the ferromagnetic phase transition at @xmath84 ( ref . ) provided some experimental evidence for short - range structural distortions close to 150 k. in recent synchrotron radiation experiments a ( 200 ) reflection , which is forbidden in the spinel structure , showed up below 200 k with increasing intensity towards 0 k. @xcite the onset of polar distortions also has been claimed in recent raman scattering experiments , @xcite where pronounced anomalies in the raman modes appear below a characteristic temperature @xmath86 k. we certainly can state that in accord with recent lsda+u calculations @xcite we found no evidence for soft - mode behavior , which usually is a fingerprint for ferroelectricity in proper ferroelectrics .
this statement is based on the observations as presented in fig .
[ fig2 ] , but also from the almost constant lst relations as presented in figs .
[ fig5]((c ) , ( f ) ) .
however , we would like to recall that the soft - mode concept , which couples a soft transverse mode to a divergent static dielectric susceptibility in multiferroics may not be the adequate description . for the perovskite - type family of mutliferroics
it has been shown theoretically @xcite and proven experimentally @xcite that electromagnons , i.e. coupled spin - phonon excitations , are the relevant excitations .
the ferroelectric transition in these compounds is then rather driven by a soft magnetic excitation .
nevertheless , in this work we provide evidence for strong spin - phonon coupling .
the onset of these effects ( figs . [ fig2 ] and [ fig3 ] ) certainly can be detected far above the magnetic transition temperature and it is unlikely that these effects are driven by ferromagnetic fluctuations . in both compounds mode 2 behaves rather unusual , with a broad cusp in the temperature dependence of the damping constant and specifically a significant anomaly in the plasma frequency .
both anomalies are located close to 150 k. in summary , the main outcome of this work is , that + i ) within the experimental resolution , there is no splitting of phonon modes in afm hgcr@xmath0s@xmath1 , and it behaves very similar like fm cdcr@xmath0s@xmath1 , demonstrating that both compounds are dominated by ferromagnetic exchange interactions .
+ ii ) there are strong frequency shifts in the phonon modes from normal anharmonic behavior , which are positive for modes 1 and 2 , but negative for modes 3 and 4 . for both compounds
these shifts appear well above the onset of magnetic order . for cdcr@xmath0s@xmath1 ,
a detailed phonon study in zero magnetic field on ceramic samples has been reported already by wakamura and arai some decades ago .
@xcite the results of both studies are in good agreement , however , with differences in the temperature dependence of the dielectric constants @xmath25 and @xmath35 .
+ iii ) in afm hgcr@xmath0s@xmath1 we find significant shifts of the eigenfrequencies by external magnetic fields , which scale perfectly well with the macroscopic magnetization .
+ iv ) the plasma frequencies of both compounds reveal a moderate increase on decreasing temperature with the exception of mode 2 , which exhibits a strong decrease below 150 k. + v ) the lyddane - sach - teller relations of all modes behave rather normal as function of temperature and magnetic field . within experimental uncertainty , only mode 4 reveals a moderate increase in the lst relation in both compounds .
+ vi ) the overall ionicity as determined from the sum over all ionic effective plasma frequencies is 28% for hgcr@xmath0s@xmath1 and 33% for cdcr@xmath0s@xmath1 , under the assumption that the normal valency would produce 100% ionic bonds .
stimulating discussions and helpful comments by p. lemmens are gratefully acknowledged .
this work partly was supported by the deutsche forschungsgemeinschaft through the german research collaboration sfb 484 ( university of augsburg ) .
10 p. k. baltzer , p. j. wojtowicz , m. robbins , and e. lopatin , phys . rev . * 151 * , 367 ( 1966 ) .
t. rudolf , ch .
kant , f. mayr , j. hemberger , v. tsurkan , and a. loidl , new j. phys .
* 9 * , 76 ( 2007 ) .
a. p. ramirez , _ handbook of magnetic materials _ , edited by k.
h. j. buschow ( elsevier science , north - holland , 2001 ) , vol .
13 , p. 423 .
y. yamashita and k. ueda , phys .
lett . * 85 * , 4960 ( 2000 ) .
o. tchernyshyov , r. moessner , and s. l. sondhi , phys .
lett . * 88 * , 067203 ( 2002 ) .
a. b. sushkov , o. tchernyshyov , w. ratcliff , s. w. cheong , and h. d. drew , phys .
lett . * 94 * , 137202 ( 2005 ) .
j. hemberger , t. rudolf , h .- a .
krug von nidda , f. mayr , a. pimenov , v. tsurkan , and a. loidl , phys .
lett . * 97 * , 087204 ( 2006 ) .
t. rudolf , ch .
kant , f. mayr , j. hemberger , v. tsurkan , and a. loidl , phys .
rev . b * 75 * , 052410 ( 2007 ) .
j. hemberger , p. lunkenheimer , r. fichtl , h .- a .
krug von nidda , v. tsurkan , and a. loidl , nature * 434 * , 364 ( 2005 ) .
s. weber , p. lunkenheimer , r. fichtl , j. hemberger , v. tsurkan , and a. loidl , phys .
96 * , 157202 ( 2006 ) .
n. w. grimes , j. phys . c * 6 * , l78 ( 1973 ) ; n. w. grimes , philos .
* 26 * , 1217 ( 1972 ) ; l. hwang , a. h. heuer , t. e. mitchell , philos . mag .
* 28 * , 241 ( 1973 ) ; a. h. heuer , t. e. mitchell , j. phys . c * 8 * , l541 ( 1975 )
. h. schmid and e. ascher , j. phys .
c * 7 * , 2697 ( 1974 ) . c. j. fennie and k. m. rabe , phys .
b * 72 * , 214123 ( 2005 ) .
j. hemberger , p. lunkenheimer , r. fichtl , s. weber , v. tsurkan , and a. loidl , phase trans .
* 79 * , 1065 ( 2006 ) .
v. tsurkan , unpublished .
j. bosman and c. crevecoeur , j. phys .
chem . solids * 29 * , 109 ( 1968 ) ; c. crevecoeur and h. j. de wit , solid state commun . * 6 * , 843 ( 1968 ) .
lunkenheimer , a. loidl , c. r. ottermann , and k. bange , phys .
b * 44 * , 5927 ( 1991 ) .
v. gnezdilov , p. lemmens , yu . g. pashkevich , p. scheib , ch .
payen , k. y. choi , j. hemberger , a. loidl , and v. tsurkan , to be published , cond - mat/0702362 .
k. wakamura and t. arai , j. appl .
63 * , 5824 ( 1988 ) .
v. tsurkan , j. hemberger , a. krimmel , h .- a .
krug von nidda , p. lunkenheimer , s. weber , v. zestrea , and a. loidl , phys .
rev . b * 73 * , 224442 ( 2006 ) .
l. c. chapon , p. g. radaelli , y. s. hor , m. t. f. telling , and j. f. mitchell , _ preprint _ cond - mat/0608031 .
b. gorshunov , a. volkov , i. spektor , a. prokhorov , a. mukhin , m. dressel , s. uchida , and a. loidl , int .
j. infrared and millimeter waves * 26 * , 1217 ( 2005 )
. h. d. lutz , g. wschenbach , g. kliche , and h. haeuseler , j. sol .
state chem . *
48 * , 196 ( 1983 ) ; h. d. lutz , b. mller , and h. j. steiner , j. sol
. state chem .
* 90 * , 54 ( 1991 ) ; j. zwinscher and h. d. lutz , j. sol .
state chem . * 118 * , 43 ( 1995 ) ; j. zwinscher and h. d. lutz ; j. alloys and compounds * 219 * , 103 ( 1995 ) .
k. wakamura , t. ogawa , t. arai , jap .
* 19 * , supplement 19 - 3 , 249 ( 1980 ) . t. h. lee , j. appl .
phys . * 42 * , 1441 ( 1971 ) .
reffit by a. kuzmenko , university of geneva , version 1.2.44 ( 2006 ) , http://optics.unige.ch/alexey/reffit.html .
j. f. scott , phys .
b * 4 * , 1360 ( 1971 ) . k. wakamura and t. arai , phys .
b * 24 * , 7371 ( 1981 ) .
r. a. cowley , adv .
* 12 * , 421 ( 1963 ) .
w. baltensperger and j. s. helman , helv .
acta * 41 * , 668 ( 1968 ) .
w. baltensperger , j. appl .
phys . * 41 * , 1052 ( 1970 ) .
p. bresch and f. dambrogio , phys .
status solidi b * 50 * , 513 ( 1972 ) .
d. j. lockwood and m. g. cottam , j. appl .
* 64 * , 5876 ( 1988 ) .
j. m. wesselinowa and a. t. apostolov , j. phys . : condens .
matter * 8 * , 473 ( 1996 ) .
h. w. lehmann and g. harbeke , phys .
rev . b * 1 * , 319 ( 1970 ) .
g. harbeke and h. pinch , phys .
* 17 * , 1090 ( 1966 ) .
s. wittekoek and p. f. bongers , solid state commun . * 7 * , 1719 ( 1969 ) .
r. f. mamin , i. k. bdikin , and a. l. kholkin , _ preprint _ cond - mat/0706.1164 .
y. zhang , h. x. yang , c. ma , h. f. tian , and j. q. li , phys . rev .
lett . * 98 * , 247602 ( 2007 ) .
h. gbel , j. magn .
* 3 * , 143 ( 1976 ) .
a. krimmel , unpublished .
h. katsura , a. v. balatsky , and n. nagaosa , phys .
lett . * 98 * , 027203 ( 2007 ) .
i. e. chupis , _ preprint _ cond - mat/0702636 .
a. pimenov , a. a. mukhin , v. yu .
ivanov , v. d. travkin , a. m. balbashov , and a. loidl , nature phys .
* 2 * , 97 ( 2006 ) .
a. pimenov , t. rudolf , f. mayr , a. loidl , a. a. mukhin , and a. m. balbashov , phys .
b * 74 * , 100403(r ) ( 2006 ) . | polar phonons of hgcr@xmath0s@xmath1 and cdcr@xmath0s@xmath1 are studied by far - infrared spectroscopy as a function of temperature and external magnetic field .
eigenfrequencies , damping constants , effective plasma frequencies and lyddane - sachs - teller relations , and effective charges are determined .
ferromagnetic cdcr@xmath0s@xmath1 and antiferromagnetic hgcr@xmath0s@xmath1 behave rather similar . both compounds
are dominated by ferromagnetic exchange and although hgcr@xmath0s@xmath1 is an antiferromagnet , no phonon splitting can be observed at the magnetic phase transition .
temperature and magnetic field dependence of the eigenfrequencies show no anomalies indicating displacive polar soft mode behavior .
however , significant effects are detected in the temperature dependence of the plasma frequencies indicating changes in the nature of the bonds and significant charge transfer .
in hgcr@xmath0s@xmath1 we provide experimental evidence that the magnetic field dependence of specific polar modes reveal shifts exactly correlated with the magnetization showing significant magneto - dielectric effects even at infrared frequencies . |
important structural and dynamical properties of networked systems can be obtained from the eigenvalues and eigenvectors of matrices associated to their graph representations .
the spectra of the adjacency , laplacian and normalized laplacian matrices of a graph provide information on the diameter , degree distribution , community structure , paths of a given length , local clustering , total number of links , number of spanning trees and many more invariants @xcite .
also , dynamic aspects of a network , such as its synchronizability and random walk properties , can be obtained from the eigenvalues of the laplacian and normalized laplacian matrices from which it is possible to calculate some interesting graph invariants like the estrada index and the laplacian energy @xcite . in the last years
, there has been a particular interest in the study of the eigenvalues and eigenvectors of the normalized laplacian matrix , since many measures for random walks on a network are linked to them .
these include the hitting time , mixing time and kemeny s constant which can be considered a measure of the efficiency of navigation on the network , see @xcite .
this paper is organized as follows .
first , in sec [ pre ] , we recall an operation called triangulation that can be applied to any simple connected graph . in sec
[ sec3 ] , we determine the spectra of the normalized laplacians of iterated triangulations of any simple connected graph and we discuss their structure .
finally , in sec [ app ] , we use the results found to calculate three significant invariants for an iterated triangulation of a graph : the multiplicative degree - kirchhoff index , its kemeny s constant and the number of spanning trees .
let @xmath0 be any simple connected graph with vertex set @xmath1 and edge set @xmath2 .
let @xmath3 denote the number of vertices of @xmath4 and @xmath5 its number of edges .
we consider a graph operation known in different contexts as henneberg 1 move @xcite , @xmath6-extension @xcite , vertex addition @xcite and triangulation @xcite . here
we will use the latter name as it is more common in recent literature .
[ define ] .
the triangulation of @xmath4 , denoted by @xmath7 , is the graph obtained by adding a new vertex corresponding to each edge of g and by joining each new vertex to the end vertices of the edge corresponding to it .
we denote @xmath8 .
the @xmath9-triangulation of @xmath4 is obtained through the iteration @xmath10 and @xmath11 and @xmath12 will be the total number of vertices and edges of @xmath13 . from this definition clearly
@xmath14 and @xmath15 and thus we have : @xmath16 figure [ fig.1 ] shows an example of iterated triangulations where the initial graph is @xmath17 or the triangle graph . in this particular case , the resulting graph is known as scale - free pseudofractal graph @xcite that exhibits a scale - free and small - world topology .
the study of its structural and dynamic properties has produced an abundant literature , since it is a good deterministic model for many real - life networks , see @xcite and references therein . .
red vertices denote the initial @xmath17 vertices while green and black vertices have been introduced to obtain @xmath18 and @xmath19 , respectively.,scaledwidth=40.0% ] the triangulation operator @xmath20 turns each edge of a graph into a triangle , i.e. a cycle of length @xmath21 . as we will see in the next section
, this property has a relevant impact on the structure of the spectrum of the normalized laplacian matrix of @xmath13 .
we label the vertices of @xmath13 from @xmath22 to @xmath11 .
let @xmath23 be the degree of vertex @xmath24 of @xmath13 , then @xmath25 denotes its diagonal degree matrix and @xmath26 the adjacency matrix , defined as a matrix with the @xmath27-entry equal to @xmath22 if vertices @xmath24 and @xmath28 are adjacent and @xmath6 otherwise .
the laplacian matrix of @xmath13 is @xmath29 .
the probability transition matrix for random walks on @xmath13 or markov matrix is @xmath30 .
@xmath31 can be normalized to obtain a symmetric matrix @xmath32 . @xmath33
the @xmath27th entry of @xmath32 is @xmath34 .
matrices @xmath32 and @xmath31 share the same eigenvalue spectrum .
the normalized laplacian matrix of @xmath13 is @xmath35 where @xmath36 is the identity matrix with the same order as @xmath32 .
we denote the spectrum of @xmath37 by @xmath38 where @xmath39 .
the zero eigenvalue is unique , due to the existence of a stationary distribution for random walks .
we denote the multiplicity of @xmath40 as @xmath41 .
. shows a one - to - one correspondence between the spectra of @xmath37 and @xmath32 .
the spectrum of the normalized laplacian matrix of a graph can provide us with important structural and dynamical information about the graph , see for example @xcite .
if we replace each edge of a simple connected graph @xmath4 by a unit resistor , we obtain an electrical network @xmath42 associated with @xmath4 .
the resistance distance @xmath43 between vertices @xmath24 and @xmath28 of @xmath4 is equal to the effective resistance between the two corresponding vertices of @xmath42 @xcite .
the multiplicative degree - kirchhoff index of @xmath4 is defined , as in @xcite : @xmath44 this index is different from the classical kirchhoff index @xcite , @xmath45 , as it considers also the degree distribution of the graph . see also @xcite for recents results on these and related indices . for a @xmath46-regular graph @xmath47 .
it is known @xcite that @xmath48 can be expressed in terms of the spectrum @xmath49 of the normalized laplacian matrix @xmath50 of @xmath4 . thus @xmath51 where @xmath52 .
for @xmath53 , we can obtain the degree - kirchhoff index of the @xmath9-triangulation of @xmath4 as : @xmath54 where @xmath39 are the eigenvalues of @xmath37 .
the knowledge of @xmath55 also allows the calculation of an important graph invariant associated with random walks .
given a graph @xmath4 , the kemeny s constant @xmath56 , or average hitting time , is the expected number of steps required for the transition from a starting vertex @xmath24 to a destination vertex , which is chosen randomly according to a stationary distribution of unbiased random walks on @xmath4 ( see @xcite for more details ) .
@xmath56 is a constant which is independent of the selection of the starting vertex @xmath24 @xcite .
moreover , the kemeny s constant can be computed as the sum of all reciprocal eigenvalues of the normalized laplacian of @xmath4 , except @xmath57 , see @xcite . therefore ,
given the spectrum @xmath55 of @xmath13 we can write : @xmath58 moreover , we have : @xmath59 this equation reflects the fact that , for a connected graph , the resistance distance can be related to random walks @xcite .
the last graph invariant considered is the number of spanning trees of a graph @xmath4 .
a spanning tree is a subgraph of @xmath4 that includes all the vertices of @xmath4 and is a tree . by using a result from chung
@xcite we can find the total number of spanning trees of @xmath13 in terms of @xmath55 , its normalized laplacian spectrum , and the degrees of all the vertices : @xmath60 in the next section we give a closed analytical expression for this invariant for any value of @xmath53 .
in this section we find an exact analytical expression for the spectrum @xmath55 of @xmath37 , the normalized laplacian of the @xmath9-triangulation graph @xmath13 .
we show that this spectrum can be obtained iteratively from the spectrum of any simple connected graph @xmath4 .
[ eig1 ] let @xmath61 be any eigenvalue of @xmath37 , @xmath62 , such that @xmath63 and @xmath64 .
then @xmath65 is an eigenvalue of @xmath66 and its multiplicity , denoted @xmath67 , is the same than the multiplicity @xmath68 of the eigenvalue @xmath61 of @xmath37 .
moreover , @xmath69 .
let @xmath70 be the set of all the newly added vertices in @xmath13 and @xmath71 be its complement .
this is also equivalent to saying that @xmath71 contains all the vertices inherited from @xmath72 . for convenience , in the following when we refer to any vertex of @xmath71 , we also refer to the corresponding vertex of @xmath72 .
let @xmath73 be an eigenvector with respect to the eigenvalue @xmath61 of @xmath37 . then , as @xmath74
, @xmath75 is also an eigenvector of @xmath32 corresponding to the eigenvalue @xmath76 .
hence @xmath77 for any vertex @xmath78 , let @xmath79 denote the set of the new neighbors of vertex @xmath24 in @xmath13 and @xmath80 the set of its old neighbors
. there exists , from the definition of triangulation , a bijection between @xmath81 and @xmath82 .
eq . can be written as @xmath83 if @xmath84 denotes the degree of vertex @xmath24 of @xmath13 , this equation leads to @xmath85 for any @xmath86 , we have a similar relationship @xmath87 where vertex @xmath88 is the other neighbor of vertex @xmath89 in @xmath13 ( excluding the vertex @xmath24 ) . combining eq . and
eq . we have : @xmath90 therefore , @xmath91 holds for @xmath92 .
. shows that @xmath93 is an eigenvalue of @xmath94 .
thus @xmath65 is an eigenvalue of @xmath66 and @xmath95 one associated eigenvector .
@xmath75 can be totally determined by @xmath96 using eq . and @xmath97 .
suppose now that @xmath98 .
this means that ithere should exist an extra eigenvector @xmath99 associated to @xmath65 without a corresponding eigenvector in @xmath37 .
. provides @xmath99 with an associated eigenvector of @xmath32 since @xmath100 , and this contradicts our assumption .
therefore , @xmath101 . when @xmath102 , @xmath103 can be determined by eq .
and we find that eq
. holds for any @xmath104 and @xmath105 , and thus @xmath106 .
+ [ eig2 ] let @xmath61 be any eigenvalue of @xmath66 such that @xmath107 . then @xmath108 is an eigenvalue of @xmath37 . besides , @xmath109 .
this is a direct consequence of lemma [ eig1 ] + [ remark1 ] lemmas [ eig1 ] and [ eig2 ] show that a large fraction of the eigenvalues of @xmath37 are obtained from the spectrum of @xmath66 as each eigenvalue , except @xmath110 , of @xmath66 will generate a unique eigenvalue of @xmath37 . as we know the multiplicity of @xmath111
, we conclude that we have @xmath112 eigenvalues of @xmath55 .
the rest of the spectrum consists of @xmath22s . from the perron - frobenius theorem
@xcite we know that the largest absolute value of the eigenvalues of @xmath32 is 1 .
since there exists a unique stationary distribution for random walks on @xmath13 , the multiplicity of eigenvalue @xmath22 of @xmath32 is @xmath22 .
thus the multiplicity of the smallest eigenvalue of @xmath37 is also 1 , i.e. @xmath113 .
we wish to find the multiplicity of 2 , the largest eigenvalue of @xmath37 , in the context of markov chains .
a markov chain is aperiodic if and only if the smallest eigenvalue of the markov matrix is different from @xmath114 .
thus , the largest eigenvalue of @xmath37 is not equal to @xmath110 if and only if random walks on @xmath13 are aperiodic .
since each edge of @xmath13 , @xmath115 , belongs to an odd - length cycle ( a triangle ) , the graph is aperiodic @xcite .
thus @xmath116 holds for @xmath115 .
note , however , that the value of @xmath117 depends on the structure of the initial graph @xmath4 , since it may be periodic @xcite .
let @xmath118 be any finite multiset of real number .
the multiset @xmath119 is defined as @xmath120 [ theo1 ] the spectrum @xmath55 of @xmath37 is @xmath121 where @xmath122 from results obtained so far , the multiplicity of the eigenvalue @xmath22 of @xmath37 can be determined indirectly : @xmath123 since @xmath124 for @xmath125 , the theorem is true . the two particular eigenvalues @xmath22 and @xmath111 of @xmath37 are called exceptional eigenvalues for the family of matrices @xmath126 , for which the spectra shows self - similarity characteristics @xcite .
theorem [ theo1 ] confirms a previous result on the spectrum of the scale - free pseudofractal graph , obtained with other related methods by one of the authors @xcite .
from the spectrum of @xmath37 , the normalized laplacian of the @xmath9-triangulation graph @xmath13 , we compute in this section some relevant invariants related to its structure . we give closed formulas for the multiplicative degree - kirchhoff index , kemeny s constant and the number of spanning trees of @xmath13 .
the results depend only on @xmath9 and some invariants of graph @xmath4 .
[ stan ] for a simple connected graph @xmath4 , the multiplicative degree - kirchhoff indices of @xmath13 and @xmath72 , @xmath115 , are related as follows : @xmath127 thus , the general expression for @xmath128 is @xmath129 we use eq . and theorem [ theo1 ] : @xmath130 where @xmath131 are eigenvalues of @xmath37 .
finally : @xmath129 [ kemny ] the kemeny s constant for random walks on @xmath13 can be obtained from @xmath132 through @xmath133 the general expression is @xmath134 this result is a direct consequence of theorem [ stan ] and eq . .
[ sptree ] the number of spanning trees of @xmath13 , @xmath115 , is : @xmath135 where @xmath136 . from eq . and the definition of triangulation of a graph : @xmath137 where @xmath138 are the eigenvalues of @xmath139 .
we obtain , for @xmath115 : @xmath140 therefore , the following equality @xmath141 holds for any @xmath115 , and finally we have : @xmath142 where @xmath143 .
we recall that former expressions for the multiplicative degree - kirchhoff index , kemeny s constant and the total number of spanning trees of an @xmath9-triangulation graph are valid given any initial simple connected graph @xmath4 .
we note here that the multiplicative degree - kirchhoff index of @xmath13 has been obtained recently by yang and klein @xcite by using a counting methodology not related with spectral techniques and also that some results for the particular case of @xmath46-regular graphs have been obtained in @xcite from the characteristic polynomials of the graphs .
our result confirms both their calculations and the usefulness of the concise spectral methods described here .
finally , the values for kemeny s constant and the number of spanning trees of a scale - free pseudofractal graph , which is another particular case of @xmath9-triangulation graph , were obtained by one of the authors in @xcite .
we find the same expressions by using @xmath144 in eq .
and eq . .
we have obtained a closed - form expression for the normalized laplacian spectrum of a @xmath9-triangulation of any simple connected graph .
this was possible through the analysis of the eigenvectors in relation to adjacent vertices at different iteration steps .
our method could be also applied to find the spectra of other graph families which are constructed iteratively .
we have also studied the spectrum structure .
the knowledge of this spectrum makes very easy , in relation to formerly known methods , the calculation of several invariants related with structural and dynamic characteristics of the iterated triangulations of a graph . as an example
, we find the multiplicative degree - kirchhoff index , kemeny s constant and the number of spanning trees .
this work was supported by the national natural science foundation of china under grant no .
f.c . was supported by the ministerio de economia y competitividad ( mineco ) , spain , and the european regional development fund under project mtm2011 - 28800-c02 - 01 .
m. bianchi , a. cornaro , j. l. palacios , a. torriero , bounding the sum of powers of normalized laplacian eigenvalues of graphs through majorization methods , match commun .
70 ( 2013 ) 707716 .
d. bonchev , a. t. balaban , x. liu , d. j. klein , molecular cyclicity and centricity of polycyclic graphs .
i. cyclicity based on resistance distances or reciprocal distances , int .
50 ( 1994 ) 120 .
s. butler , algebraic aspects of the normalized laplacian , in : a. beveridge , j. griggs , l. hogben , g. musiker , p. tetali ( eds . ) , recent trends in combinatorics , vol . to appear of the i m a volumes in mathematics and its applications , i m a , 2016 .
a. nixon , e. ross , one brick at a time : a survey of inductive constructions in rigidity theory , in : r. connelly , a. i. weiss , w. whiteley ( eds . ) , rigidity and symmetry , vol .
70 of fields institute communications , springer , 2014 , pp . 303324 | the eigenvalues of the normalized laplacian of a graph provide information on its topological and structural characteristics and also on some relevant dynamical aspects , specifically in relation to random walks . in this paper
we determine the spectra of the normalized laplacian of iterated triangulations of a generic simple connected graph . as an application
, we also find closed - forms for their multiplicative degree - kirchhoff index , kemeny s constant and number of spanning trees .
complex networks , normalized laplacian spectrum , graph triangulations , degree - kirchhoff index , kemeny constant , spanning trees |
the riemann zeta function is defined , for @xmath5 , as @xmath6 .
this function has a singularity ( in fact , a simple pole ) at @xmath7 , which corresponds to the divergence of the harmonic series . for real values of @xmath8 , @xmath9
, the series converges to a real number between @xmath10 and @xmath11 , according to the integral test . for positive integer values of @xmath8 , @xmath9
, one has a well - known formula by euler ( see ref .
@xcite and references therein ) : @xmath12 @xmath13 being a positive integer . here
, @xmath14 are bernoulli numbers , i.e. the rational coefficients of @xmath15 in the taylor series expansion of @xmath16 , @xmath17 . for @xmath18 ,
on the other hand , no analogous expression is currently known .
this scenario has a ` reverse ' counterpart on the values of the dirichlet beta function , defined as @xmath19 , @xmath20 , in the sense that , for integer values of @xmath21 , the following analogue of eq .
is known : , since @xmath22 and @xmath23 . ]
@xmath24 where @xmath25 are euler numbers , i.e. the ( integer ) coefficients of @xmath15 in the taylor expansion of @xmath26 , @xmath27 . for @xmath28 ,
no analogous expression is known , not even for @xmath29 , known as the catalan s constant @xmath30 .
some progress in this direction was reached by klbig ( 1996 ) , who proved that @xcite : @xmath31 where @xmath32 is the polygamma function , i.e. the @xmath13-th derivative of @xmath33 , the digamma function .
, in turn , is defined as the logarithm derivative of @xmath34 , the classical gamma function . ]
although this identity resembles euler s formula , the arithmetic nature of @xmath35 is currently unknown . in a very recent paper ,
i have succeeded in applying the dancs and he series expansion method , as introduced in ref .
@xcite , to find similar formulae for @xmath36 @xcite .
i could then prove that @xmath37 since both @xmath38 and @xmath39 converge absolutely , the zeta series in eq .
equals the difference of these individual series .
however , the best i could do there in ref .
@xcite was to investigate the pattern of the analytical results found for the first series , eq .
, for small values of @xmath13 .
this did lead me to _ conjecture _ a formula for its summation which should be valid for any positive integer @xmath40 . here in this work ,
i make use of a classical wilton s formula to prove the above mentioned conjecture .
the proof of another conjecture , involving the zeta series in eq .
, follows from eq . .
finally , on aiming at a generalization of these formulae i substitute the fractions @xmath41 and @xmath42 by @xmath43 , @xmath44 being any non - null real number with @xmath45 . this has led me to develop a simpler , more direct proof of a theorem by katsurada ( 1999 ) @xcite .
the proof of the first conjecture in ref .
@xcite , involving the zeta series in eq .
, above , follows from a classical result by wilton for a rapidly convergent series representation for @xmath47 @xcite .
hereafter , we define @xmath48 as the @xmath13-th harmonic number . [
le : wilton ] let @xmath49 be a positive integer
. then @xmath50 for a skeleton of the proof , see the original work of wilton @xcite . for a more complete proof , see sec .
4.2 ( in particular , pp .
412413 ) of ref .
@xcite , a systematic collection of zeta series recently published by srivastava and choi .
this lemma allows us to prove the first conjecture of ref .
@xcite , namely that in its eq .
( 29 ) .
[ teo : zetaser1 ] let @xmath49 be a positive integer
. then @xmath51 .\end{aligned}\ ] ] from lemma [ le : wilton ] , we know that @xmath52 by isolating the last term , one has @xmath53 by substituting @xmath54 in the above equation , one finds @xmath55 a division by @xmath11 completes the proof .
for instance , on putting @xmath56 in this theorem we get a rapidly converging series representation for @xmath57 , namely @xmath58 this formula converges to @xmath1 much faster than @xmath59 and even than @xmath60}}$ ] , a central binomial series used by apry @xcite .
numerical computation using _ mathematica _ shows that only ten terms of the above zeta series are enough for a ten decimal places accuracy .
now we can use eq . to prove the other conjecture raised in ref .
@xcite , namely that in its eq .
( 30 ) .
[ teo : zetaser2 ] let @xmath49 be a positive integer . then @xmath61 \ !
.\end{aligned}\ ] ] from eq .
, we know that @xmath62 on substituting the first zeta series at the left - hand side by the result proved in theorem [ teo : zetaser1 ] , one has @xmath63 by isolating the remaining zeta series , one finds @xmath64 a division by @xmath11 completes the proof .
on putting @xmath65 in theorem [ teo : zetaser2 ] we get a rapidly converging series representation for @xmath66 , namely @xmath67 this formula converges much faster than @xmath68 and even faster than @xmath69}$ ] , a rapidly converging central binomial series discovered by ramanujan @xcite .
numerically , only six terms of the above zeta series are enough for a result accurate to ten decimal places . after an extensive search for similar zeta series in literature ,
i have found a formula by srivastava and tsumura ( 2000 ) in ref .
@xcite ( see also ref .
@xcite , p. 421
in fact , this formula could be taken into account for an independent proof of our theorem [ teo : zetaser2 ] , after some simple manipulations , as the reader can easily check .
for this , it will be useful to know that @xmath70 is the hurwitz zeta function , for which it is well - known that @xmath71 ( see , e.g. , eq . ( 25.11.12 ) of ref .
@xcite ) . then @xmath72 and , from our eq .
, it follows that @xmath73 , where @xmath74 was substituted by @xmath75 . ] on investigating the substitution of @xmath76 and @xmath77 by @xmath43 , @xmath78 being any non - null real number with @xmath79 , in the zeta series treated in the previous theorems , i have found the following general result .
[ teo : geral ] let @xmath49 be a positive integer and @xmath44 be any real number with @xmath80 . then @xmath81 \ !
.\end{aligned}\ ] ] from the euler s product formula for the sine function , we know that , for all non - null real @xmath82 with @xmath83 , the following identity holds : @xmath84 on taking the logarithm on both sides , we have @xmath85 } = \sum_{n=1}^\infty{\ln{\left ( 1-\frac{z^2}{n^2 } \right ) } } = - \ , \sum_{n=1}^\infty { \left ( \sum_{k=1}^\infty { \frac{z^{2k}}{k \ : n^{2k } } } \right ) } , \ ] ] where the logarithm was expanded in a taylor series in the last step .
this implies that @xmath86 } = - \ , \sum_{k=1}^\infty { \left ( \ , \sum_{n=1}^\infty { \frac{1}{n^{2k } } } \right ) \ , \frac{z^{2k}}{k } } = -\sum_{k=1}^\infty{\frac{z^{2k}}{k } \ , \ , \zeta{(2k)}}\ ] ] and then @xmath87 on multiplying both sides by @xmath88 , @xmath13 being a positive integer , and integrating from @xmath89 to @xmath90 , @xmath90 being any non - null real in the interval @xmath91 $ ] , we have @xmath92 let us solve each definite integral carefully .
the first integral in eq . can be expanded in a trigonometric series as follows . since , for all @xmath93 , @xmath94 } = - \sum_{k=1}^\infty{\frac{\cos{(k \ , \theta)}}{k } } \ , , \ ] ] then , on substituting @xmath95 in the integral , one finds @xmath96 on integrating by parts , one has @xmath97 on integrating by parts again and again , we find , after some algebra , that @xmath98 the second integral in eq
. is readily solved by noting that @xmath99 } \nonumber \\ = \ , \frac{x^{2n+1}}{2n+1 } \ ,
\left [ \ln{|2\,\pi\,x| } + \sum_{\ell=1}^{2n+1}{\frac{(-1)^\ell}{\ell } \ : \binom{2n+1}{\ell } } \right ] .
\quad \label{eq : zint2}\end{aligned}\ ] ] the third integral , i.e. the one at the right - hand side of eq . , can be written in the form of a zeta series on integrating it by parts directly .
this yields @xmath100 finally , on substituting the results in eqs . , , and on the integrals in eq .
, we find @xmath101 = - 2 \ , ( 2n ) ! \ , \sum_{k=1}^\infty{\frac{(2k-1 ) ! \ , \zeta(2k)}{(2k+2n+1 ) ! } \ :
x^{2k } } .
\qquad \label{eq : zexpr}\end{aligned}\ ] ] on dividing both sides by @xmath102 $ ] , we have @xmath103 where we have made use of the binomial sum @xmath104 , which is easily proved by induction on @xmath40 . on extracting the last term of the sum involving odd zeta values ( i.e. , that for @xmath105 ) we bring up the desired result . our proof of theorem [ teo : geral ] , above ,
has allowed us to detect some typos in theorem 2 of ref .
@xcite , as well as some mistakes in its proof , which is based upon the mellin transform technique .
in fact , the formula as printed in katsurada s paper can not be correct because as @xmath106 the left - hand side approaches @xmath107 whereas the right - hand side approaches zero .
unfortunately , that incorrect formula is reproduced on p. 442 of ref .
@xcite , which is currently the main reference on zeta series in literature .
our theorem [ teo : geral ] corrects both the absence of a modulus in the argument of the logarithm and the ` @xmath108 ' sign preceding the sine series in eq .
( 1.6 ) of ref .
@xcite .
interestingly , the formula in our theorem [ teo : geral ] can be reduced to a more appropriate form for both symbolic and numerical computations .
this is easily obtained by noting that @xmath109 for all real @xmath90 , where @xmath110 is the clausen function of order @xmath111 . by substituting @xmath112 and @xmath113 in theorem [ teo : geral ]
, we readily find @xmath114 \ !
, \label{eq : cl2m}\end{aligned}\ ] ] which holds for any non - null real @xmath115 with @xmath116 .
advantageously , this form remains valid for all _ positive integer _
values of @xmath117 , as long as the sum at the right - hand side is taken as null when @xmath118 , as usual . since @xmath119 and @xmath120 , eq .
allows for prompt proofs of theorems [ teo : zetaser1 ] and [ teo : zetaser2 ] , respectively , showing that they are special cases of theorem [ teo : geral ] , as expected .
this is indeed the case for a number of rapidly convergent zeta series in literature , as e.g. some of the zeta series given in refs .
@xcite and many zeta series presented in chaps . 3 and 4 of ref .
this reflects the generality of our theorem [ teo : geral ] .
equation can thus be viewed as a source of rapidly converging zeta series for odd zeta values and clausen functions of even order .
for instance , on taking @xmath121 in eq . , one finds @xmath122 for @xmath123 , this reduces to @xmath124 whereas for @xmath125 one has @xmath126 similarly , for @xmath127 one has @xmath128 for @xmath129 and @xmath130 for @xmath131 .
i could not find these formulae explicitly in literature .
all zeta series investigated here belong to the class embraced by theorem [ teo : geral ] .
their rate of convergence can be analyzed as follows . for convenience ,
denote by @xmath132 the summand of the zeta series at the right - hand side of the formula in our theorem [ teo : geral ] . by applying stirling s formula , namely @xmath133 , and noting that @xmath134 as @xmath135 , one finds @xmath136 where we have made use of @xmath137 and the binomial approximation @xmath138 .
therefore , @xmath139 which is valid for any real @xmath44 with @xmath79 .
this allows for a direct comparison of rates of convergence for the zeta series investigated here in this paper .
for instance , the zeta series in wilton s formula has @xmath140 , hence @xmath141 , whereas the zeta series in our theorem [ teo : zetaser2 ] has @xmath142 , hence @xmath143 .
this is why the latter converges somewhat faster than the former . | in a recent work on euler - type formulae for even dirichlet beta values , i.e. @xmath0 , i have derived an exact closed - form expression for a class of zeta series . from this result , i have conjectured closed - form summations for two families of zeta series . here in this work ,
i begin by using a known formula by wilton to prove those conjectures .
as example of applications , some special cases are explored , yielding rapidly converging series representations for the apry constant , @xmath1 , and the catalan constant , @xmath2 .
interestingly , our series for @xmath3 converges faster than that used by apry in his irrationality proof ( 1978 ) .
also , our series for @xmath4 converges faster than a celebrated one discovered by ramanujan ( 1915 ) .
at last , i present a simpler , more direct proof for a recent theorem by katsurada which generalizes the above results . |
the common age , metallicity and distance of their members make galactic open star clusters favourable environments in which to examine fundamental issues in stellar and galactic astrophysics e.g. the shape of the initial mass function ( imf ) or the form of the initial mass - final mass relationship ( e.g. weidemann 1987 )
. the modestly rich and well studied praesepe ( ngc2632 ) cluster at a distance of 177pc , as determined from hipparcos astrometric measurements ( mermilliod et al .
1997 ) , appears particularly suited to such investigations .
its members share a distinct proper motion so it is comparatively straightforward to discriminate them from the vast majority of field objects along this line of sight .
for example , hambly et al .
( 1995 ) performed an astrometric survey of 19 sq .
degrees centered on the cluster and found the proper motions of members tightly clumped around @xmath14 mas yr@xmath15 and @xmath16mas yr@xmath15 .
a more recent hipparcos based study of praesepe finds mean values of @xmath17 mas yr@xmath15 and @xmath18mas yr@xmath15 ( van leeuwen 1999 ) .
[ cols="<,<,^,^,^,^,^,^,^ , > " , ] a steep powerlaw ( @xmath19 ) was the only shape of imf of the four investigated by w04 consistent with his assumption of seven white dwarf members of praesepe , for reasonable values of the maximum progenitor mass ( 6m@xmath20 m@xmath21m@xmath5 ) .
the `` naylor '' imf , a broken powerlaw with index @xmath22 ( m@xmath23m@xmath5 ) and @xmath24 ( m@xmath25m@xmath5 ) was found to be consistent only for m@xmath26m@xmath5 .
however , in a subsequent paper , williams , bolte & liebert ( 2004 ) , show lb6037 and lb6072 to be qsos , reducing the number of bona fide white dwarf members of praesepe unearthed to date to only five . in this case , only the steep powerlaw form with m@xmath27m@xmath5 can be considered consistent with the observations . nevertheless
, we have presented evidence here the number of praesepe white dwarfs is at least seven . as a further check of the membership status of our two new spectroscopic candidates
we constrain their distances using the measured t@xmath28s and log g s and radii derived from evolutionary tracks . from the `` thick h layer '' models of wood ( 1995 ) we determine r@xmath29r@xmath5 and r@xmath30r @xmath5 for lb5959 and wd0837 + 218 respectively .
subsequently , we estimate the absolute visual magnitude of lb5959 to be m@xmath31 and that of wd0837 + 218 to be m@xmath32 . refering to the synthetic photometry of bergeron et al .
( 1995 ; private comm . ) , we estimate b - v @xmath33 and + 0.05 for the cooler and hotter degenerate respectively . thus from the supercosmos data , b@xmath34 and b@xmath35 , we determine v magnitudes of @xmath36 and @xmath37 .
hence , we estimate these white dwarfs to reside at @xmath38pc and @xmath39pc respectively .
indeed , the favourable success rate of our preliminary survey in unearthing cluster white dwarfs suggests several of the remaining four new candidates will also prove to be degenerate members .
however , as confirmation of this must await further spectroscopy , for now we consider the observed number of praesepe white dwarfs , n , to lie in the range @xmath40 n @xmath41 . based on the williams simulations we find the steep powerlaw form of the imf is consistent with the observed number for any reasonable value of m@xmath11 . similarly , if n@xmath42 the `` naylor '' form can be considered consistent for any reasonable value of m@xmath11 .
we note , one requires to detect at least ten white dwarf members for the salpeter form of the imf to be regarded at best , in williams scheme , as mildly inconsistent with the observed number ( p@xmath43 ) .
as our preliminary results indicate white dwarf members are to be found beyond the well studied inner regions of the cluster ( figure 1 shows wd0837 + 218 lies at a projected separation of @xmath44 ) more likely await discovery .
a detailed survey extending out to at least the tidal radius should be undertaken before any firm conclusions are drawn regarding the form of the imf .
prior to this work only the `` naylor '' form of the imf was found to be consistent with the non - detection of cluster white dwarfs residing in unresolved binary systems ( w04 ) . however , there is no evidence to suggest that either of the two new confirmed white dwarfs resides in a binary .
we find a white dwarf with t@xmath45k and log g @xmath46 , typical of the praesepe population , has m@xmath47=11.0 , m@xmath48=11.0 and m@xmath49=10.2 ( bergeron et al . 1995 ) and a young disc m3 dwarf has m@xmath47=10.7 , m@xmath48=12.3 and m@xmath49=13.4 ( leggett et al .
an unresolved binary consisting of these two objects has u - b@xmath50 , b - v@xmath51 .
it thus lies on the fringes of the w04 criteria for being `` observable '' ( u - b @xmath52 , b - v @xmath53 ) , which were chosen to approximate the limits of the ubv surveys of c01 and anthony - twarog .
white dwarfs with more massive main sequence companions are unlikely to have been identified by these surveys .
similarly , since both luyten s and the current work utilised blue and red photographic plates , selecting objects with colours i.c.@xmath54 ( @xmath55 b - v@xmath56 ) and o - e@xmath57 respectively , these too are likely biased against finding white dwarfs residing in binaries with stars of spectral type earlier than mid - m .
farihi et al ( 2003 ) report a deficit of objects of spectral type later than mid - m paired to field white dwarfs at separations of @xmath58 few 100au .
further , radial velocity surveys point towards a drop in the relative frequency of main sequence binaries with mass ratios m@xmath59m@xmath60 , at separations @xmath61 au ( halbwachs et al . 2003 ; marcy & butler 2000 ) .
hence , it seems plausible , particularly as the progenitors of the praesepe and hyades white dwarfs had m@xmath62m@xmath5 , that by including a population of zero age binaries consisting of randomly paired stars , the simulations of w04 overpredict the number of detectable white dwarfs in unresolved binaries with main sequence companions .
this assumption was acknowledged by w04 as a possible shortcoming in their modelling .
our forthcoming galex survey of praesepe will expand the parameter space in which we are able to search for white dwarfs to include those in unresolved systems with k , g , and f type companions .
it will reveal important additional information on the form of the imf and the binary fraction of this cluster .
we have estimated the masses of the two new white dwarfs by comparing their measured t@xmath28s and log g s to the predictions of evolutionary calculations ( wood 1995 ) . as expected for their progenitor masses , m @xmath62 m , these two objects have masses greater than the canonical white dwarf value of m @xmath63m@xmath5 ( see table 2 ) .
we have also used the evolutionary tracks to estimate the cooling times of these objects , determining @xmath8myrs and @xmath6myrs for lb5959 and wd0837 + 218 respectively .
as the age of the cluster is rather uncertain we have used three different estimates encompassing the likely value ( log @xmath64 ) to derive the progenitor lifetimes . using cubic splines to interpolate between the lifetimes calculated for stars of solar composition by girardi et al .
( 2000 ) , we constrain the masses of the progenitors of lb5959 and wd0837 + 218 to be @xmath9 m @xmath10 m@xmath11 m@xmath5 and @xmath12 m @xmath13 m@xmath5 respectively . examining the location of these objects in figure 11 of c01 we find while lb5959 fits in comfortably with a monotonic relationship between initial mass and final - mass , like lb5893 , wd0837 + 218 appears to be too hot and hence too young for its high mass .
we note there are five blue straggler members of praesepe ( e.g. andrievsky 1998 ) .
the evolution of these objects appears to have been delayed , either through binary interaction or another as yet unidentified mechanism .
we suggest that lb5893 and wd0837 + 218 may be related to this population .
alternatively , as suggested by reid ( 1996 ) , perhaps , at least some stars , do not subscribe to a simple monotonic relationship between their mass and the mass of their resulting white dwarf remnant .
pdd and dpi are sponsored by pparc postdoctoral grants .
rn and mbu acknowledge the support of pparc advanced fellowships . the wht is operated on the island of la palma by the isaac newton group in the spanish observatorio del roque de los muchachos of the instituto de astrofisica de canarias .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
we thank the anonymous referee for timely and useful comments .
99 adams j.d . ,
stauffer j.r . ,
skrutskie m.f . ,
monet d.g . , portegies zwart s.f .
, janes k.a . ,
beichman c.a . , 2002 , aj , 124 , 1570 allard n.f . , royer a. , kielkopf j.f . ,
feautrier n. , 1999 , phrva , 60 , 1021 allen c.w . , 1973 , astrophysical quantities . the athlone press , london andievsky s.m . , 1998 ,
a&a , 334 , 139 anthony - twarog b.j . , 1982 , apj , 255 , 245 anthony - twarog b.j . , 1984 ,
aj , 89 , 267 bergeron p. , wesemael f. , beauchamp a. , 1995 , pasp , 107 , 1047 boesgaard a. m. , budge k. g. , 1988 , apj , 332 , 410 claver c.f .
, liebert j. , bergeron p. , koester d. , 2001 , apj , 563 , 987 eggen o.j . , 1960 , mnras , 120 , 540 eggen o.j .
, greenstein j.l . , 1965 ,
apj , 141 , 83 farihi j. , becklin e.e . , zuckerman b. , 2003 , in brown dwarfs , eds .
martin , e.l .
, iau 211 , asp fellhauer m. , lin d.n.c . , bolte m. , aarseth s.j .
, williams k.a .
, 2003 , apj , 595 , 53 friel e.d .
, boesgaard a.m. , 1992 , apj , 387 , 170 girardi l. , bressan a. , bertelli g. , chiosi c. , 2000 , a&as 141 , 371 .
halbwachs j.l . , mayor m. , udry s. , arenou f. , 2003 , a&a , 397 , 159 hambly n.c . ,
steele i.a .
, hawkins m.r.s . ,
jameson r.f . , 1995 , a&as , 109 , 29 hambly n.c . , irwin m.j . ,
macgillivray h.t . , 2001 ,
mnras , 326 , 1295 holberg j.b . , oswalt t.d . , sion e.m . , 2002 ,
apj , 571 , 512 hubeny i. , 1988 , comp .
comm . , 52 , 103 hubeny i. , hummer d. , lanz t. , 1994 , a&a , 282 , 151 hubeny i. , lanz t. , 1995 , apj , 439 , 875 hummer d. , mihalas d. , 1988 , apj , 331 , 794 lemke m. , 1997 , a&as , 122 , 285 luyten w.j . , 1962 , the observatory , univ .
minnesota , 31 , 1 marcy m. , butler b. , 2000 , pasp , 112 , 137 mathieu r.d .
, mazeh t. , 1988 , apj , 326 , 256 mazzei p. , pigatto l. , 1988 , a&a , 193 , 148 mermilliod j.c . , turon c. , robichon n. , arenou f. , lebreton y , 1997 , in hipparcos venice 97 , esa sp-402 , p643 monet d.g . , et al . , 2003 ,
aj , 125 , 984 oke j.b , 1974 , apjs , 27 , 21 oke j.b , 1990 , aj , 99 , 1621 reid i.n . , 1996 , aj , 111 , 2000 shafer r.a . , et al . , 1991 , esa tm-09 skrutskie , m. f. , et al . 1997 , in the impact of large scale near - ir sky surveys , ed .
f. garzon et al .
( dordrecht : kluwer ) , 25 tinsley b.m .
, 1974 , pasp , 86 , 554 van leeuwen f. , 1999 , a&a , 341 , 71 weidemann v. , 1987 , a&a , 188 , 74 williams k. , 2004 , apj , 601 , 1067 , w04 williams k. , bolte m. , liebert j. , 2004 , astro - ph/0406671 wood m.a . , 1995 , in white dwarfs , eds .
d. koester & k werner , springer , berlin , p41 | we report the discovery , from our preliminary survey of the praesepe open cluster , of two new spectroscopically confirmed white dwarf candidate members . we derive the effective temperatures and surface gravities of wd0837 + 218 and wd0837 + 185 ( lb5959 ) to be @xmath0k and log g @xmath1 and @xmath2k and log g @xmath3 respectively . using theoretical evolutionary tracks we estimate the masses and cooling ages of these white dwarfs to be @xmath4m@xmath5 and @xmath6myrs and @xmath7m@xmath5 and @xmath8myrs respectively . adopting reasonable values for the cluster age we infer the progenitors of wd0837 + 218 and wd0837 + 185 had masses of @xmath9 m @xmath10 m@xmath11m@xmath5 and @xmath12 m @xmath13 m@xmath5 respectively , where m@xmath11 is the maximum mass of a white dwarf progenitor .
we briefly discuss these findings in the context of the observed deficit of white dwarfs in open clusters and the initial mass - final - mass relationship .
[ firstpage ] stars : white dwarfs ; galaxy : open clusters and associations : praesepe |
interstellar dust is still a dominant astrophysical unknown in cosmological distance estimates @xcite and models of how starlight is re - processed within a galaxy ( e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) because some 10 - 30% of all the starlight is re - emitted by the dust in the far - infrared @xcite .
interstellar dust can be found in two ways ; by its emission or through the extinction of stellar light .
characterization of emission has made great strides with the _ spitzer _ and _ herschel space observatories _
( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a library of far - infrared and sub - mm images of nearby galaxies is currently being collated and more insight into the physics and distribution of interstellar dust in nearby galaxies can be expected with the great improvements in spectral coverage , sensitivity and spatial resolution .
extinction measures of dust have some specific advantages over emission : they do not depend on the dust temperature , allowing for the detection of much colder dusty structures , and typically have the high resolution of the optical imaging observations .
the single drawback is that one needs a known background light source . in the case of the transparency of spiral galaxies , two techniques
have just such a proven background source : background galaxies counts and occulting galaxy pairs .
the technique that uses the number of background galaxies @xcite is nearing obsolescence as its inherent resolution and accuracy , limited by the intrinsic cosmic variance of background sources , are now surpassed by the accuracy and sensitivity of _ herschel space observatory _ observations of dust surface density in nearby galaxies .
the occulting galaxies technique , however , has increased steadily in accuracy and usefulness , owing in a large part to the increasing sample sizes . estimating dust extinction and mass from differential photometry in occulting pairs of galaxies was first proposed by @xcite .
their technique was then applied to all known pairs using ground - based optical images @xcite and spectroscopy @xcite .
subsequently , some pairs were imaged with the _ hubble space telescope _ ( hst * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these initial results , however , were limited by sample sizes ( @xmath0 pairs ) .
more recently , new pairs were found in the sdss spectroscopic catalogue ( 86 pairs in * ? ? ?
* ) and through the galaxyzoo project @xcite : 1993 pairs reported in @xcite .
this wealth of new pairs provided opportunities for follow - up with ifu observations @xcite and galex @xcite .
a greatly expanded occulting galaxy catalog improves accuracy as ideal pairs " an elliptical partially occulted by a late - type galaxy can be selected for follow - up .
ellipticals are the optimal background source as their light profile is smooth and very symmetric .
results from the occulting galaxy pairs include : ( 1 ) a mean extinction profile @xcite , ( 2 ) an indication that the dust may be fractal @xcite and ( 3 ) the observation that the colour - extinction relation is grey , i.e. there is little or no relation between the reddening of the stellar populations and total extinction .
the latter is due to the coarse physical sampling of ground - based observations .
the galactic extinction law returns as soon as the physical sampling of the overlap region resolves the molecular clouds in the foreground disk ( @xmath1 pc * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . a very reliable way to identify occulting galaxy pairs , i.e. , purely serendipitous overlaps of galaxies is through blended spectra . in @xcite
, we used the rejects from the strong lenses with acs survey ( slacs * ? ? ?
* ; * ? ? ?
* ) , a highly successful search for strong lenses , confirmed with hst @xcite , with spectroscopic selection extended now to the boss survey @xcite .
both types of blended spectral sources have two things in common : very close association on the sky ( within a sdss spectroscopic fibre of 3 diameter ) and clear spectroscopic signal from both galaxies at distinct redshifts . in this paper
, we present the blended spectra catalogue based on the galaxy and mass assembly ( gama ) survey @xcite as candidates for either strong lensing follow - up or occulting galaxy analysis ( e.g. , hst imaging or spectroscopy ) .
the gama data is an improvement over sdss as the target galaxies can be fainter , the aperture is smaller ( i.e. , a closer overlap of the galaxies ) and the autoz detection algorithm is a marked improvement on the sdss detections .
the paper is organized as follows :
[ s : gama ] briefly introduces the gama survey , [ s : selection ] the redshift determination and selection of blended spectra , [ s : vis ] presents the visual classifications of the blended objects , [ s : cat ] presents the blended spectra catalogue and we discuss the pair classification and their possible future uses in
[ s : disc ] .
the gama survey has obtained over 250000 galaxy redshifts selected to @xmath2mag over 290 deg@xmath3 of sky @xcite . at the heart of this survey
is the redshift survey with the upgraded 2df spectrograph aaomega @xcite on the anglo - australian telescope .
the gama survey extends over three equatorial survey regions of 60 deg@xmath3 each ( called g09 , g12 and g15 ) and two southern regions of similar area ( g02 , g23 ) .
see @xcite for a detailed description of the gama input catalogueue for the equatorial regions .
the redshift survey in combination with a wealth of imaging data has led to many science results already .
we use for this work the gama ii redshifts @xcite , which were obtained using a robust cross - correlation method for spectra with and without strong emission lines @xcite .
galaxy redshifts were initially determined by a supervised fit @xcite but a recent upgrade to the gama survey pipeline includes a fully - automated template - based redshift determination ( autoz , * ? ? ? * ) . in certain cases ,
the fits for different templates resulted in two high - fidelity , but different redshifts ; these are the candidate blended objects of interest to us here .
the autoz code obtains cross - correlation redshifts against stellar and galaxy templates with varying strength of emission and absorption line features .
the height and position of the first four peaks of normalized cross - correlation functions are obtained .
these are called @xmath4 , @xmath5 , @xmath6 and @xmath7 each with a corresponding redshift and template number , with the peaks separated by at least 600km / s .
high values of @xmath4 and @xmath5 , particularly relative to @xmath6 and @xmath7 , can then be used to select candidate blended spectra .
the autoz algorithm marks a significant improvement in the identification of blended spectra over that which could be obtained from gama i redshifts . in the initial redshift campaign ,
`` redshifters '' the gama team members identifying the redshift with runz were focused on attaining a reliable redshift for single objects .
in such an approach , only those spectra with wildly different redshifts by two redshifters or spectra remarked upon during visual inspection would be selected . with autoz , blended spectra
are identified as different redshifts using normalized cross - correlation functions , a much more objective and complete approach .
double redshift selection using the autoz approach was to require that two different redshifts had high cross - correlation peaks ( @xmath4 and @xmath5 ) , while the next two redshifts had significantly lower peak values . in order to address this
, we defined the ratio between the second redshift peak value and subsequent , third and fourth , redshift peaks to be : @xmath8 to select the double - z candidates , we required @xmath9 ( to avoid aliasing and a clean selection of real blends , see figures [ f : double - z ] and [ f : z - z - plot ] ) and for the first two redshifts to be from galaxy spectral templates .
the galaxy spectral templates used in autoz were from sdss .
an early version of the code used the sdss dr2 templates , while a later version used templates derived from the @xcite galaxy eigenspectra .
these later templates were numbered 40 - 47 in order of increasing emission - line strength . to broadly classify the templates , we select templates 40 - 42 as ` passive galaxies ' ( pg ) and 43 - 47 as ` emission - line galaxies ' ( elg ) . initially , we selected candidates with @xmath10 and @xmath11 , with these values determined using the early version of the code .
after the code was updated , the value of @xmath12 changed as a result of the new templates and a modest increase in the allowed redshift range from 0.8 to 0.9 . using the new values , we selected candidates with @xmath10 and with no restriction on @xmath5 .
old candidates were retained subject to a couple of criteria : the new value of @xmath12 was still greater than 1.35 , and @xmath13 was not near a problematic cross - correlation alias .
aliases can result from the matching of different emission lines in the templates to a strong line in the data .
all candidates , old and new , near the alias of @xmath14 ( @xmath15 ) or the inverse were removed from the sample .
figure [ f : double - z ] shows the distribution of candidates in @xmath5 versus @xmath12 , i.e. the selection parameters .
figure [ f : z - z - plot ] shows where the aliases lie , in @xmath16 versus @xmath17 , with respect to the candidates . versus @xmath17 : for all gama spectra with @xmath19 and @xmath18 shown with black points , and for the selected candidates ( @xmath9 ) shown with red squares .
most candidates lie away from the alias lines that are appear parallel when plotting the logarithm of one plus redshift on each axis.,scaledwidth=45.0% ] the selection resulted in galaxy pair candidates ( from 299 blended spectra some source locations were observed more than once ) . depending on which template matched best for both redshifts in the blend , we classified the type of blends as follows : two passive - template galaxies ( pg+pg ) , two emission - line templates ( elg+elg ) , a passive template at low redshift and emission line template at higher redshift ( pg+elg ) , or vice versa ( elg+pg ) .
table [ t : spectype ] summarizes the resulting classifications .
the autoz results for our blended objects are listed in table [ t : cat ] . | we present the catalogue of blended galaxy spectra from the galaxy and mass assembly ( gama ) survey .
these are cases where light from two galaxies are significantly detected in a single gama fibre .
galaxy pairs identified from their blended spectrum fall into two principal classes : they are either strong lenses , a passive galaxy lensing an emission - line galaxy ; or occulting galaxies , serendipitous overlaps of two galaxies , of any type .
blended spectra can thus be used to reliably identify strong lenses for follow - up observations ( high resolution imaging ) and occulting pairs , especially those that are a late - type partly obscuring an early - type galaxy which are of interest for the study of dust content of spiral and irregular galaxies .
the gama survey setup and its autoz automated redshift determination were used to identify candidate blended galaxy spectra from the cross - correlation peaks .
we identify blended spectra with a minimum velocity separation of 600 km / s , of which are lens pair candidates , emission - line - passive pairs , are pairs of emission - line galaxies and and are pairs of galaxies with passive spectra .
we have visually inspected the candidates in the sloan digital sky survey ( sdss ) and kilo degree survey ( kids ) images .
many blended objects are ellipticals with blue fuzz ( _ ef _ in our classification ) .
these latter ef " classifications are candidates for possible strong lenses , massive ellipticals with an emission - line galaxy in one or more lensed images .
the gama lens and occulting galaxy candidate samples are similar in size to those identified in the entire sdss .
this blended spectrum sample stands as a testament of the power of this highly complete , second - largest spectroscopic survey in existence and offers the possibility to expand e.g. , strong gravitational lens surveys .
[ firstpage ] catalogues galaxies : statistics galaxies : distances and redshifts ( ism : ) dust , extinction gravitational lensing : strong |
this work is funded by the australian research council centre of excellence program ( ce110001027 ) .
jg and mtc are supported by the chinese scholarship council overseas scholarship . | we introduce a scheme for the parallel storage of frequency separated signals in an optical memory and demonstrate that this dual - rail storage is a suitable memory for high fidelity frequency qubits .
the two signals are stored simultaneously in the zeeman - split raman absorption lines of a cold atom ensemble using gradient echo memory techniques .
analysis of the split - zeeman storage shows that the memory can be configured to preserve the relative amplitude and phase of the frequency separated signals . in an experimental demonstration dual - frequency pulses
are recalled with @xmath0% efficiency , @xmath1% interference fringe visibility , and @xmath2 phase stability .
the fidelity of the frequency - qubit memory is limited by frequency - dependent polarisation rotation and ambient magnetic field fluctuations , our analysis describes how these can be addressed in an alternative configuration .
99 a. i. lvovsky , b. c. sanders , and w. tittel , ( 12 ) , 706714 ( 2009 ) .
m. hosseini , b. m. sparkes , g. t. campbell , p. k. lam , and b. c. buchler , ( 2 ) , 174 ( 2011 ) . m. p. hedges , j. j. longdell , y. li , and m. j. sellars , ( 7301 ) , 10526 ( 2010 ) .
b. m. sparkes , j. bernu , m. hosseini , j. geng , q. glorieux , p. a. altin , p. k. lam , n. p. robins , and b. c. buchler , ( 8) , 085027 ( 2013 ) .
j. j. longdell , e. fraval , m. j. sellars , and n. b. manson , ( 6 ) , 063601 ( 2005 ) .
y. o. dudin , l. li , and a. kuzmich , ( 3 ) , 031801 ( 2013 ) g. heinze , c. hubrich , and t. halfmann , ( 3 ) , 033601 ( 2013 ) . e. saglamyurek , n. sinclair , j. jin , j. a. slater , d.l oblak , f. bussires , m. george , r. ricken , w. sohler , and w. tittel , ( 7331 ) , 5125 ( 2011 ) .
m. afzelius , i. usmani , a. amari , b. lauritzen , a. walther , c. simon , n. sangouard , j. min , h. de riedmatten , n. gisin , and s. krll , ( 4 ) , 40503 ( 2010 ) .
m. hosseini , b. m sparkes , g. htet , j. j. longdell , p. k. lam , and b. c. buchler , ( 7261 ) , 2415 ( 2009 ) .
y - w cho and y - h kim . ( 25 ) , 2578693 ( 2010 ) .
s. riedl , m. lettner , c. vo , s. baur , g. rempe , and s. drr , ( 2 ) , 22318 ( 2012 ) . c. kupchak , t. mittiga , b. jordaan , m. namazi , c. nlleke , and e. figueroa , , 7658 ( 2015 ) . c. clausen , f. bussires , m. afzelius , and n. gisin , ( 19 ) , 190503 ( 2012 ) .
d. g. england , p. s. michelberger , t. f. m. champion , k. f. reim , k. c. lee , m. r. sprague , x - m jin , n. k. langford , w. s. kolthammer , j. nunn , and i. a. walmsley , ( 12 ) , 124008 ( 2012 ) .
a. nicolas , l. veissier , l. giner , e. giacobino , d. maxein , and j. laurat , ( 3 ) , 234238 ( 2014 ) . a. l. alexander , j. j. longdell , m. j. sellars , and n. b. manson , .
( 4 ) , 14 ( 2006 ) .
j. j. longdell , g. htet , p. k. lam , and m. j. sellars , ( 3 ) , 032337 ( 2008 ) .
g. htet , m. hosseini , b. m sparkes , d. oblak , p. k. lam , and b. c. buchler , ( 20 ) , 23235 ( 2008 ) .
j. roslund , r. m. de arajo , s. jiang , c. fabre , and n. treps , ( 2 ) , 109112 ( 2013 ) .
p. c. humphreys , w. s. kolthammer , j. nunn , m. barbieri , a. datta , and i. a. walmsley , , 130502 ( 2014 ) .
a. macrae , g. t. campbell , and a. i. lvovsky , ( 22 ) , 2659 ( 2008 ) .
m. hosseini , g. t. campbell , b. m. sparkes , p. k. lam , and b. c. buchler , ( 10 ) , 794798 ( 2011 ) .
b. m. sparkes , , phd thesis ( 2013 ) .
d. a. steck , ( 2001 ) .
j. geng , g. t. campbell , j. bernu , d. b. higginbottom , b. m. sparkes , s. m. assad , w. p. zhang , n. p. robins , p. k. lam , and b. c. buchler , ( 11 ) , 113053 ( 2014 ) .
optical quantum memories for the coherent and reversible transfer of quantum information between light and matter are an essential component of quantum communication and quantum computation proposals @xcite . despite substantial challenges , quantum memories are now demonstrating efficiencies approaching 90@xmath3 @xcite , storage times of many seconds @xcite , bandwidths of hundreds of mhz @xcite and large multi - modal capacities @xcite .
the performance of a practical quantum memory is contingent on the qubit encoding that is used .
any linear , single - mode memory is capable of storing photon number qubits , however inefficiency and loss in the memory or other network elements will decrease output state fidelity .
for this reason an alternate encoding is often used , such as the polarisation , orbital angular momentum , path , or arrival time ( time - bin ) of a single photon .
memories for polarisation qubits have been demonstrated using electromagnetically induced transparency ( eit ) @xcite , atomic frequency comb ( afc ) @xcite memories , and raman absorption @xcite .
eit has also been used to store orbital angular momentum qubits with high fidelity @xcite .
temporally multimode memories , such as afc @xcite and gradient echo memories ( gem ) @xcite , are suitable for time - bin qubits . here
we present work based on the three - level gem protocol @xcite towards a quantum memory for frequency qubits @xcite .
we extend the gem protocol to make use of zeeman sub - levels to store and recall two frequency channels , this provides a basis to store frequency qubits with a high fidelity .
these split zeeman levels have previously been used to store matched pulses with eit @xcite . in the three - level gem scheme a strong optical coupling field and a weak signal field form an off - resonant raman transition in an ensemble of @xmath4-type three - level atoms . by applying a linear atomic frequency gradient along the optical axis of the ensemble
a signal pulse can be transferred into a coherent atomic spin wave excitation between the two ground states .
signals absorbed by the broadened ensemble can be re - emitted on demand by reversing the frequency gradient such that the atomic coherence rephases . because gradient - echo memories can limit losses by operating with optical fields far detuned from the excited state , storage efficiencies can be made significantly higher than similar eit memories .
gems have been operated with 87@xmath3 efficiency in warm atomic vapour @xcite and 80@xmath3 efficiency in a cold atomic ensemble with longer storage time @xcite . under constant uniform bias field @xmath5 and reversible gradient field
@xmath6 illustrating the operation of the dual - rail gradient echo memory . ] in our system the atomic frequencies are manipulated by applying a tailored magnetic field along the optical propagation axis . a bias magnetic field @xmath5 lifts the degeneracy of the zeeman sub - levels by introducing a zeeman shift given by @xmath7 , where @xmath8 is the zeeman level number , @xmath9 is the land factor for hyperfine state @xmath10 , @xmath11 is the bohr magneton and @xmath12 is the planck constant . for the d1 line of @xmath13rb atoms the land factors are respectively @xmath14 and @xmath15 .
a single coupling field detuned from the @xmath16 transition will produce three raman absorption lines for a weak signal field on the @xmath17 transition , corresponding to signal field detunings of 0 ( insensitive line ) and @xmath18 relative to the unsplit raman transition , where @xmath19 is given by @xmath20 a reversible magnetic field gradient @xmath21 is applied along the ensemble in addition to the uniform bias field @xmath5 to broaden the two sensitive lines around @xmath22 for light storage using the gem protocol , as illustrated in fig [ fig.simple_level ] .
previous gem experiments used only one of these two magnetically sensitive lines to store and recall pulses . in this paper we show that the sub - levels responsible for the two absorption lines can be utilised to store pulses of light at frequencies @xmath22 in parallel such that the two lines form a memory for frequency encoded qubits .
the same scheme readily extends to higher mode frequency states in atomic ensembles with high ground state angular momentum .
for example , the four magnetically sensitive zeeman levels of the @xmath23rb ground state ( f=2 ) could be used to store four mode frequency states using the same approach that we demonstrate here with @xmath13rb . to achieve high fidelity frequency qubit storage in such a dual - rail gem
it is necessary to balance the effective optical depth of the two raman lines shown in fig .
[ fig.raman_lines](b ) and match the output polarisation of each frequency rail .
the optical depth of each line depends on the dipole matrix elements associated with the transitions for the coupling and signal fields involved in the relevant two - photon transitions , as well as the polarisation of the two beams .
in general , for an ensemble of atoms that has a population distributed uniformly across the zeeman sub - levels , this results in unequal optical depths for the @xmath22 raman transitions . to balance the efficiencies of the two frequency rails we use linear polarisations for the coupling and signal fields which , by symmetry , ensures equal optical depths . in this configuration
, however , frequency dependent polarisation rotation will produce output modes that are not matched .
polarisation rotation occurs as a result of interference between coherences driven by the @xmath24 and @xmath25 components of each signal field . to examine the effect of this rotation
, we consider below the level structure of the rb@xmath13 atoms , shown in fig .
[ fig.raman_lines](a ) .
off - resonant dispersion of @xmath24 and @xmath25 will further rotate the signal fields , but the relative faraday rotation between signals 1 and 2 is negligible so long as their splitting , @xmath19 , is much smaller than the detuning , @xmath26 .
we assume that the population initially rests entirely in the @xmath27 ground state , evenly distributed between zeeman sub - levels , and that there are no coherences within that manifold .
these assumptions are valid provided that the probe field is weak and that the splitting , @xmath19 , is large compared to the raman linewidth .
we proceed by treating each of the signals independently , starting with the signal 1 . referring to fig .
[ fig.raman_lines](a ) , signal 1 will produce a coherence , @xmath28 , between the @xmath29 and @xmath30 states via three @xmath4 systems with excited states @xmath31 , @xmath32 and @xmath33 . this last @xmath4 system reduces the absorption asymmetry between @xmath24 and @xmath25 components of the signal . consequently the @xmath24/@xmath25 absorption ratio approaches unity as the detuning @xmath26 is increased .
the @xmath4 system formed by @xmath34 , @xmath35 and @xmath36 makes a small contribution to the memory process due to weak effective coupling strength .
we neglect this contribution in the analysis below because the optical depth associated with this process is @xmath37 times smaller than the optical depth of the three @xmath4 schemes contributing to the @xmath38 coherence .
the equations of motion for signal 1 are @xmath39 with effective coupling strengths @xmath40 where @xmath41 is the dipole matrix element corresponding to the @xmath42 transition , @xmath43 are the rabi frequencies associated with the right and left circularly polarised components of the control field , @xmath44 is the control field detuning from the @xmath45 transition and @xmath46 mhz @xcite is the control field detuning from the @xmath47 transition .
we can further simplify this system by introducing the signal 1 coupled mode @xmath48 @xmath49 from eqs .
( [ maxwell_bloch:1 ] ) and ( [ maxwell_bloch:2 ] ) we derive eqs . of motion for the coupled mode @xmath50 the component of the input perpendicular to the coupled mode polarisation
is not stored in the memory . given a horizontally polarised signal and vertically polarised control , @xmath51 and @xmath52 determine the signal 1 coupled mode polarisation @xmath53 .
analogous eqs .
exist for signal 2 ; by symmetry ( assuming detunings @xmath54 and @xmath55 are approximately the same for signals 1 and 2 , and given the same beam polarisations ) the coupled mode polarisation for signal 2 is @xmath56 .
the input and output polarisations of both signals are indicated schematically in fig [ fig.setup ] .
the portion of the input orthogonal to the coupled mode is not stored in the memory , therefore the highest efficiency possible in this memory configuration is @xmath57 .
the interference fringe visibility between the two output modes is limited to the coupled mode overlap @xmath58 . to eliminate the polarisation rotation described above and improve fidelity
the memory can be configured with circularly polarised signal and control beams . in this alternative configuration
the memory fidelity is limited by recall efficiency asymmetry between the rails , which goes to zero in the high optical depth limit . with low optical depth one
could still balance these efficiencies directly by preparing an ensemble population unevenly distributed between the zeeman sub - levels of the ground state .
the experiment is carried out in a cold atomic cloud of @xmath59rb produced by a magneto - optical trap .
the ensemble is initially prepared in the @xmath27 hyperfine ground state with a resonant optical depth of @xmath60 on the unsplit d@xmath61 @xmath17 transition .
details of the trapping and preparation of the ensemble can be found in @xcite .
because memory experiments require precise control of the magnetic field that is applied to the ensemble , the magnetic fields used for trapping must be switched off prior to performing the experiment .
eddy currents induced in the optical bench are allowed to die away for 1 ms after the trapping fields are removed before the memory experiment begins .
the coupling field is turned on 400 @xmath62s before the writing stage to pump the atoms to the @xmath27 hyperfine ground state and remains on during the writing and reading stages .
the experimental setup is shown schematically in fig [ fig.setup ] .
the coupling and signal fields are derived from a ti : sapphire laser that is 200 mhz blue - detuned from the @xmath13rb d@xmath61 @xmath16 transition .
the signal fields are produced by a fibre - coupled phase - modulator followed by a locked cavity .
the modulator generates sidebands at @xmath63 ghz relative to the coupling field and the cavity is used to select only the correct sideband , which is then split into the two signal fields .
all three fields pass through separate acousto - optic modulators to allow for fine frequency adjustment as well as gating and pulsing .
the signal fields are combined with the same linear polarisation ( horizontal ) , passed through an optical fibre and focused to match the size of the atomic cloud ( beam waist = @xmath64 @xmath62 m ) .
the coupling field is polarised orthogonally to the signal fields and is collimated to a larger diameter of 7 mm so that it illuminates the entire ensemble uniformly .
the coupling field propagates with a small angle relative to the signal beams and after passing through the cell it is extinguished by @xmath65 45 db using a razor blade and a pinhole .
this spatial filtering incidentally reduces the signal field efficiency by 1 db .
the results of the dual - rail storage and recall experiment are shown in fig .
[ fig.echo ] .
panels ( a ) and ( b ) show independent storage and recall of each of the signal fields in the absence of the other . panel ( c ) shows storage of the two signal fields simultaneously .
the input signal pulses have a gaussian profile with 10 @xmath62s width .
we record reference traces of the inputs ( black ) in the absence of the atomic ensemble .
traces showing storage and recall in the atom ensemble ( red , blue , purple ) are produced by reversing the magnetic field gradient across the mot as per the gem scheme described above such that the pulse is stored for approximately one pulse width .
the polarisation of the control field is tuned to balance the efficiencies of the echoes to compensate for misalignment of the magnetic gradient relative to the optical axis and polarisation errors caused by the glass vacuum cell walls .
the recall efficiencies are @xmath66 and @xmath67 for signal fields 1 and 2 respectively .
we demonstrate dual - rail storage by sending both signal pulses into the memory simultaneously with matched temporal profiles and amplitudes .
the combined signals are recalled with 35@xmath3 efficiency and the interference fringe visibility between the recalled pulses is @xmath68 , compared to an expected visibility of @xmath69 given the output mode overlap derived earlier ( @xmath70 ) , the measured rail efficiencies ( @xmath71,@xmath72 ) and the input temporal matching ( @xmath73 ) .
the relative phase between the echo signals is measured by comparing the beat signal phase of the recalled pulses to earlier reference pulses that are not stored and not shown in fig .
[ fig.echo ] .
the storage process produces an overall shift in the phase difference between the two signals with a standard deviation of 15@xmath74 , fig .
[ fig.echo](d ) .
this variation is due to magnetic field background noise . by triggering the experimental sequence from the mains power supply , which improves the magnetic field stability ,
this phase noise was later reduced to 6@xmath74 , fig .
[ fig.echo](d ) .
the output visibility in this second measurement was reduced by unrelated experimental considerations including exhaustion of our rb source .
these results demonstrate that the dual - rail gradient echo memory can be used to store and recall two frequency separated signals in parallel while preserving the relative phase and amplitude of the optical states .
this technique is therefore a suitable memory for frequency qubits with fidelity limited by the frequency - dependent polarisation rotation of two signals and the magnetic field stability of the atom ensemble .
the alternative circular polarisation scheme described above is free of polarisation rotation .
instead the fidelity in this configuration is limited by unbalanced absorption of the two signals , which goes to zero as the optical depth increases and could be addressed even in a low optical depth ensemble by preparing asymmetric initial populations .
achievable improvements to the atom ensemble optical depth and ambient field isolation promise to improve both the efficiency and fidelity of such a frequency - qubit memory . |
in constructing models beyond the standard model ( sm ) such as technicolor or supersymmetric theories , it is important to test their viability by looking for possible conflicts with currently available experimental data .
a popular procedure has been the @xmath0 oblique correction analysis of ref .
however , the method can only be applied to theories which satisfy three conditions , namely ( 1 ) the electroweak gauge group is the standard @xmath2 , ( 2 ) vertex and box corrections from new physics are negligible , and ( 3 ) the scale of new physics is large compared to the electroweak scale .
furthermore , the oblique correction analysis requires ( 4 ) the model to be completely specified since any particle with electroweak gauge quantum numbers will show up in a vacuum polarization loop .
many theories of interest do not fall into this category .
for instance , in the topcolor assisted technicolor model of ref .
@xcite , ( 1 ) the electroweak gauge group is @xmath3 resulting in an extra @xmath4 boson , ( 2 ) there are potentially large corrections to the @xmath5 vertex for third generation fermions coming from @xmath4 and coloron exchange , ( 3 ) the scale of new physics may be as low as a tev , and ( 4 ) the technisector of the theory remains unspecified .
the solution to our predicament lies in the observation that a large number of electroweak observables have been measured accurately at lep and sld _ on the @xmath1pole _ , and that the majority of them are just _ ratios of coupling constants _ of the @xmath1 to quarks and leptons .
this means that they are only sensitive to oblique corrections through the effective value of @xmath6 _ at that particular energy scale_. therefore , one can use just one of the observables , @xmath7 say , to _ fix _ the value of @xmath6 , and use it to predict all the rest for the sm . any deviations of the data from those predictions must be due to _ vertex corrections _ from new physics .
this method of comparing sm predictions to @xmath1pole electroweak observables solves all of the problems mentioned above since ( 1 ) mixing of the @xmath1 with extra gauge bosons can be treated as a vertex correction , ( 2 ) vertex corrections need not be small while box corrections are still naturally suppressed on the @xmath1pole , ( 3 ) the new physics scale need not be large since only observables at one scale are used in the analysis , and ( 4 ) the theory need not be specified completely since the method is blind to oblique corrections .
this method has been used successfully to constrain the size of possible vertex corrections to the @xmath8 vertex @xcite , and to constrain the topcolor assisted technicolor model mentioned above @xcite .
we have also applied the technique to constrain : an r parity violating extension to the mssm @xcite , higgs masses in a general two higgs doublet model @xcite , and the fundamental planck scale of a model with large extra dimensions @xcite .
details are provided in the respective references .
this work was supported in part by the u.s .
department of energy , grant de
fg0592er40709 , task a. | we discuss how new physics not encompassed within the @xmath0 oblique correction framework can be constrained from precision electroweak measurements via vertex corrections to @xmath1pole observables . |
although a clear physical picture for three dimensional antiferromagnets has emerged , the understanding of one dimensional ( 1d ) antiferromagnets ( afms ) is less clear.@xcite in these heisenberg chain systems , quantum effects play an important role , but long - range order is not expected.@xcite one of the interesting effects that occurs in 1d afms is a dimerization of the spins , known as the spin - peierls transition,@xcite which leads to a decrease in the magnetic energy without long - range magnetic order as reported in cugeo@xmath5 .
however , the spin - peierls transition can also be caused by three - dimensional phonons ; in this case 3d antiferromagnetic order can occur.@xcite cusb@xmath0o@xmath6 is a green - colored@xcite insulator which orders antiferromagnetically at nel temperature @xmath3= 8.7 k.@xcite it possesses a nearly octahedral local environment of the cu@xmath7 ions in contrast to other cuprates , which have a quasiplanar fourfold cu - o coordination . @xcite it undergoes a phase transition from an ideal trirutile tetragonal structure ( @xmath8-cusb@xmath0o@xmath6 ) above 380 k to monoclinic ( @xmath9-cusb@xmath0o@xmath6 ) below while largely maintaining the quasioctahedral environment.@xcite the distortion of cusb@xmath0o@xmath6 to lower symmetry is in sharp contrast to the related trirutiles , @xmath10sb@xmath0o@xmath6 , which remain tetragonal ( @xmath10 = co and ni ) and order antiferromagnetically at 12.64 k and 2.5 k , respectively .
@xcite the magnetic sublattice of cusb@xmath0o@xmath6 has been studied previously.@xcite two investigations@xcite noted weak magnetic superlattice peaks below @xmath3 .
the first was unable to refine the structure using rietveld analysis .
two possible magnetic structures were proposed and the authors established@xcite that the magnetic moment on the copper site was 0.5@xmath11 .
the second investigation also proposed two magnetic structural models , and ultimately found only one to be consistent with magnetic susceptibility data.@xcite the single - crystal investigation by kato et al .
presents the highest - quality data , and their magnetic structure model appears to deserve the most confidence,@xcite although some magnetic reflections could not be indexed .
the suggested model involves cu moments ordered ferromagnetically along @xmath12 with a magnetic wave vector of ( @xmath13/a,0,@xmath13/c ) .
the 1d model for the magnetic susceptibility of linear chain spin-@xmath14 systems due to bonner and fisher agrees well with the magnetic susceptibility@xcite of cusb@xmath0o@xmath6 .
the interchain - to - intrachain coupling ratio was estimated@xcite at 2@xmath1510@xmath16 .
this is about 10 times smaller than reported@xcite for cugeo@xmath5 .
electronic structure calculations for cusb@xmath0o@xmath6 reveal an unusual quasi-1d magnetic ground state driven by orbital ordering , which is attributed to the presence of competing in- and out - of - plaquette orbitals and strong electronic correlations .
the 1d superexchange interaction ( cu - o - o - cu ) is along the [ 110 ] direction , via the apical oxygen ions , rather than the in - plaquette oxygens .
this unique orbital exchange is believed@xcite to be at the core of the unusual 1d magnetism in cusb@xmath17o@xmath6 . in this report
, measurements of the magnetic susceptibility , heat capacity , and thermal expansion of cusb@xmath0o@xmath6 are reported .
a broad peak at @xmath18 60 k in the magnetic susceptibility indicates the presence of one - dimensional antiferromagnetism and three - dimensional antiferromagnetism appears at @xmath3 = 8.7 k. subtraction of the heat capacity for the non - magnetic analog znsb@xmath0o@xmath6 allows determination of the magnetic entropy , which clearly shows that the one - dimensional antiferromagnetism begins to appear at @xmath18 115 k. both the heat capacity and thermal expansion coefficients reveal distinct jumps at @xmath3 , similar to what is typically observed at a normal - superconductor phase transition .
the heat capacity data reveals an energy gap below @xmath19 of magnitude @xmath4 = 17.48(6 ) k. the unusual character of this phase transition suggests that it is a crossover between one - dimensional antiferromagnetism and three - dimensional antiferromagnetic order that is associated with a spin - peierls transition occurring as a result of three - dimensional phonons coupling with _ jordan - wigner - transformed _ fermions .
the thermal expansion reveals extremely unusual and anisotropic behavior attributed to anharmonic lattice vibrations resulting from the presence of short - range antiferromagnetic order .
the thermal expansion results are compared with published@xcite x - ray data for cusb@xmath0o@xmath6 , demonstrating excellent agreement .
spin - flop transitions for magnetic fields applied along both the @xmath20 and @xmath12 axes are observed .
single crystals used in this work were grown by chemical vapor transport , as described elsewhere .
@xcite two cusb@xmath0o@xmath6 crystals were characterized by magnetic and heat capacity measurements using a quantum design physical property measurement system and they yielded identical results .
the orientation of the single crystals was determined using laue x - ray diffraction .
we observed twinning in the cusb@xmath0o@xmath6 crystals . in most cases ,
the laue diffraction on a sample selected for orientation confirmed the presence of two crystals ( primary and secondary ) with the same @xmath12 axis .
the [ 100 ] and [ 001 ] directions of the primary crystal coincided with the [ 013 ] and [ 101 ] directions of the secondary crystal .
the sample was then polished to eliminate the secondary crystal and therafter could be oriented along the principle crystallographic axes for measurements .
the final laue images of these polished samples revealed no twinning .
a polycrystalline sample of znsb@xmath0o@xmath6 was prepared@xcite for heat capacity measurements by reacting well - mixed , stoichiometric amounts , of sb@xmath17o@xmath5 and zno@xmath17 in air at @xmath21 for 12 h. the powder was reground , pelletized , and sintered at @xmath22 in air for 12 h. x - ray diffraction revealed single - phase nature with lattice parameters @xmath20 = 4.66 and @xmath23 = 9.24 , in agreement with a prior report.@xcite znsb@xmath0o@xmath6 single crystals were grown using chemical vapor transport .
polycrystalline znsb@xmath0o@xmath6 powder was sealed in an evacuated quartz tube and the hot and cold ends of the tube were kept at 800 and 750 @xmath24c , respectively for 100 h. well - faceted crystals were obtained , which exhibited good laue imaged with no twinning .
thermal expansion measurements were performed on a cusb@xmath0o@xmath6 sample with dimensions of about @xmath25 , @xmath26 , and @xmath27 @xmath28 ( @xmath20 , @xmath12 , and @xmath23 axes , respectively ) .
measurements of a second sample confirmed the results .
thermal expansion was also measured on an oriented znsb@xmath0o@xmath6 single crystal with dimensions 1.85 , 1.22 and 0.38 mm ( @xmath20 , @xmath12 , and @xmath23 axes , respectively ) .
the quartz dilatometer@xcite used for the thermal expansion measurements has a sensitivity to changes in length of 0.1 .
measurements along each axis consist of 1700 data points and were repeated 2 - 3 times and averaged .
the data are corrected for the empty - cell effect and for the relative expansion between the cell and the sample .
note that the relative resolution of our dilatometer cell is about 4 orders of magnitude higher than that possible with x - ray or neutron diffraction .
( color online ) ( a ) temperature ( _ t _ ) dependence of magnetization ( @xmath10 ) when the magnetic field is applied along the @xmath20 , @xmath12 and @xmath23 axes .
the lower inset shows @xmath29 expanded near @xmath3 .
the upper inset shows the @xmath30 dependence of derivative of magnetization near @xmath3.,width=302 ] the main panel of figure [ mvst ] shows the temperature ( _ t _ ) dependence of the magnetization ( @xmath10 ) with the magnetic field ( @xmath31 ) applied along each of the principal crystallographic axes .
while cooling from 300 k , @xmath10 increases with decreasing @xmath30 and exhibits a broad maxima near 60 k. this behavior is typical of systems consisting of 1d chains of magnetic ions .
accordingly , we have used the bonner - fisher model @xcite for the spin 1/2 afm heisenberg chain to compare to the observed behavior with good agreement .
the fits yield an exchange coupling strength @xmath32
= 47.2(2 ) k , g@xmath33= 2.20(1 ) , g@xmath34= 2.20(1 ) and g@xmath35= 2.24(2 ) , which indicates a strong spin - orbit interaction and , perhaps , some minor g - factor anisotropy .
anisotropy in the g - factor was observed through electron - spin resonance measurements.@xcite at @xmath3 = 8.7 k long - range antiferromagnetic ( afm ) order is observed , as reported previously .
@xcite thus , the transition at @xmath3 represents a crossover from one- to three - dimensional antiferromagnetic behavior prominent anisotropy is observed below @xmath3 in @xmath10 , as shown in the lower inset of figure [ mvst ] .
this anisotropy is clearly evident in @xmath36 versus @xmath30 for the three field directions ; @xmath36 for @xmath37 and @xmath38 exhibits an abrupt increase at @xmath19 at the afm transition while cooling whereas it exhibits a small decrease for @xmath39 , as seen in the upper inset of fig .
[ mvst ] ; the subscript @xmath40 in the symbol @xmath41 denotes the direction along which @xmath31 was applied .
this behavior is typical for afm systems , and is largely consistent with the proposed ordering of the magnetic moments.@xcite ( color online ) temperature dependence of magnetization at different magnetic fields applied along the ( a ) @xmath20 ( b ) @xmath12 and ( c ) @xmath23 axes .
the numbers indicate the magnetic field strength(@xmath42 ) in t. the insets show the field dependence of @xmath10 for @xmath31 applied along the respective axes .
, width=302 ] to further examine the afm ordered state , the temperature dependence of @xmath10 at different magnetic fields applied along each crystallographic axis was measured .
[ mvstdiffh ] shows @xmath29 for magnetic field along each axis . for field strengths 0.5 and 1.2
t along @xmath20 , @xmath29 exhibits a downward drop below @xmath19= 8.7 k [ fig .
[ mvstdiffh](a ) ]
. however , for 1.3 t , this downward drop is followed by an increase below 5 k while cooling , which could be attributed to a spin - flop transition where the spins turn from a parallel orientation to a perpendicular orientation , thereby lowering their energy .
furthermore , the spin - flop transition shifts towards higher temperatures with increasing field strengths .
note that the spin - flop transition is also manifested as abrupt jumps in @xmath43 , for temperatures below @xmath19 , as shown in the inset of fig .
[ mvstdiffh](a ) ; in these data , it is evident that the transition occurs at lower field for lower temperatures .
this is due to the lowering of total energy of the spin - flop configuration compared to the antiferromagnetic state with increasing anisotropy and coupling strength as temperature decreases .
interestingly , @xmath29 and @xmath43 when @xmath31 is applied along @xmath12 [ fig .
[ mvstdiffh](b ) ] exhibit similar behavior , indicating that the spin - flop transition occurs in cusb@xmath0o@xmath6 for @xmath31 along both the @xmath20 and @xmath12 axes . since heinrich et al .
[ ] observed a spin - flop transition for the @xmath12-axis and not for directions perpendicular to @xmath12 , it was suggested that the easy antiferromagnetic axis lies along the @xmath12 direction .
however , our results confirm that cusb@xmath0o@xmath6 orders below 8.7 k with easy antiferromagnetic axis along either the @xmath20 or @xmath12 directions , in agreement with a prior report,@xcite which may reflect the presence of significant magnetic disorder , or weak differences in the magnetocrystalline energy for these two directions . as expected , we do not observe any spin flop transition in @xmath29 and @xmath43 for @xmath31 along @xmath23 [ fig .
[ mvstdiffh](c ) ] .
( a ) temperature ( _ t _ ) dependence of the linear thermal expansion ( @xmath44@xmath45 ) along the @xmath20 , @xmath12 , and @xmath23 axes . each curve contains more than 1700 data points .
inset shows the expanded region around @xmath19 .
( b ) comparison of temperature dependence of lattice parameters using dilatometer ( solid lines , this work ) and x - ray diffraction ( open symbols , ref . ) techniques.,width=302 ] in figure [ ltevst](a ) , the linear thermal expansion normalized to the length at 300 k , @xmath44@xmath45 , is plotted for the @xmath20 , @xmath12 and @xmath23 axes . here , @xmath46 is the length of the sample at @xmath47 300 k. upon warming from 5 k , @xmath44@xmath45 along @xmath20 and @xmath12 increases .
the magnitude of @xmath44@xmath45 along b shows a variation of @xmath48 between 5 and 300 k. on the other hand , @xmath44@xmath45 along @xmath23 decreases while warming from 5 - 150 k and thereafter increases with further increasing temperature .
the inset of fig .
[ ltevst](a ) displays @xmath44@xmath45 along @xmath20 and @xmath23 on the left and right scale , respectively , for @xmath30 near @xmath3 .
note that a clear change of slope is visible at @xmath19 along @xmath20 ( similar behavior occurs along @xmath12 ) and @xmath23 , as expected for a continuous phase transition .
the thermal expansion data are compared to data obtained using x - ray diffraction @xcite ( open symbols ) in fig .
[ ltevst](b ) .
our @xmath44@xmath45 data were multiplied by the respective lattice parameter at room temperature , and the result was then added to that lattice parameter in order to obtain the temperature dependence which appears as _ solid lines _ in the figure , which are actually about 1700 discrete data points .
the lattice parameters @xmath20 and @xmath12 tend to merge near 350 k due to the well - known monoclinic to tetragonal structural transition@xcite at 380 k. this comparison reveals the excellent agreement with the data obtained from diffraction as well as the outstanding resolution of our data .
( a ) temperature ( _ t _ ) dependence of linear thermal expansion coefficient ( @xmath49@xmath50@xmath51/@xmath52@xmath30 ) along the three crystallographic directions for cusb@xmath0o@xmath6 .
inset shows the expanded region around @xmath19 .
dotted lines are @xmath53 for the non - magnetic analog compound znsb@xmath0o@xmath6.,width=302 ] fig .
[ ltecvst ] shows the linear thermal expansion coefficient @xmath49@xmath50@xmath51/@xmath52@xmath30 for the @xmath20 , @xmath12 and @xmath23 axes .
while @xmath53 along @xmath20 and @xmath12 is positive over the entire temperature range , along @xmath23 it is negative at temperatures below 175 k and positive at higher temperatures .
moreover , @xmath53 displays a broad peak along @xmath20 and @xmath12 and a broad minimum along @xmath23 near 50 k. the expanded view of @xmath54 near the afm transition is given in the inset of fig .
[ ltecvst ] .
abrupt jumps occur in @xmath53 at the afm transition .
while @xmath53 along @xmath20 and @xmath12 jumps downward at the paramagnetic - afm transition on cooling , @xmath53 along @xmath23 shows an upward jump .
features in @xmath53 are expected at @xmath3 for a continuous phase transition , but the shape is unusual ( see below ) .
the main panel also shows @xmath53 for znsb@xmath0o@xmath6 along the @xmath20 and @xmath23 axes of the tetragonal crystal as dotted lines .
note that this non - magnetic analog compound exhibits monotonic behavior of @xmath53 and fails to exhibit @xmath55 0 along @xmath23 .
the unusual temperature dependence of the @xmath56 data for cusb@xmath0o@xmath6 is likely associated with its 1d antiferromagnetism and lower crystallographic symmetry , this point will be addressed below .
the heat capacity ( @xmath57 ) as a function of temperature is presented for cusb@xmath0o@xmath6 in fig .
[ fig5 ] . at the highest measurement temperature , @xmath30= 395
k ( data above 100 k not shown . ) , @xmath57 reaches 202.4 j / mol@xmath58k which is close to 224.4 j / mol@xmath58k calculated from dulong - petit .
a prominent anomaly occurs at @xmath19 , which will be discussed further below .
fitting to the equation @xmath59 over the range 9.2 k @xmath60 t @xmath60 15.6 k yields @xmath61 58.7(7 ) mj / mol@xmath58k@xmath62 , @xmath63 0.171(5 ) mj / mol@xmath58k@xmath64 and a debye temperature @xmath65 = 467(5 ) k. the main panel of fig .
[ fig5 ] also shows @xmath66 of znsb@xmath0o@xmath6 , the non - magnetic analog of cusb@xmath0o@xmath6 .
fitting these data over the range 0.45 k @xmath60 t @xmath60 15.6 k to the equation @xmath67 yields @xmath61 0.02(4 ) mj / mol@xmath58k@xmath62
@xmath63 0.11(2 ) mj / mol@xmath58k@xmath64 and a debye temperature @xmath65 = 539(30 ) k. subtraction of the znsb@xmath0o@xmath6 data from the cusb@xmath0o@xmath6 data yields @xmath68 , which is shown in the insets of fig .
[ fig5 ] as either @xmath68 or @xmath69 in expanded views .
the broad peak in @xmath69 above @xmath19 corresponds to the contribution of short - range - magnetic ordering to the entropy .
integrating the @xmath69 versus @xmath30 curve ( lower inset ) over the region below 115 k ( the temperature below which the cusb@xmath0o@xmath6 and znsb@xmath0o@xmath6 data sets deviate from one another ) provides the entropy change @xmath70 , which is plotted in the inset of fig .
[ mhcvst ] . an entropy change of 3.85 j / mol@xmath58k ( 66.8% of @xmath71 ) is observed over this entire temperature range .
if one considers only the region from just above @xmath3 to 115 k , 57.8% @xmath71 is obtained , so most of the entropy change occurs between 115 k and @xmath3 .
note that for the region 50 k to 115 k , the contribution to @xmath70 is only 12% rln2 .
some measurements of @xmath66 at 8 tesla , with the magnetic field parallel to the @xmath23 axis , were also conducted .
the zero field znsb@xmath0o@xmath6 data were subtracted from the @xmath72 data to obtain @xmath73 , which is shown in the insets of fig .
these data extend only slightly above @xmath3 , where a small increase in @xmath74 is evident .
the @xmath75 region will be discussed below .
main panel displays the molar heat capacity @xmath57 versus @xmath30 for cusb@xmath0o@xmath6 and its non - magnetic analog znsb@xmath0o@xmath6 ( shown as a dotted line ) .
the upper inset shows the temperature dependence of the magnetic contribution to the heat capacity ( @xmath68 ) plotted as ln(@xmath68 ) versus 1/@xmath30 for 0 and 8 tesla ; the data uncertainties are similar to the data point size .
the solid line represents the fit to the 0 tesla data to determine the energy gap in the equation @xmath76 .
the lower inset shows @xmath69 at 0 and 8 tesla ; the area under these curves is the entropy.,width=302 ] molar heat capacity at zero magnetic field after subtracting the entropy contributions , @xmath77 ( closed symbols ) and @xmath78 versus @xmath30 ; @xmath7955@xmath1510@xmath80 j / mol@xmath58k .
inset shows @xmath70 , which was obtained by integrating the @xmath69 versus @xmath30 data over the region below 115 k. , width=302 ] considering the thermodynamics of a continuous phase transition , the heat capacity in the immediate vicinity of @xmath3 can be written as@xcite @xmath81 here , @xmath82 , @xmath83 , @xmath84 and @xmath85 are the molar entropy , pressure , molar volume and volume thermal expansion coefficient , respectively ; the subscript @xmath86 signifies that eq .
( [ equation1 ] ) is valid near the nel temperature . for this analysis
, it is assumed that @xmath85 = ( @xmath87+@xmath88+@xmath89)@xmath15sin@xmath8 , near 8.7 k , since the crystallographic angle @xmath8 tends to saturate to 91.3@xmath24 below 100 k ( see fig . 2 of ref . [ ] ) .
when the entropy contribution is eliminated by subtracting the term that is linear in @xmath30 from @xmath66 , the result , @xmath90 , is proportional to the product of @xmath85 and @xmath30 . as expected from eq .
( [ equation1 ] ) and shown in fig .
[ mhcvst ] , close to @xmath19 , @xmath90 scales with @xmath91 following the relation @xmath92 .
the excellent overlap of the two data sets is obtained with @xmath7955@xmath1510@xmath80 j / mol@xmath58k .
furthermore , the observation that the data sets scale suggests that the phase transition is continuous .
the parameter @xmath93 can now be used to determine @xmath94 since @xmath95 . the molar density at 300 k provides @xmath84 as 6.8 @xmath15 10@xmath96 m@xmath97/mol , which then yields the pressure derivative @xmath94= -0.11(1 ) k / gpa .
note that there are no direct measurements of @xmath94 available thus far .
the @xmath66 data reveal a sizable @xmath98 term of magnitude @xmath61 57.4(6 ) mj / mol@xmath58k@xmath62 , that completely vanishes below @xmath3 . since cusb@xmath0o@xmath6 is an electrical insulator , this term can not be associated with the presence of conduction electrons and must have a purely magnetic origin . for a heisenberg chain , theory@xcite reveals @xmath98 = @xmath99 120 mj / mol@xmath58k@xmath62 using the value of @xmath32 obtained from fitting our magnetic susceptibility data .
thus , the obtained @xmath98 is approximately half of the theoretical value .
the decay of @xmath66 below @xmath3 is exponential , suggesting the presence of an energy gap .
this gap is extracted from a linear fit to the data plotted as ln(@xmath68 ) versus 1/@xmath30 ( see upper inset of fig . [ fig5 ] ) , suggesting the function @xmath68 @xmath18 @xmath100 .
the gap @xmath4= 17.48(6 ) k ( @xmath181.51(1 ) mev ) is obtained from the fit .
this gap must be associated with the crossover from partially - ordered moments on non - interacting chains to three - dimensional , long - range order .
measurements of @xmath66 at an applied magnetic field of 8 t have a small influence on the obtained @xmath4 , yielding @xmath4= 18.42(6 ) k ( @xmath181.59(1 ) mev ) , which suggests that the low - temperature ordered phase is more robust in non - zero magnetic field .
the character of the phase transition at @xmath3 warrants some discussion .
normally the heat capacity and thermal expansion coefficient of a magnetic system displays @xmath93-like character at the magnetic ordering temperature , which is associated with a divergence of measurable quantities , such as @xmath66 , @xmath101 , @xmath102 and @xmath10 at the critical point.@xcite the antiferromagnetic analog system cosb@xmath0o@xmath6 , for example , clearly exhibits @xmath93-like character in @xmath66 at its afm transition.@xcite this is not the case for cusb@xmath0o@xmath6 , where @xmath66 and @xmath53 display discontinuities at @xmath3 followed by an exponential decay in @xmath66 below @xmath3 . the character of these transitions is reminiscent of that observed at the normal - superconducting phase transition , the only phase transition known to exhibit distinct jumps at the critical temperature.@xcite note that other@xcite 1d magnetic systems , such as cugeo@xmath5 and nav@xmath17o@xmath103 exhibit @xmath93-like character in @xmath66 and @xmath53 .
these systems , however , possess strong lattice distortions@xcite at the spin - peierls transition ( in the case of cugeo@xmath5 ) and charge - ordering transition ( in nav@xmath17o@xmath103 ) .
this magnetic - lattice coupling may play a role in the large @xmath93-like features in @xmath53 at their respective transitions . in the case of cusb@xmath0o@xmath6 ,
it exhibits a transition from 1d to 3d order at @xmath3 , _ without _ an obvious presence of strong magnetic - lattice coupling . a possible scenario that agrees with the observed discontinuity of specific heat , the energy gap below @xmath3 , and a weak magnetic - lattice coupling is described as follows .
the spin - peierls transition originates from the coupling between the 1-d _ jordan - wigner - transformed _ fermions and 3-d phonons .
similar to the usual peierls transition , it opens up an energy gap that is very much bcs - like.@xcite in our system , there is a collection of quasi-1d spin chains that run parallel to one another .
the same group of phonons that couple spins on one chain may also couple spins on the other chains .
this introduces an effective inter - chain coupling .
if this coupling remains marginally weak , upon the dimerization ( spin - peierls ) transition , the energy gap opened should behave essentially like the case of a single chain while the dimerization ordering of all chains will then appear as a 3-d ordering .
if this is the case , the specific - heat discontinuity occurs in the same manner as the bcs case and the spin - energy gap appears as an additional contribution of specific heat the same way as the heat capacity contribution of superconducting electron pairs .
this ordered state is likely an alignment of the 1d chains , rather than a true 3d ordered state ; this may be the source of difficulty in identification@xcite of the magnetic lattice below @xmath3 and the disorder suggested from the spin - flop transition s occurrence along both the @xmath20 and @xmath12 axes .
the nearly tetragonal crystal structure of cusb@xmath0o@xmath6 results in almost identical positive thermal expansion behavior observed along the @xmath20 and @xmath12 axes below 200 k. above this temperature @xmath56 differs for these two axes due to the well - known monoclinic to tetragonal structural transition at 380 k.@xcite in contrast , the thermal expansion along @xmath23 is negative at temperatures below about 170 k. negative thermal expansion can originate from unusual phonon modes that become active at low temperatures@xcite and are often observed in low - d systems.@xcite however , consideration of @xmath56 for the non - magnetic analog compound znsb@xmath0o@xmath6 , along with the fact that 1d antiferromagnetism sets in below 115 k ( see @xmath70 versus @xmath30 in the inset of fig . [ mhcvst ] ) suggests that the unusual temperature dependence of @xmath56 in cusb@xmath0o@xmath6 is associated with the 1d antiferromagnetism .
the thermal expansion in solids is due to anharmonic vibrations , and in turn , anharmonic contributions to the elastic potential .
this is a many - body potential , but pair potentials between neighboring atoms can play a significant role.@xcite in cusb@xmath17o@xmath6 the onset of local 1d magnetic order occurs on cooling below @xmath18115 k , as determined from the heat capacity @xmath68 .
local order would strongly affect the pair potentials in those regions of the sample where the order occurs . this could lead to stretching or tilting of bonds as well as changes in bond strength in these regions .
even if the affected pair potentials are harmonic , the result can be a potential for the _ sample _ that is anharmonic.@xcite thus , the increase in magnitude of the thermal expansion coefficients below @xmath18 115 k is most likely associated with changes in anharmonic contributions to the elastic potential of the sample as a result of the formation of short - range 1d antiferromagnetic order . below @xmath1850
k the thermal expansion coefficients change behavior because a large proportion of the spins are ordered , and the anharmonic contributions reduce upon further cooling leading to a decrease in the magnitude of @xmath53 .
the negative thermal expansion along @xmath23 is in strong contrast to the behavior along @xmath20 and @xmath12 , and is certainly connected to the 1d behavior between 8.7 k and 115 k and anharmonic lattice vibrations that specifically affect @xmath53 along this direction .
other types of transitions can also lead to the development of anharmonic contributions to the thermal expansion , such as in li@xmath104mo@xmath1o@xmath105 where at 28 k a crossover in dimension has been associated with two - particle hopping and the formation of bosons.@xcite cugeo@xmath5 also exhibits@xcite a prominent correlation between thermal expansion and the formation of 1d magnetic correlations .
the thermodynamic and magnetic properties of cusb@xmath0o@xmath1 were investigated .
one - dimensional antiferromagnetic order among the spin @xmath14 cu atoms is observed to occur below @xmath18115 k. this order has a strong impact on anharmonic lattice vibrations , and leads to a sizable change in the thermal expansion coefficients . at @xmath3 = 8.7 k , a crossover from one- to three - dimensional antiferromagnetic behavior occurs .
this long - range antiferromagnetic order leads to distinct jumps in the heat capacity and thermal expansion coefficients , which are reminiscent of those observed at the normal - superconducting phase transition .
the heat capacity data reveals the presence of a gap below @xmath19 of magnitude @xmath106 = 17.48(6 ) k. the occurrence of three - dimensional antiferromagnetic order is attributed to a spin - peierls transition occurring as a result of three - dimensional phonons coupling with _ jordan - wigner - transformed _ fermions .
spin - flop transitions for magnetic field applied in @xmath20 and @xmath12 axes below @xmath19 indicate that the easy afm axis lies in both the @xmath20 and @xmath12 directions , which may reflect the presence of significant magnetic disorder , or weak differences in the magnetocrystalline energy for these two directions
. 1 i. affleck , j. phys . : condens .
matter * 1 * , 3047 ( 1989 ) j. c. bonner and m. e. fisher , phys . rev . * 135 * , a640 ( 1964 ) .
m. hase , i. terasaki , and k. uchinokura , phys .
lett . * 70 * , 3651 ( 1993 ) .
j. p. pouget , l. p. regnault , m. ain , b. hennion , j. p. renard , p. veillet , g. dhalenne , and a. revcolevschi , phys .
. lett . * 72 * , 4037 ( 1994 ) .
e. pytte , phys .
b * 10 * , 4637 ( 1974 )
. i. s. jacobs , j. w. bray , h. r. hart , jr .
, l. v. interrante , j. s. kasper , g. d. watkins , d. e. prober , and j. c. bonner , phys . rev .
b 14 , 3036 ( 1976 ) .
a. bystrom , b. hok , and b. mason , ark .
b * 15 * , 1 ( 1941 ) .
a. nakua , h. yun , j. n. reimers , j. e. greedan , and c. v. stager , j. solid state chem .
* 91 * , 105 ( 1991 ) .
a. m. nakua and j. e. greedan , j. solid state chem .
* 118 * , 1999 ( 1995 ) .
b. j. gibson , r. k. kremer , a. v. prokofiev , w. assmus , and b. ouladdiaf , j. magn .
magn . mat . *
272 * , 927 ( 2004 ) .
m. kato , k. kajimoto , k. yoshimuar , k. kosuge , m. nishi , and k. kakurai , j. phys .
jpn . * 71 * suppl . , 187 ( 2002 ) .
a. v. prokofiev , f. ritter , w. assmus , b. j. gibson , and r. k. kremer , j. crystal growth * 247 * , 457 ( 2003 ) .
h. mller - bauschbaum , agnew.chem .
* 89 * , 704 ( 1977 ) .
giere , a. brahimi , h. j. deiseroth , and d. reinen , j. solid state chem .
* 131 * , 263 , ( 1997 ) .
j. d. donaldson , a. kjekshus , d. g. nicholson , and t. rakke , acta chem .
scand . * 29 * , 803 ( 1975 ) . j. n. reimers , j. e. greedan , c. v. stager , and r. kremer , j. solid state chem . , * 83 * , 20 ( 1989 ) ; e. ramos , m. l. veiga , f. fernndez , and r. saez - puche , j. solid state chem . , * 91 * , 113 ( 1991 ) . e. ramos , f. fernndez , a. jerez , c. pico , j. rodriguez - carvajal , r. saez - puche , and m. l. veiga , mater .
* 27 * , 1041 ( 1992 ) . h. ehrenberg , g. wltschek , j. rodriguez - carvajal , and t. vogt , j. magn .
magn . mat . * 184 * , 111 ( 1998 ) .
m. hase , j. magn . mag .
materials * 177 - 181 * , 611 ( 1998 ) .
d. kasinatham , k. koepernik , and h. rosner , phys .
* 100 * , 237202 ( 2008 ) .
m. heinrich , h .- a .
krug von nidda , a. krimmel , a. loidl , r. m. eremina , a. d. ineev , b. i. kochelaev , a. v. prokofiev , and w. assmus , phys .
b * 67 * , 224418 ( 2003 ) .
a. katsui , and h. matsushita , phys .
( a ) * 203 * , 2832 ( 2006 ) .
s. matsushima , t. tanizaki , h. nakamura , m. nonaka , and m. arai , chem .
lett * 10 * , 1010 ( 2001 ) .
j. j. neumeier , r. k. bollinger , g. e. timmins , c. r. lane , r. d. krogstad , and j. macaluso , rev .
instrum . *
79 * , 033903 ( 2008 ) .
j. a. souza , y .- k .
yu , j. j. neumeier , h. terashita , and r. f. jardim , phys . rev . lett . * 94 * , 207209 ( 2005 ) .
d. c. johnston , r. k. kremer , m. troyer , x. wang , a. klmper , s. l. budko , a. f. panchula , and p. c. canfield , phys .
b * 61 * , 9558 ( 2000 ) .
b. d. white , j. a. souza , c. chiorescu , j. j. neumeier , and j. l. cohn , phys .
b * 79 * , 104427 ( 2009 ) .
a. christian , m. g. smith , a. rebello and j. j. neumeier , unpublished .
h. winkelmann , e. gamper , b. buchner , m. braden , a. revcolevschi , and g. dhalenne , phys .
b * 51 * , 12884 ( 1995 ) .
g. d. barerra , j. a. o. bruno , t. h. k. barron , and n. l. allan , j. phys . : condens .
matter * 15 * , r217 ( 2005 ) c. a. m. dos santos , b. d. white , yi - kuo yu , j. j. neumeier , and j. a. souza , phys .
rev . lett .
* 98 * , 266405 ( 2007 ) .
a. rebello , j. j. neumeier , z. gao , y. qi , and y. ma , phys .
b * 86 * , 104303 ( 2012 ) . | measurements of magnetic susceptibility , heat capacity and thermal expansion are reported for single crystalline cusb@xmath0o@xmath1 in the temperature range @xmath2 k. the magnetic susceptibility exhibits a broad peak centered near 60 k that is typical of one - dimensional antiferromagnetic compounds .
long - range antiferromagnetic order at @xmath3 = 8.7 k is accompanied by an energy gap ( @xmath4 = 17.48(6 ) k ) .
this transition represents a crossover from one- to three - dimensional antiferromagnetic behavior . both heat capacity and the thermal expansion coefficients exhibit distinct jumps at @xmath3 , which are similar to those observed at the normal - superconducting phase transition in a superconductor .
this behavior is quite unusual , and is presumably associated with a spin - peierls transition occurring as a result of three - dimensional phonons coupling with _ jordan - wigner - transformed _ fermions . |
the euler - heisenberg effective action @xcite has been widely investigated in the last decades , in various contexts ( see @xcite and references therein ) .
nevertheless , studies related to the lorentz - symmetry violation @xcite have been considered so far in the context of photon splitting @xcite .
the effective theory used in these analyses was mainly the lorentz - violating qed @xcite , a subset of the minimal standard - model extension @xcite .
for an overview and references on this extended qed , see the review @xcite and the incomplete list of works @xcite . in this paper , we are interested in studying the radiative generation of the lorentz - violating euler - heisenberg action , from the cpt - even derivative term @xmath3 of the lorentz - violating qed extension .
the cpt - odd term @xmath4 also generates nonlinear corrections to the maxwell theory , however , these calculations well be presented in a next work . with the exact expressions for these results we will be able to calculate specific scattering amplitudes , e.g. , the photon scattering in the electromagnet field of a nucleus ( delbrck scattering ) @xcite and the photon splitting in a strong magnetic field @xcite , in order to numerically estimate the coefficients for the lorentz violation .
recently it has been argued that the photon - photon scattering ( also a process calculated from the results of the euler - heisenberg action ) can be observed at the large hadron collider @xcite as an opportunity to discuss the noncommutative interactions @xcite , among others , and therefore also the lorentz violation effects .
the structure of the paper is as follows . in sec .
[ nonperturbative ] we consider a nonperturbative calculation in the coefficient @xmath0 , by assuming rotational invariance in which @xmath0 is reduced to the product of timelike vectors .
we also recover the results of the amplitude for vacuum photon splitting , previously obtained in @xcite . in sec .
[ perturbative ] we take into account the perturbative approach , where the coefficient @xmath0 is treated as a insertion in the propagator as well as a new vertex . sec . [ summary ] contains a summary of out results .
the starting fermion lagrangian that we are interested is given by @xmath5 where we have written @xmath6 and @xmath7 . as we have mentioned above , in our nonperturbative approach the coefficient @xmath0
is reduced to the product of two unit timelike vectors , i.e. , @xmath8 , where @xmath9 and @xmath10 is the coefficient that determine the scale of lorentz violation .
hence , the corresponding feynman rules are the fermion propagator , @xmath11[1.0cm,.5cm]{ferm.eps } } } = \frac{i}{\slashed{\tilde p}-m},\ ] ] and the fermion - photon vertex , @xmath11[1.0cm,1.5cm]{vert.eps } } } = -ie\gamma^\mu,\ ] ] with @xmath12 and @xmath13 .
thus , the resulting effective action takes the form @xmath14 where @xmath15 with @xmath16[3.5cm,3.0cm]{quartic1.eps } } } & = & i\,t_1^{\mu_1\mu_2\mu_3\mu_4}(k_1,k_2,k_3,k_4 ) \nonumber\\ \nonumber\\ & = & -(-ie)^4\int \frac{d^4p}{(2\pi)^4}{\rm tr}\frac{i}{\slashed{\tilde p}-m}\gamma^{\mu_1}\frac{i}{\slashed{\tilde p}_1-m}\gamma^{\mu_2}\frac{i}{\slashed{\tilde p}_{12}-m}\gamma^{\mu_3}\frac{i}{\slashed{\tilde p}_{123}-m}\gamma^{\mu_4},\end{aligned}\ ] ] @xmath17[3.5cm,3.0cm]{quartic2.eps } } } & = & i\,t_2^{\mu_1\mu_2\mu_3\mu_4}(k_1,k_2,k_3,k_4 ) \nonumber\\ \nonumber\\ & = & -(-ie)^4\int \frac{d^4p}{(2\pi)^4}{\rm tr}\frac{i}{\slashed{\tilde p}-m}\gamma^{\mu_2}\frac{i}{\slashed{\tilde p}_2-m}\gamma^{\mu_1}\frac{i}{\slashed{\tilde p}_{12}-m}\gamma^{\mu_3}\frac{i}{\slashed{\tilde p}_{123}-m}\gamma^{\mu_4},\end{aligned}\ ] ] @xmath18[3.5cm,3.0cm]{quartic3.eps } } } & = & i\,t_3^{\mu_1\mu_2\mu_3\mu_4}(k_1,k_2,k_3,k_4 ) \nonumber\\ \nonumber\\ & = & -(-ie)^4\int \frac{d^4p}{(2\pi)^4}{\rm tr}\frac{i}{\slashed{\tilde p}-m}\gamma^{\mu_4}\frac{i}{\slashed{\tilde p}_4-m}\gamma^{\mu_2}\frac{i}{\slashed{\tilde p}_{24}-m}\gamma^{\mu_3}\frac{i}{\slashed{\tilde p}_{234}-m}\gamma^{\mu_1}.\end{aligned}\ ] ] in the above expressions , @xmath19 , the global factor @xmath20 in eq .
( [ g ] ) means the two orientations of the fermion loop , and @xmath21 , @xmath22 , and so on . here
, we observe that the contributions ( [ t2 ] ) and ( [ t3 ] ) can be obtained from ( [ t1 ] ) , when we perform the interchanges : [ t2t3 ] @xmath23 therefore , we just need to calculate the expression ( [ t1 ] ) , in which , by using first the feynman parameterization , we arrive at the expression @xmath24,\end{aligned}\ ] ] where @xmath25 and @xmath26 with @xmath27 , @xmath28 , and @xmath29 . in order to calculate the momentum integrals in ( [ t1a ] )
, we will use the fact that @xmath30 where @xmath31 and @xmath32 .
consequently , the effect of lorentz violation will be concentrated in the overall factor @xmath33 and into the mass parameter @xmath34 .
thus , by calculating the trace over dirac matrices and the integrals , we obtain , up to order @xmath35 , the following results : @xmath36 where @xmath37(1+\kappa)^{-1 } \nonumber\\ & & \times ( g^{\mu_1\mu_2}g^{\mu_3\mu_4}-2g^{\mu_1\mu_3}g^{\mu_2\mu_4}+g^{\mu_1\mu_4}g^{\mu_2\mu_3}),\end{aligned}\ ] ] with @xmath38 and @xmath39 , and [ t1gg ] @xmath40,\end{aligned}\ ] ] @xmath41,\end{aligned}\ ] ] @xmath42,\end{aligned}\ ] ] @xmath43.\end{aligned}\ ] ] the other contributions of eq .
( [ t1b ] ) , @xmath44 and @xmath45 , have been omitted for the sake of brevity .
note that we have adopted dimensional regularization by extending the spacetime from @xmath46 to @xmath47 dimensions , so that @xmath48 goes to @xmath49 , where @xmath50 is an arbitrary parameter that identifies the mass scale .
as we will see below , from these results ( [ t1b ] ) we will be able to obtain the lorentz - violating euler - heisenberg action in both nonperturbative and perturbative approaches .
let us now recover the ( perturbative in @xmath0 ) amplitude obtained in @xcite for the photon triple splitting , from the above nonperturbative results ( [ t1b ] ) , in the collinear limit .
for this , if we consider that the incident on - shell photon has energy @xmath51 and momentum @xmath52 , from the momentum conservation @xmath53 , we have the inequation @xmath54 then , as @xmath55 or @xmath56 ( with @xmath57 ) , in order to also satisfy the energy conservation @xmath58 , all the momenta @xmath59 must be aligned
. therefore , the incident photon and the decay photons must be collinear , such that the four - momenta of all photons are mutually orthogonal , @xmath60 . by considering that these four - momenta are proportional to some four - momentum @xmath61 ,
satisfying @xmath62 , we can write @xmath63 , so that @xmath64 , in which @xmath65 are now scalar coefficients , instead four - momenta . with regard to the transversality condition of the polarization four - vectors , as usual , @xmath66 ( or @xmath67 ) . thus , due to the requirement of collinearity
, we also have that @xmath68 ( or @xmath69 ) .
taking into account these considerations , we find @xmath70 \nonumber\\ & & + { \cal o}(\tilde k_0 ^ 4)+{\cal o}(\tilde k_0^{\mu_i}),\end{aligned}\ ] ] where we have interchanged @xmath71 and @xmath72 and thereafter @xmath71 and @xmath73 , however without changing the momentum indices , in order to obtain @xmath74 and @xmath75 ( eqs .
( [ t2 ] ) and ( [ t3 ] ) ) from @xmath76 ( eq . ( [ t1b ] ) ) , following the prescription given in ref . @xcite .
note that , as expected , the divergent term ( [ t1ge ] ) vanishes , and , for simplicity , we have not included the terms of @xmath77 and @xmath78 .
now , when we expand up to first order in @xmath10 , the lorentz - violating contributions become @xmath79(1+\kappa)^{-1 } \nonumber\\ & = & 2\kappa(k_0 ^ 0)^2+{\cal o}(\kappa^2).\end{aligned}\ ] ] therefore , the perturbative result is readily recovered ( see eq .
( 10 ) of ref .
@xcite ) , which is given by @xmath80,\end{aligned}\ ] ] with @xmath81 , however , for @xmath82 and @xmath2 . we will now discuss the generation of the lorentz - violating euler - heisenberg action , in the first order correction , @xmath83 .
for this , we must calculate @xmath84 ( eq . ( [ g ] ) ) without regard to the collinear limit .
thus , as pointed out at the beginning , @xmath74 and @xmath75 are obtained from @xmath76 , when we interchange the uncontracted indices as well as the momentum indices , see eq .
( [ t2t3 ] ) .
proceeding in this way , we obtain @xmath85 where @xmath86,\end{aligned}\ ] ] [ ggk ] @xmath87,\end{aligned}\ ] ] @xmath88,\end{aligned}\ ] ] @xmath89,\end{aligned}\ ] ] @xmath90,\end{aligned}\ ] ] @xmath91,\end{aligned}\ ] ] @xmath92,\end{aligned}\ ] ] @xmath93,\end{aligned}\ ] ] with @xmath94 .
we observe that all the contributions of order @xmath95 completely vanish , as expected .
in fact , the euler - heisenberg actions are all proportional to order @xmath35 , in the first order correction .
therefore , by considering these results ( [ ga ] ) , the effective action ( [ seff ] ) takes the following form @xmath96 where @xmath97 so that we can write @xmath98 with @xmath99 , and so on . then , inverting the fourier transform in the expression ( [ seh ] ) , the lorentz - violating euler - heisenberg action becomes @xmath100 where now @xmath101 .
the above result is somewhat unusual , although it resembles the conventional euler - heisenberg action @xcite . from this expression ( [ seh2 ] ) ( or from the equations ( [ t1b ] ) ) we are now able to calculate some scattering amplitudes of interest in the presence of lorentz violation ( see sec .
[ intro ] , for some proposals ) .
it is worth mentioning that the lagrangian of eq .
( [ seh2 ] ) is the leading term of the expansion of the euler - heisenberg effective lagrangian , in the weak - field limit , which , written in terms of the dual field strength @xmath102 and the critical field strength @xmath103 v / cm , can be reexpressed as @xmath104 where @xmath105 with @xmath106 ( note that we have reinstated @xmath107 and @xmath108 ) . in this limit ,
the electric and magnetic fields @xmath109 and @xmath110 are weak in comparison with the critical field @xmath111 .
in fact , the limit @xmath112 is essential for the validity of the effective lagrangian as well as imposes severe restrictions on the real ( electron - positron ) pair production .
we refer the reader to ref .
@xcite , for a more complete discussion .
while the above expression ( [ leh2 ] ) is sufficient to analyze the processes of ( four - photon ) box diagrams , for more external legs , we can infer from the euler - heisenberg effective lagrangian the other expressions , e.g. , for ( six - photon ) hexagon diagrams , we can write @xmath113 where @xmath114 in particular , the lagrangiana ( [ leh3 ] ) is that responsible for the first non - zero contribution to the photon splitting in an ( constant and spatially uniform ) external magnetic field @xmath115 @xcite , in which @xmath116 g ( i.e. , a strong field ) .
the limit of very strong magnetic fields ( @xmath117 ) is also interesting , however , it must be taken into account in the complete euler - heisenberg effective lagrangian .
we believe that the lorentz - violating version of this effective lagrangian resembles its original shape , except for the overall factor @xmath33 and for the field strength @xmath118 and its dual . nevertheless , this will be confirmed in a forthcoming study . in general , calculations involving lorentz violation are performed in the first order of the coefficient that controls the violation .
thus , by expanding the action ( [ seh2 ] ) up to first order in @xmath10 , we obtain @xmath119 where we have introduced the coefficient @xmath120 , given by @xmath121 in the next section we will see that the second and third contributions of eq .
( [ seh3 ] ) are in fact found from the perturbative approach , however , for a generic coefficient @xmath0 .
an interesting point about the decomposition of @xmath120 ( eq . ( [ kf ] ) ) , in terms of a symmetry tensor @xmath0 , is that it is also found in the born - infeld electrodynamics , with an ( lorentz - violating ) external field @xcite . in particular
, this decomposition restricts the lorentz - violating electrodynamics to a nonbirefringence sector @xcite .
in order to take into account the perturbative approach , let us first rewrite the fermion lagrangian ( [ l ] ) as follows @xmath122 in which , from now on , @xmath0 is a generic coefficient for lorentz violation .
now , by expanding the propagator up to first order in @xmath0 , @xmath123 so that we can consider @xmath124 as a insertion in the propagator @xmath125 , and treating the last term in ( [ l2 ] ) as a new vertex , the effective action ( [ seff ] ) then becomes @xmath126 with @xmath127 where now each graph above is subdivided into eight graphs , @xmath128 given by @xmath129 @xmath130 @xmath131 @xmath132 once more , it is easy to see that @xmath133 and @xmath134 are obtained from @xmath135 , as well as @xmath136 , @xmath137 , and @xmath138 from @xmath139 , when we also perform the ciclic interchanges : [ tc1234 ] @xmath140 therefore , we must focus our attention on the two graphs of @xmath139 ( eq .
( [ tc11 ] ) ) , in which , by considering first the feynman parameterization , we obtain @xmath141\nonumber\\ & & - \int_0 ^ 1 dx_1 \int_0^{1-x_1 } dx_2 \int_0^{1-x_{12 } } dx_3 \int \frac{d^4p}{(2\pi)^4}\frac{24ixe^4}{(p^2-m^2)^5}\nonumber\\ & & \times{\rm tr}[(\slashed{q}+m)c_{\mu\nu}q^\mu\gamma^\nu(\slashed{q}+m)\gamma^{\mu_1}(\slashed{q}_1+m)\gamma^{\mu_2}(\slashed{q}_{12}+m)\gamma^{\mu_3}(\slashed{q}_{123}+m)\gamma^{\mu_4}].\end{aligned}\ ] ] following the standard procedure , after we calculate the trace over dirac matrices and the corresponding integrals , up to order @xmath35 , we arrive at @xmath142 where @xmath143(-g^{\mu _ 1\mu _ 2 } c^{\mu _ 3\mu _ 4}+c^{\mu _ 1\mu _ 2 } g^{\mu _ 3\mu _ 4}-4 c^{\mu_1\mu _ 3}g^{\mu _
2\mu _ 4}\nonumber\\ & & + 2 c^{\mu_1\mu_4}g^{\mu _ 2\mu _ 3}+2g^{\mu
_ 1\mu _ 4 } c^{\mu _ 2\mu _ 3}+c_{\mu\nu}g^{\mu\nu } g^{\mu _ 1\mu _ 3 } g^{\mu _ 2\mu _ 4}-c_{\mu\nu}g^{\mu\nu } g^{\mu _ 1\mu _ 4}g^{\mu _ 2\mu _ 3}),\end{aligned}\ ] ] [ c11gg ] @xmath144,\end{aligned}\ ] ] @xmath145,\end{aligned}\ ] ] @xmath146,\end{aligned}\ ] ] @xmath147.\end{aligned}\ ] ] in the above results we have omitted some contributions of @xmath148 and all of @xmath149 and @xmath150 , because they are very lengthy .
in fact , the expressions ( [ c11gg ] ) are also extensive , however , it is worth considering them here , as we will see below .
finally , by taking into account the interchanges ( [ tc1234 ] ) , the first graph @xmath135 takes the form : @xmath151 where @xmath152\nonumber\\ & & \times[2 g^{\mu_1\mu_4 } c^{\mu_2\mu_3}+2 g^{\mu_1\mu_2 } c^{\mu_3\mu_4}-4 g^{\mu_1\mu_3 } c^{\mu_2\mu_4}\nonumber\\ & & + 2 c^{\mu_1\mu_2 } g^{\mu_3\mu_4}-4 g^{\mu_2\mu_4 } c^{\mu_1\mu_3}+2g^{\mu_2\mu_3 } c^{\mu_1\mu_4}\nonumber\\ & & -c_{\mu\nu}g^{\mu\nu}(g^{\mu_1\mu_4 } g^{\mu_2\mu_3}-2 g^{\mu_1\mu_3 } g^{\mu_2\mu_4}+g^{\mu_1\mu_2 } g^{\mu_3\mu_4})]\end{aligned}\ ] ] and @xmath153 thus , comparing the above perturbative expressions ( [ tc1gg ] ) with the nonperturbative ones ( [ t1gg ] ) we observe that they are similar , being the results of eq .
( [ tc1 ] ) an expansion up to first order in @xmath0 of the equation ( [ t1b ] ) , with @xmath1 and @xmath2 .
note that , as expected , the expression ( [ gcoll ] ) for the vacuum photon splitting is obtained from ( [ tc1gg ] ) , when we consider the collinear limit .
therefore , we can easily deduce the action coming from @xmath135 ( [ tc1 ] ) whose expression is in fact the perturbative lorentz - violating euler - heisenberg action , previously obtained in ( [ seh3 ] ) , nevertheless here , for a generic coefficient @xmath0 , given by @xmath154 where @xmath120 is defined in eq .
( [ kf ] ) .
in this work , we have studied the radiative generation of the lorentz - violating euler - heisenberg action , in the weak field approximation .
firstly , we have considered a nonperturbative calculation in the coefficient @xmath0 , however , by assuming rotational invariance , such that only @xmath82 .
the preliminary results are presented in eq .
( [ t1b ] ) , resulting then in the action ( [ seh2 ] ) . from these expressions
, we are now able to calculate some scattering amplitudes relating to four - photon diagrams , in order to numerically estimate the coefficient for lorentz violation . within this approach
, we have also recovered the results of the amplitude for the photon triple splitting ( [ gcoll ] ) , previously obtained in the literature .
finally , we have taken into account the perturbative approach , where @xmath0 is treated as a insertion in the propagator and a new vertex .
the partial results are shown in eq .
( [ tc1 ] ) , which are in fact an expansion up to first order in @xmath0 of eq .
( [ t1b ] ) , with @xmath1 and @xmath2 .
this suggest that the complete results ( [ t1b ] ) obtained in the nonperturbative approach can be used in both treatments .
an extension of this work would be to consider higher - derivative terms , e.g. , see the lorentz - violating qed with operator of mass dimension six of ref .
@xcite .
w. heisenberg and h. euler , z. phys .
* 98 * , 714 ( 1936 ) [ physics/0605038 ] . j. s. schwinger , phys . rev . *
82 * , 664 ( 1951 ) .
g. v. dunne , int .
j. mod .
a * 27 * , 1260004 ( 2012 ) [ arxiv:1202.1557 [ hep - th ] ] .
d. colladay and v. a. kostelecky , phys .
d * 58 * , 116002 ( 1998 ) [ hep - ph/9809521 ] .
d. colladay and v. a. kostelecky , phys .
d * 55 * , 6760 ( 1997 ) [ hep - ph/9703464 ] .
v. a. kostelecky , phys .
d * 69 * , 105009 ( 2004 ) [ hep - th/0312310 ] .
t. jacobson , s. liberati and d. mattingly , phys .
d * 67 * , 124011 ( 2003 ) [ hep - ph/0209264 ] . c. adam and f. r. klinkhamer , nucl .
b * 657 * , 214 ( 2003 ) [ hep - th/0212028 ] .
v. a. kostelecky and a. g. m. pickering , phys .
lett . * 91 * , 031801 ( 2003 ) [ hep - ph/0212382 ] .
g. gelmini , s. nussinov and c. e. yaguna , jcap * 0506 * , 012 ( 2005 ) [ hep - ph/0503130 ] . c. kaufhold and f. r. klinkhamer , nucl .
b * 734 * , 1 ( 2006 ) [ hep - th/0508074 ] .
f. a. brito , e. passos and p. v. santos , europhys .
lett . * 95 * , 51001 ( 2011 ) [ arxiv:1101.5390 [ hep - th ] ] .
v. a. kostelecky , c. d. lane and a. g. m. pickering , phys .
d * 65 * , 056006 ( 2002 ) [ hep - th/0111123 ] .
r. bluhm , lect .
notes phys .
* 702 * , 191 ( 2006 ) [ hep - ph/0506054 ] .
r. jackiw and v. a. kostelecky , phys .
lett . * 82 * , 3572 ( 1999 ) [ hep - ph/9901358 ] .
v. a. kostelecky and m. mewes , phys .
* 87 * , 251304 ( 2001 ) [ hep - ph/0111026 ] .
v. a. kostelecky and m. mewes , phys .
d * 66 * , 056005 ( 2002 ) [ hep - ph/0205211 ] .
v. a. kostelecky and m. mewes , phys .
d * 80 * , 015020 ( 2009 ) [ arxiv:0905.0031 [ hep - ph ] ] .
r. casana , m. m. ferreira , jr , j. s. rodrigues and m. r. o. silva , phys .
d * 80 * , 085026 ( 2009 ) [ arxiv:0907.1924 [ hep - th ] ] .
m. gomes , t. mariz , j. r. nascimento , a. y. .petrov , a. f. santos and a. j. da silva , phys .
d * 81 * , 045013 ( 2010 ) [ arxiv:0910.4560 [ hep - th ] ] .
r. casana , m. m. ferreira , a. r. gomes and f. e. p. dos santos , phys . rev .
d * 82 * , 125006 ( 2010 ) [ arxiv:1010.2776 [ hep - th ] ] .
t. mariz , phys .
d * 83 * , 045018 ( 2011 ) [ arxiv:1010.5013 [ hep - th ] ] .
g. jarlskog , l. joensson , s. pruenster , h. d. schulz , h. j. willutzki and g. g. winter , phys .
d * 8 * , 3813 ( 1973 ) .
a. i. milstein and m. schumacher , phys .
* 243 * , 183 ( 1994 ) .
z. bialynicka - birula and i. bialynicki - birula , phys .
d * 2 * , 2341 ( 1970 ) .
s. l. adler , j. n. bahcall , c. g. callan and m. n. rosenbluth , phys .
lett . * 25 * , 1061 ( 1970 ) .
s. l. adler , annals phys . * 67 * , 599 ( 1971 ) .
s. l. adler and c. schubert , phys .
lett . * 77 * , 1695 ( 1996 ) [ hep - th/9605035 ] .
d. denterria and g. g. da silveira , phys .
lett . * 111 * , 080405 ( 2013 ) [ arxiv:1305.7142 [ hep - ph ] ] .
j. l. hewett , f. j. petriello and t. g. rizzo , phys .
d * 64 * , 075012 ( 2001 ) [ hep - ph/0010354 ] .
r. karplus and m. neuman , phys .
80 * , 380 ( 1950 ) . v. b. berestetskii , e. m. lifshitz and l. p. pitaevskii , _ quantum electrodynamics _
( pergamon press , oxford , 1982 )
. s. villalba - chavez and a. e. shabad , phys .
d * 86 * , 105040 ( 2012 ) [ arxiv:1206.4491 [ hep - th ] ] .
f. r. klinkhamer and m. schreck , phys .
d * 78 * , 085026 ( 2008 ) [ arxiv:0809.3217 [ hep - ph ] ] .
g. rubtsov , p. satunin and s. sibiryakov , phys .
d * 86 * , 085012 ( 2012 ) [ arxiv:1204.5782 [ hep - ph ] ] . | in this work , we study the radiative generation of the lorentz - violating euler - heisenberg action , in the weak field approximation . for this , we first consider a nonperturbative calculation in the coefficient @xmath0 , however , by assuming rotational invariance . within this approach
, we also obtain the results of the amplitude for the photon triple splitting . in the following ,
we take into account the perturbative approach , where @xmath0 is treated as a insertion in the propagator and a new vertex .
the partial results are in fact an expansion up to first order in @xmath0 of the nonperturbative ones , with @xmath1 and @xmath2 .
this suggest that the complete results obtained in the nonperturbative approach can be used in both treatments . |
when a liquid or polymer melt is ( super)cooled towards the glass transition , its viscosity and relaxation time increase with many orders of magnitude over a relatively small temperature range .
more generally , the dynamics of a viscous liquid depends on two variables , density @xmath1 and temperature @xmath2 ( or pressure and temperature ) . understanding what exactly controls the viscous slowing down upon cooling and/or compression remains one of the main challenges related to the glass transition @xcite .
an indication that a single , underlying quantity determines the viscous slowing down of supercooled liquids was published in 1998 by tlle et al .
they showed that the dynamics of _ ortho_-terphenyl , measured at different densities and temperatures , collapses on a single curve when plotted against a function of density over temperature @xmath4 .
more specifically , these neutron scattering data were found to collapse for @xmath5 .
later , a similar scaling was found to work for other organic glass formers , including polymers , showing that the relaxation time is a function of @xmath4 @xcite .
there was some debate over the functional form of @xmath6 and whether it could be uniquely determined given the limited density changes experimentally available @xcite . in a famous review roland et al .
@xcite demonstrated that scaling with @xmath7 with a material specific scaling exponent @xmath3 works well for a large group of organic glass formers , including polymers .
we refer to this scaling as power - law density scaling . to date , many more molecular liquids have been shown to obey power - law density scaling to a good approximation , including polymers , but also ionic liquids @xcite and liquid crystals @xcite .
the recently developed isomorph theory@xcite explains and generalizes power - law density scaling .
the isomorph theory predicts that liquids which obey the theory have curves ( isomorphs ) in their phase diagrams along which structure and dynamics are invariant in the appropriate units .
the isomorphs are identified by @xmath4 being constant on an isomorph , where @xmath6 is a material specific function .
consequently relaxation times and transport coefficients are predicted @xcite to be functions of @xmath4 . for sufficiently small density changes
@xmath6 may be approximated by a power law : @xmath8 , which is equivalent to power law density scaling . in typical experiments
, it is possible to change density around 5% , but recently it has been shown in experiments that @xmath6 is not well approximated by a power law for larger density changes of up to 20% @xcite .
moreover , the theory provides an independent method of determining the scaling exponent @xmath3 for a small density range .
other predictions of the theory are that certain thermodynamical quantities including the excess entropy and isochoric specific heat are invariant on the isomorph .
since both excess entropy and the relaxation times are predicted to be constant on an isomorph , the isomorph theory provides an explanation for rosenfeld s excess entropy scaling @xcite , according to which a liquid s relaxation times and transport coefficients are functions of excess entropy only .
the isomorph theory has so far only been tested in detail for atomic systems @xcite , and for some small rigid molecules @xcite
. however , many organic glass formers are large molecules or have bulky side groups , because this makes it harder for the liquid to crystallize .
these larger molecules , and polymers in particular , inherently have intra molecular degrees of freedom that influence the liquid structure and dynamics . here
, we aim to bridge the gap between the simple models already shown to obey they isomorph theory , and larger flexible glass formers shown experimentally to obey power law density scaling . since
both alkanes @xcite and polymers @xcite have been shown to obey power - law density scaling , we simulated a general viscous model liquid of linear , flexible lennard - jones chains ( ljc ) .
the model has been used extensively for viscous polymer melts close to the glass transition @xcite .
we show that the ljc liquid has isomorphs in its phase diagram , and we study the effect of the intra molecular degrees of freedom on the applicability of the isomorph theory . in section [ sec : theory ]
we give a short overview over the relevant aspects of the isomorph theory .
we explain the ljc model in section [ sec : model ] and present the details of our simulation method .
we start our discussion of the results by showing how the isomorphs were obtained for the ljc model ( section [ ssec : making_isomorphs ] ) .
we then verify that the dynamics ( section [ ssec : dynamics ] ) and some aspects of the structure ( section [ ssec : structure ] ) are invariant on the isomorph . as predicted by the isomorph theory ,
we show in section [ ssec : scaling ] that isomorph scaling can be used to collapse the dynamics along different isochores onto a single master curve .
an isomorph is a curve in the phase diagram that consists of state points that are isomorphic to each other .
if one takes two state points with @xmath9 and @xmath10 , then pairs of microconfigurations exist with the same coordinates when scaled with density @xmath11 here @xmath12 denotes the coordinates of all particles .
two state points are now defined to be isomorphic if these pairs of microconfigurations have proportional boltzmann weights @xcite : @xmath13 with @xmath14 being a proportionality constant that is the same for all physically relevant pairs of microconfigurations , depending only on the two state points .
thus , if two state points are isomorphic , they have the same probability distributions of their reduced unit configurations . from this definition it can be shown that various dynamical and structural properties are invariant on an isomorph , as well as the excess entropy @xcite .
it should be noted that our model system is expected to only obey the isomorph definition approximately , since the rigid bonds in the molecule do not scale with density .
therefore , the equilibrium configurations at different densities in general are not the same .
the development of the isomorph theory was preceded by the discovery that some liquids have strong correlations in the equilibrium fluctuations of the configurational parts of their energy and pressure .
the correlations can be quantified by the standard correlation coefficient @xcite @xmath15 where @xmath16 is the potential energy , @xmath17 is the virial , @xmath18 denotes deviation from thermal average , and brackets @xmath19 denote average in the canonical ensemble . for liquids where the pair potential is an inverse power law ( ipl ) , @xmath20 , the correlation is perfect ( @xmath21 ) , but a large group of liquids have a correlation coefficient close to one , indicating strong correlation . liquids with a correlation coefficient larger than 0.9 were referred to as `` strongly correlating '' , but since this term was often confused with strongly correlated quantum systems , we now refer to this class of liquids as `` roskilde - simple '' liquids .
the standard linear regression slope @xmath22 of the fluctuations is given by @xmath23 it can be shown using the standard fluctuation formulae that this slope is equal to the logarithmic density derivative of the temperature on a curve of constant excess entropy @xmath24 , where @xmath25 is the entropy of an ideal gas at the same temperature and density @xcite @xmath26 this slope @xmath22 is equal to the density scaling exponent @xmath27 mentioned in the introduction , as long the change of density is small enough . one can use the `` slope '' @xmath22 calculated from the fluctuations to trace out a curve of constant excess entropy in the phase diagram .
first , one calculates @xmath22 at a certain state point ( 1 ) with temperature @xmath28 and density @xmath29 using eq .
( [ eq : gamma1 ] ) . if one then increments density by a sufficiently small amount to density @xmath30 , it is possible to calculate the temperature @xmath31 ( eq . ( [ eq : gamma2 ] ) ) that has the same excess entropy at this new density .
this can be done many times in an iterative fashion to obtain a set of state points that have the same excess entropy .
since @xmath22 may change with density , it is necessary to increment density by a sufficiently small amount ( e.g. , 2% ) , which can be checked by comparing the effect of a different density increment .
here we use this method to trace out curves in the phase diagram with invariant excess entropy and check if the predicted isomorphic invariance of other properties is fulfilled .
it should be noted that the invariance is only predicted to hold when quantities are considered in the appropriate reduced units , e.g. , using @xmath32 as the unit of length , and @xmath33 as the unit of energy @xcite .
we denote reduced units with a tilde .
the isomorph theory predicts `` isomorph scaling '' , i.e. , that the dynamics is a function of @xmath4 , where @xmath6 depends on the system @xcite . for atomic systems interacting via a pair potential that is the sum of ipl potentials @xmath34 , @xmath6 is given by @xmath35 , where the constants @xmath36 are the fractional contributions of each term to the heat capacity @xcite .
this includes for example the celebrated lennard - jones potential @xcite . for molecular liquids @xmath6
is not known analytically .
we performed molecular dynamics simulations of flexible lennard - jones chains ( ljcs ) consisting of 10 rigidly bonded segments .
segments in different molecules and non - bonded segments within a molecule interact via the standard lj potential , cutting and shifting the potential at @xmath37 .
we simulated 200 chains in a cubic bounding box with periodic boundary conditions in the nvt ensemble using a nos - hoover thermostat .
for the time step we used @xmath38 , and the time constant of the thermostat was 0.2 .
the simulations were performed with our rumd @xcite software utilizing state of the art gpu computing .
the model has been derived from a model by kremer and grest @xcite , who did not include the attractive part of the lj potential .
later , the attractive part has usually been included .
short lj chains of around ten segments have been used extensively to simulate glassy polymer melts @xcite , even though real polymers easily consist of thousands of monomers .
the reason for this is threefold .
firstly , the ljc is a coarse - grained model , meaning that a single lennard - jones particle may correspond to several monomers .
secondly , increasing the chain length in general increases the total system size , which in turn increases the simulation time .
most importantly , it is often the equilibrium ( viscous ) liquid that is of interest .
both increasing the chain length and approaching the glass transition increase the equilibration time , meaning that there is always a trade - off between chain length and viscosity @xcite .
often , the neighboring segments in the chain are bonded by a fene potential , although harmonic springs @xcite and rigid bonds @xcite have also been used . here
, the bond length @xmath39 was kept constant using the time symmetrical central difference algorithm @xcite . like other constraint algorithms ,
these bonds contribute to the virial @xcite : @xmath40 , but not to the energy . with our purpose in mind ,
the model is of special interest since it has already been shown to obey power - law density scaling , using @xmath22 as an empirical scaling parameter .
moreover , the ljc liquid has been shown to obey rosenfeld s excess entropy scaling @xcite .
to generate an isomorph , a nvt simulation was performed at a state point @xmath41 , and the scaling exponent @xmath22 was calculated using eq .
( [ eq : gamma1 ] ) .
we then changed density with 0.02 and used equation eq .
( [ eq : gamma2 ] ) to find the temperature at the new state point for which the excess entropy @xmath42 is the same .
applying this procedure iteratively we obtain a curve with constant @xmath42 .
if the model conforms to the isomorph theory , this curve will be an isomorph , i.e. , have invariant dynamics and structure in reduced units .
five prospective isomorphs were generated using this procedure with @xmath43 and @xmath44 , calculated from the instantaneous values of the virial @xmath17 and the potential energy @xmath16 using eq .
( [ eq : r ] ) .
each data set corresponds to an isomorph , obtained as described in the text ( see fig . [
fig : rhot ] for the corresponding temperatures ) .
the correlation coefficient is high , albeit lower than for the single component lennard - jones liquid @xcite .
( b ) the isomorphic scaling exponent @xmath22 as defined by eq .
( [ eq : gamma1 ] ) .
the values found are significantly higher than for the single component lennard - jones liquids @xcite , and show a clear density dependence .
the logarithmic derivatives of @xmath22 on the isochore and isotherm confirm that @xmath22 is much more dependent on the density than on temperature : @xmath45 and @xmath46 , as predicted by the isomorph theory . ] in fig .
[ fig : rgamma](a ) , the correlation coefficient @xmath47 is plotted as a function of density for the five isomorphs . for the densities we simulated
, the correlation coefficient varies between 0.81 and 0.87 , which is lower then the ( somewhat arbitrary ) 0.9 limit for simple liquids .
however , we show with this paper that the ljc model has clear isomorphs in its phase diagram . in fig .
[ fig : rgamma](b ) we plot the values of @xmath22 calculated from eq .
( [ eq : gamma1 ] ) .
the isomorph theory predicts @xmath22 to depend on density but not temperature@xcite .
this is seen to be fulfilled to a good approximation ; @xmath22 changes much more by increasing density by 25% than by increasing temperature by 60% .
the density dependence of @xmath22 means that we can only use eq .
( [ eq : gamma2 ] ) for small density changes , which indicates that simple power - law density scaling is an approximation that only works for small density changes .
the @xmath22 values found for the ljc model ( 6.17.9 ) are higher than for a single component lj liquid ( 5.36.7 ) @xcite .
this increase in @xmath22 is due to the fixed constraints , which can be seen as a very steep repulsion between bonded segments . on the other hand ,
the high @xmath22 values is in contrast to the values found from power - law density scaling , which in experiments are generally lower for polymers than for small molecular liquids @xcite .
tsolou et al @xcite found @xmath48 from power - law density scaling of simulation data of a united atom model of _ cis_-1,4-polybutadiene .
a possible explanation for this low value of @xmath22 has been given by xu @xcite who showed using the generalized entropy theory that polymer rigidity significantly decreases the density scaling exponent @xmath22 .
xu quantified polymer rigidity y the bending energy of the angle between two bonds . , as well as the normalized orientational autocorrelation function of the end - to - end vector @xmath49 .
we used @xmath50 ( the position of the first peak of the segmental structure factor ) .
( a ) the data for 7 isomorphic state points collapse on a single master curve when plotted in reduced units , and this is the case for all three relaxation functions .
( b ) for isothermal state points , the curves do not collapse but are spread over a larger dynamical range.,title="fig : " ] + , as well as the normalized orientational autocorrelation function of the end - to - end vector @xmath49 .
we used @xmath50 ( the position of the first peak of the segmental structure factor ) .
( a ) the data for 7 isomorphic state points collapse on a single master curve when plotted in reduced units , and this is the case for all three relaxation functions .
( b ) for isothermal state points , the curves do not collapse but are spread over a larger dynamical range.,title="fig : " ] in the following , we test a number of isomorph predictions focusing on the @xmath51 isomorph , before returning to the question of the overall scaling properties of the model .
the isomorph theory predicts dynamics and structure to be invariant on an isomorph .
this invariance applies to data in reduced units , which means that distance and time are scaled using @xmath52 and @xmath53 , where @xmath54 is the mass of a segment .
the dynamics are of particular interest here , because the dependence on state point becomes large upon cooling and/or compression . in fig .
[ fig : dynamics](a ) , different dynamical quantities are plotted .
the self part of the segmental and the center of mass intermediate scattering function @xmath55 , as well as the normalized orientational autocorrelation of the end - to - end vector @xmath56 are plotted as a function of reduced time .
the values of @xmath57 were kept constant in reduced units : @xmath58 ( @xmath59 ) .
all these measures of the dynamics collapse well for the isomorphic state points compared to an isothermal density change ; increasing the density by 11% while keeping temperature constant significantly changes the dynamics , whereas increasing the density 25% while following the isomorph keeps the dynamics invariant . the data in fig .
[ fig : dynamics](a ) are in agreement with power law density scaling of segmental and chain relaxation times of simulated polybutadieen @xcite .
our data extend these results by showing that the shape of the entire relaxation curves is invariant .
vector has been varied to obtain different measures of the relaxation time .
each value was kept constant in reduced units for the different densities all relaxation time measures are invariant for isomorphic state points
( filled symbols ) .
an isotherm is included for comparison ( open red symbols ) . ]
we define a relaxation time for the dynamical quantities as the time where the correlation function reaches 0.2 .
these relaxation times are plotted in fig .
[ fig : tau ] , this time also varying @xmath60 .
the different relaxation times characterizing the dynamics covers more than 4 decades in time , but each of them are to a good approximation invariant on the isomorph .
in contrast , the relaxation times on the isotherm shown ( open red symbols ) shows a clear dependence on density .
the collapse of the rouse modes is good , especially for the lower modes .
( b ) data for the same isothermal state points as in fig .
[ fig : dynamics](b ) .
there is no collapse of the dynamics for isothermal state points.,title="fig : " ] + ( a ) .
the collapse of the rouse modes is good , especially for the lower modes .
( b ) data for the same isothermal state points as in fig . [
fig : dynamics](b ) .
there is no collapse of the dynamics for isothermal state points.,title="fig : " ] the dynamics of flexible chains are often expressed in terms of correlation functions of rouse modes , @xmath61 @xcite .
the zeroth mode @xmath62 describes the position of the center of mass of the chain , while the higher modes with @xmath63 describe the local motion of a subchain of @xmath64 segments . in fig .
[ fig : rouse ] some of the rouse mode auto correlation functions are plotted for the isomorphic state points . for the lower modes , there is an excellent collapse of the correlation functions , whereas the invariance decreases somewhat for the higher modes .
the variance of the highest modes is somewhat surprising considering that the segmental intermediate scattering function shows such a good collapse .
it should however be noted that the amplitude of the rouse modes is predicted to scale as @xmath65 , so the contribution of the higher modes is very small @xcite .
moreover , the @xmath66 rouse modes represent the conformation of the ( sub)chain , and the less than perfect collapse of the highest modes thus indicates that the deviation from isomorph theory is specific to the local intramolecular dynamics .
it is well known that reducing the local intramolecular degrees of freedom by including bond and torsional potentials leads to dynamics that are less rouse - like @xcite .
the local degrees of freedom affect mostly the higher modes , giving the standard rouse behavior for the lower modes representing longer subchains @xcite + fig . [
fig : msd ] shows the isomorphic invariance of the mean square displacement of both the segments and the center of mass in all regime , including the subdiffusive regimes which is specific for polymers and other flexible molecules . .
the inset shows the direction of the jumps in the phase diagram , plotted in the @xmath67-plane .
black line : a jump between isomorphic state points .
the energy shows no relaxation since the system is immediately in equilibrium .
red and green lines : two jumps from the same isomorph to another isomorph show the same relaxation behavior .
the data of the relaxation plots are averages of 8 independent starting configurations . ]
not only equilibrium dynamics , but also out of equilibrium dynamics is predicted to be invariant on an isomorph .
we test this by changing density and temperature instantaneously during a simulation .
the center of mass positions of the molecules are scaled together with the box , but the intramolecular distances were kept constant . in fig .
[ fig : jump ] the relaxation of the potential energy is plotted after different instantaneous jumps .
although the energy is not the same at two isomorphic state points , no relaxation is visible in the energy when jumping between two the to state points ( black line ) .
this is predicted by the isomorph theory : two state points on the same isomorph are equivalent with regard to aging @xcite .
likewise , when jumping from two state points on the same isomorph to a third state point that is not on that isomorph , the relaxation curve is the same for the two jumps .
when the density is changed , the system is immediately in equilibrium at the isomorphic state point with the new density .
any relaxation after the density jump then takes place on the isochore @xcite . .
dashed lines correspond to isothermal density changes and are included for comparison .
the data show that intermolecular structure is invariant , while intramolecular structure is not invariant on the isomorph due to the constant bond length .
( a ) the intermolecular ( segmental ) radial distribution function @xmath68 in reduced units on the isomorph @xmath69 .
the intermolecular @xmath70 is to a good degree invariant for isomorph state points , especially when compared to a ( small ) density change on an isotherm .
( b ) the intramolecular @xmath70 is clearly not invariant on an isomorph . ]
also the structure is predicted to be invariant on an isomorph @xcite .
however , not all structural quantities are necessarily equally invariant when molecular liquids are considered . since the length of the rigid bonds is constant in normal units and does not change with density , the bond length in reduced units will not be constant on the isomorph in reduced units .
for that reason we plot the inter- and intramolecular contribution to the segmental radial distribution function @xmath70 separately in fig .
[ fig : rdf ] .
the intermolecular structure is quite constant on the isomorph , while the intramolecular structure is clearly not .
the center of mass @xmath68 was also found to be invariant on the isomorph when plotted in reduced units ( data not shown ) , but it is also invariant on the isochore and isotherm within the liquid ( fluid ) phase . and the mean square radius of gyration ( b ) as a function of density .
the temperature dependence of these quantities is similar on the isomorph and isotherms ( dashed lines ) .
it should be mentioned that when these quantities are plotted in real units , they show an ( intuitive ) decrease with density .
( c ) and ( d ) the same data for the isomorphic state points , now plotted as a function of temperature and compared with an isochore .
these intramolecular quantities are actually more constant on the isochore . ] to investigate the difference in inter- and intramolecular structure further , we plot the mean square radius of gyration @xmath71 and the mean square end - to - end vector @xmath72 in fig . [
fig : gyradius ] .
these intramolecular quantities are clearly not invariant on the isomorph , changing as much with density as on the isotherm . on an isochore these quantities are even more constant than on the isomorph .
the lack of temperature dependence of these quantities was already noted for a similar bead - spring model @xcite . ,
open symbols : same data with temperatures divided by @xmath73 , showing a good collapse as predicted by the isomorph theory . dashed lines :
the function @xmath74 was found by fitting to the open symbols ( see text ) . ] finally , we return to the question of the overall scaling of the dynamics of the model . as mentioned in the introduction , the isomorph theory predicts that each relaxation time characterizing the dynamics is a function of @xmath4 where @xmath6 is system dependent function . for atomic systems with pair potentials being sums of power laws
, we have an analytical expression for @xmath6 @xcite . due to the presence of the bonds ,
we unfortunately do _ not _ have an analytical expression for @xmath6 in the model studied here .
[ fig : rhot ] shows the five studied isomorphs in the @xmath75 plane ( filled symbols ) .
the open symbols show the same data , except that the temperatures are divided by @xmath73 ( the temperature at @xmath76 ) .
the scaled data is predicted to collapse on a single curve , @xmath6 , which is indeed seen to be the case .
we have found that the @xmath6 from the single component lennard - jones liquid @xcite does not describe the shape of lennard - jones chain isomorphs correctly due to the rigid bonds ( fit not shown ) .
instead , we have fitted the shape of the isomorphs with a function of the form @xmath77 where @xmath78 and @xmath79 are fitting parameters .
the choice of the functional form is rather arbitrary ; it was found to fit the data well with only two fitting parameters , but there is no a priori reason why @xmath6 should be a sum of two power laws and we do not ascribe any meaning to this functional form .
nonetheless , the shape of the isomorphs is well described by @xmath80 shown as the dashed pink line in fig .
[ fig : rhot ] . ) .
( a ) and ( b ) the power - law density scaling approach for two different values of @xmath22 ( 7.7 and 6.7 ) , collapsing the low and high density isochores respectively .
neither value gives a good collapse of all the data .
( c ) isomorph scaling approach , using the function @xmath81 ( see fig . [
fig : rhot ] ) to scale the relaxation times , giving a much better collapse . ] fig .
[ fig : scaling ] compares for three isochores the power - law density scaling and the scaling predicted by the isomorph theory .
[ fig : scaling](a ) and fig .
[ fig : scaling](b ) show that the two smallest densities collapse using power - law density scaling with @xmath82 , whereas the two highest densities collapse using @xmath83 .
notice that the values of @xmath22 found by this empirical scaling is consistent with the values found from the @xmath84 fluctuations in the respective density intervals ( see fig .
[ fig : rgamma ] ) .
the power - law density scaling is an approximation that works well for ( relatively ) small density changes , and the scaling exponent @xmath22 can be determined independently from the @xmath84-fluctuations .
the more general form of scaling is the one predicted by the isomorph theory , which is tested in fig .
[ fig : scaling](c ) , using the @xmath6 determined empirically in fig .
[ fig : rhot ] .
the collapse is seen to be excellent .
notice that the isomorph scaling also captures the different shapes of the segmental and chain dynamics , which is also well known for power - law density scaling in a small density range @xcite .
to summarize , we have shown that the predictions of the isomorph theory apply to a flexible chain - like model system , despite the fact that the system is not entirely `` roskilde - simple '' because the correlation coefficient of the instantaneous @xmath67 fluctuations is less than 0.9 . however , the collapse of the dynamics at different time and length scales is unmistakable , and works for the segmental dynamics as well as the chain dynamics .
we see a slight deviation from invariance for the highest rouse modes , but we attribute this to a specific intramolecular effect related to the ( local ) conformation of the chain . the rigid bonds in the model can not scale with density and the structure can therefore not be constant on the isomorphs .
we have shown that this is only the case for intramolecular structure , while the intermolecular structure stays invariant on the isomorph .
our results indicate that the isomorph theory may be extended to include flexible molecules . in particular this explains the experimentally observed power - law density scaling for alkanes and many polymers - and predicts that it should break down at larger density variations where isomorph scaling is needed . | the isomorph theory provides an explanation for the so - called power law density scaling which has been observed in many molecular and polymeric glass formers , both experimentally and in simulations .
power law density scaling ( relaxation times and transport coefficients being functions of @xmath0 , where @xmath1 is density , @xmath2 is temperature , and @xmath3 is a material specific scaling exponent ) is an approximation to a more general scaling predicted by the isomorph theory .
furthermore , the isomorph theory provides an explanation for rosenfeld scaling ( relaxation times and transport coefficients being functions of excess entropy ) which has been observed in simulations of both molecular and polymeric systems . doing molecular dynamics simulations of flexible lennard - jones chains ( ljc ) with rigid bonds , we here provide the first detailed test of the isomorph theory applied to flexible chain molecules .
we confirm the existence of isomorphs , which are curves in the phase diagram along which the dynamics is invariant in the appropriate reduced units .
this holds not only for the relaxation times but also for the full time dependence of the dynamics , including chain specific dynamics such as the end - to - end vector autocorrelation function and the relaxation of the rouse modes .
as predicted by the isomorph theory , jumps between different state points on the same isomorph happen instantaneously without any slow relaxation . since the ljc is a simple coarse - grained model for alkanes and polymers ,
our results provide a possible explanation for why power - law density scaling is observed experimentally in alkanes and many polymeric systems .
the theory provides an independent method of determining the scaling exponent , which is usually treated as a empirical scaling parameter . |
the problem of bacterial photosynthesis has attracted a lot of recent interest since the structures of the photosynthetic reaction center ( rc ) in purple bacteria _
rhodopseudomonas viridis _ and _ rhodobacterias sphaeroides _ have been determined@xcite .
much research effort is now focused on understanding the relationship between the function of the rc and its structure .
one fundamental theoretical question concerns the actual mechanism of the primary electron transfer ( et ) process in the rc , and two possible mechanisms ( or a combination of both ) have emerged out of the recent theoretical work @xcite .
the first is an incoherent two - step mechanism where the charge separation involves a sequential transfer from the excited special pair ( p@xmath0 ) via an intermediate bacteriochlorophyll monomer ( b ) to the bacteriopheophytin ( h ) .
the other is a coherent one - step superexchange mechanism , with p@xmath1b@xmath2 acting only as a virtual intermediate . from experimental studies to date
, one can not easily rule out either possibility .
in fact , conflicting interpretations have been derived from different experiments .
the detection of transient population in the p@xmath1b@xmath2 state has been considered by many to be the key experimental evidence that can differentiate between the two mechanisms .
many transient absorption spectroscopic experiments @xcite could not detect bleaching in the b@xmath2 band , lending support to the coherent superexchange mechanism . on the other hand ,
some other experiments @xcite point toward a two - step process since a fast second rate constant corresponding to the p@xmath1b@xmath2h @xmath3 p@xmath1bh@xmath2 transition was detected .
the picture emerging from a comparison of the molecular dynamics ( md ) simulations for the rc carried out so far @xcite is equally murky .
different groups arrive at conflicting conclusions . despite the emphasis placed on the experimental detection of population in the p@xmath1b@xmath2 band , we have shown in a recent real - time path integral simulation@xcite that observation of transient population in p@xmath1b@xmath2 can not definitively rule out either mechanism .
in fact , significant population in the p@xmath1b@xmath2 state ( up to 20% ) can be observed in both the sequential and the superexchange regimes in the simulations .
this suggests that transient population on the p@xmath1b@xmath2 state , even if detectable , may not really be useful for differentiating between the two mechanisms@xcite .
in fact , almost all experimentally observed dynamical features of the rc can be reproduced nicely in either regime . in this letter , we report a detailed path integral analysis of the mechanism of the primary charge separation process .
we have computed the dynamics of the primary et ( i.e. , the time - dependence of the occupation probabilities on the three relevant chromophores ) as a function of the p@xmath1b@xmath2 energy for many combinations of the electronic coupling and the reorganization energy .
the primary objective is to determine the parameter values that are required for the dynamics to be consistent with characteristic experimental observations , and hence to provide a map of the various possible parameter regions in which the rc could operate .
our previous simulations have been carried out for a relatively small energy range for the p@xmath1b@xmath2 state ( between @xmath4 @xmath5 and @xmath6 @xmath5 relative to the p@xmath0 state ) . in this letter , we present new results for a much larger range of energies , up to 3000 @xmath5 above and 1333 @xmath5 below the p@xmath0 state .
this spans the region usually considered to be relevant for the rc , except for the region with extremely high p@xmath1b@xmath2 energy ( a value of @xmath7 @xmath5 has been suggested in ref.@xcite ) .
the model for our studies has been discussed previously in great detail @xcite . for the sake of completeness
, we will summarize some of the relevant features here .
for a model of the rc , we consider hamiltonians of the form @xmath8 where @xmath9 describes the bare three - state system , @xmath10 corresponds to a gaussian bath describing the protein environment , and @xmath11 is a bilinear system - environment coupling describing the interaction of the electric dipole moment with the bath polarization .
the states 1 , 2 , and 3 correspond to the electronic configurations p@xmath0bh , p@xmath1b@xmath2h and p@xmath1bh@xmath2 , respectively , and we allow for arbitrary binding energies of these states . the bare three - state hamiltonian is @xmath12 where the electronic coupling ( tunnel matrix element ) between states @xmath13 and @xmath14 is denoted by @xmath15 .
this spin - boson model is an idealized model for the rc .
its limitations have been discussed in ref.@xcite .
the most significant aspect of this model , however , is its ability to capture the effects of the low - frequency protein modes on the tunneling electron from first principles@xcite . in this sense ,
the protein casing in the rc presents itself to the electron as a gaussian fluctuating polarization field with a continuous spectral density .
previous md simulations have shown that this spectral density is smooth , featureless , and of ohmic form with a characteristic frequency @xmath16 of about 160 @xmath5 .
they have also confirmed the assumption of gaussian statistics for the bath polarization since the free energy curves were always found to be strictly parabolic . in the classical ( high temperature ) limit
, the spectral density has a simple relationship with the familiar reorganization energy 13 between states 1 and 3 , and one can characterize the ohmic spectral density completely by the two parameters @xmath16 and 13 .
we fix a number of parameters in our model according to reasonable experimental and theoretical estimates@xcite : 1 . with the energy scale fixed by setting 1 = 0
, the energy of the p@xmath1h@xmath2 state 3 is set at @xmath17 @xmath5 .
the bath frequency @xmath16 is set at 166 @xmath5 according to md results .
the ratio 23/12 is set at 4 , and 13 is assumed to be negligible .
all electronic couplings are assumed to be independent of the actual protein configuration ( condon approximation ) .
the technical details of our real - time quantum monte carlo ( qmc ) simulations have been reported in great detail elsewhere @xcite and will not be repeated here .
this numerical technique allows for a computation of the time - dependent occupation probabilities on the electronic states up to a certain time limit .
qmc simulations can not be carried out to very long times due to the well - known dynamical sign problem .
however , the simulations reported here have been carried out up to a few picoseconds , the same timescale as the experimentally determined et rate . in these simulations
, we search for parameter regions in which the charge separation proceeds in qualitative agreement with experimental observations .
we consider the following to be key experimental characteristics : ( 1 ) the p@xmath1b@xmath2 population ( 2 ) should be small throughout ( 20% ) ; ( 2 ) the charge separation rate should be about 3 ps at room temperature ; and ( 3 ) the rate should increase about twofold at cryogenic temperatures . for each value of the p@xmath1b@xmath2 state energy 2 , we attempt to find a combination of 12 and 13 that would yield dynamics with these experimental characteristics .
from the qmc simulations , we have compiled a `` phase diagram '' for the rc , which is a map of the various regions in parameter space which generate dynamics with all the key experimental characteristics . this phase diagram is shown in fig . 1 , with the proper values of 12 correlated with the energy of the p@xmath1b@xmath2 state 2 .
on the phase diagram , we have also indicated the values of 12 predicted by conventional nonadiabatic theory to give a room - temperature et rate of 3 ps ( without demanding that they also conform to the inverse temperature dependence ) . from the phase diagram
, there are clearly two distinct regions which yield charge separation dynamics with the correct experimental characteristics .
the first is characterized by a p@xmath1b@xmath2 state energy 2 lower than about @xmath18 @xmath5 and down to about @xmath19 @xmath5 . within this region ,
the dynamics largely agrees with predictions from conventional nonadiabatic theory for the sequential mechanism .
simulation results for the transient populations on the three electronic states are shown in fig . 2 at two temperatures for 2 = @xmath4 @xmath5 and @xmath20 @xmath5 .
the 1 @xmath3 3 transfer is efficient in both cases , with 2 remaining below 15% throughout . for 2 = @xmath4
@xmath5 , the charge separation rate increases from 3.0 ps at 298 k to 2.0 ps at 140 k. this should be contrasted with 2 = @xmath20 @xmath5 , for which the rate increases from 3.3 ps at 298 k to only 2.7 ps at 140 k. although the et rate still increases with lower temperature for 2 = @xmath20 @xmath5 , this temperature - dependence is somewhat weaker than that observed in experiments . for 2 much below @xmath20
@xmath5 , the temperature dependence is not strong enough to be consistent with experiments . therefore , in the phase diagram , the sequential region terminates around 2 = @xmath20 @xmath5 .
the second region is characterized by a p@xmath1b@xmath2 state energy 2 higher than @xmath6 @xmath5 ( we have performed simulations for 2 up to @xmath21 @xmath5 ) . for 2 = @xmath6
@xmath5 and @xmath22 @xmath5 , conventional nonadiabatic superexchange theory does _ not _ give reliable predictions for the dynamics . the transfer mechanism in this region
is predominantly superexchange , although the precise rates are not well described by nonadiabatic theory . as the energy of the p@xmath1b@xmath2 state moves up to 2 = @xmath23 @xmath5 and beyond , the dynamics becomes increasingly nonadiabatic .
we mention in passing that in this limit the dynamical sign problem is quite severe , and one needs to sample extremely long monte carlo trajectories rendering qmc simulations for 2 much higher than @xmath21 @xmath5 very costly .
nevertheless , our present data show that for sufficiently high - lying p@xmath1b@xmath2 state , the conventional golden rule superexchange formula becomes increasingly accurate , making qmc simulations less valuable for 2 @xmath24 @xmath21 @xmath5 . for 2 @xmath25 @xmath5
, however , nonadiabatic theory seems to severely underestimate the magnitude of the electronic couplings required to achieve superexchange rates consistent with experiments .
simulation results for the transient populations on the three electronic states are shown in fig . 3 at two temperatures for 2 = @xmath6 @xmath5 , @xmath22 @xmath5 and @xmath23 @xmath5 .
for all three energies , the charge separation rates at 298 k are about 3 ps . these rates increase to about 2.2 to 2.5 ps at 140 k. as was pointed out in our previous work @xcite , the dynamics in this regime is not simply monoexponential .
there is a fast short - time component for the first @xmath26 ps , followed by a slower component .
the fast component always obeys the experimentally observed inverse temperature dependence , but the rate and the proportion of the slower component vary nonmonotonically with temperature , as measured in many mutant rcs @xcite .
the physical reason for this phenomenon has been given by gehlen et al.@xcite , and our simulations provide direct numerical support for their arguments . the phase diagram in fig .
1 also reveals another interesting aspect of the superexchange regime : the rc seems to be able to operate under a wide range of p@xmath1b@xmath2 state energies with the same conditions , i.e. , with approximately the same electronic coupling 12 ( about 140 to 150 @xmath5 ) and the same reorganization energy 13 ( about 2000 @xmath5 ) .
this means that within the superexchange regime , the system is robust against variation in the p@xmath1b@xmath2 state energy .
the sequential region is not as robust .
although approximately the same value of 12 ( about 28 @xmath5 ) would work for 2 between about @xmath20
@xmath5 and @xmath4 @xmath5 , the optimal value of the reorganization energy 13 changes from 4000 @xmath5 at 2 = @xmath4 @xmath5 to 5300 @xmath5 at 2 = @xmath20 @xmath5 . finally , it is important to stress that the rc does _ not _ seem to operate via a combined sequential / superexchange mechanism as proposed in refs.@xcite .
our simulations indicate clearly that it is either a pure sequential or a pure superexchange transfer .
this is indicated by the grey region in fig . 1 , which we refer to as the `` gap '' region .
we emphasize that this does not imply that in this region it is impossible to achieve a fast et rate . on the contrary , it is possible to achieve a 3 ps transfer at room temperature with only minor population accumulation on the p@xmath1b@xmath2 state .
however , we have not found any parameter set in this region that can simultaneously satisfy all the experimentally observed characteristics _ both _ at high and low temperatures @xcite .
by invoking a simple and reasonable approximation for the formally exact path - integral expressions , we have previously derived a set of nonlocal master equations for the time - dependent occupation probabilities @xcite .
our `` noninteracting - cluster approximation '' ( nica ) expresses the time - dependent transition rates between the various sites in terms of cluster functions @xmath27 .
these cluster functions give the amplitude of all paths going from diagonal state @xmath13 to diagonal state @xmath14 of the reduced density matrix without touching the diagonal in between .
since each hop on the @xmath28 lattice representing the possible states of the reduced density matrix gives a factor @xmath29 , the rates can be expressed as a power series in the electronic couplings . by confining this expansion to the lowest - order terms and taking the long - time limit ( which makes the master equations local in time ) ,
one obtains the nonadiabatic rate expressions referred to in sec .
iii . the rates @xmath30 and @xmath31 ( plus backward rates ) describing a sequential transfer
are just golden rule rates .
furthermore , the lowest - order cluster @xmath32 gives the nonadiabatic superexchange rate .
there is only one pathway contributing to this fourth - order cluster ( plus the complex conjugate ) , namely the one going along the edge of the @xmath28 lattice .
the resulting rate formula can be further simplified in the classical limit by taking a short - time approximation for the bath kernel , or for a high - lying state where one can derive the conventional golden rule rate for the emerging two - state system spanned by states 1 and 3 @xcite .
these golden rule rates are consistent with the qmc results for very high p@xmath1b@xmath2 state energy ( @xmath33 @xmath5 ) as well as for low - lying p@xmath1b@xmath2 states ( @xmath34 @xmath5 ) .
but in the intermediate region @xmath6 @xmath52 @xmath23 @xmath5 , nonadiabatic theory breaks down ( see fig . 1 ) .
we will now study the origin of the apparent discrepancy between nonadiabatic theory and the qmc results in this intermediate region . to that purpose ,
we have computed the next - order corrections to the superexchange cluster , @xmath35 .
here one has to consider the two pathways with six hops which go from diagonal state 1 to state 3 without hitting the diagonal . in the classical limit
, we obtain for the activationless situation considered in the simulations ( @xmath36 ) @xmath37 where @xmath38 and @xmath39 . for a high - lying intermediate state ,
the bracket gives unity because @xmath40 .
therefore the sixth - order correction decreases the nonadiabatic estimate by @xmath41 for @xmath42 @xmath5 . with the short - time bath kernel
, it would actually be possible to sum the whole power series and thus to obtain the full classical nica superexchange rate without stopping at sixth order .
however , we have not done so because this short - time approximation is inappropriate for high - order diagrams . in any case
, it is expected that these higher - order contributions lead to a _ reduction _ of the nonadiabatic superexchange rate .
moreover , we also expect that adiabatic corrections beyond nica can lead to substantial renormalizations for large electronic couplings . to conclude , we have given a path - integral analysis of the primary electron transfer step in bacterial photosynthesis .
the mechanism appears to be either a sequential or a superexchange transfer but not a combined one . whereas conventional nonadiabatic theory appears to be accurate for both the sequential and the ultimate superexchange regime , adiabatic corrections
reduce the superexchange rate substantially for intermediate energies of the p@xmath1b@xmath2 state .
we wish to thank david chandler and uli weiss for illuminating discussions .
this work was supported in part by the national science foundation ( che-9216221 ) and the nsf young investigator awards program ( che-9257094 ) , the camille and henry dreyfus foundation under the camille teacher - scholar award program , and the sloan foundation through a alfred p. sloan fellowship .
computational resources from ibm are acknowledged .
j. deisenhofer , o. epp , k. miki , r. huber and h. michel , _
_ 180 ( 1984 ) 385 ; j.p .
allen , g. feher , t.o .
yeates , h. komiya and d.c .
rees , _ proc .
usa _ 84 ( 1987 ) 5730 ; c.h .
chang , d.m .
tiede , j. tang , u. smith , j.r . norris and m. schiffer , _ febs lett .
_ 205 ( 1986 ) 82 .
marcus , _ chem
_ 133 ( 1987 ) 471 ; _ ibid . _ 146 ( 1988 ) 13 .
m. bixon , j. jortner , m.e .
michel - beyerle , a. ogrodnik and w. lersch , _ chem
_ 140 ( 1987 ) 626
. m. bixon , j. jortner and m.e .
michel - beyerle , _ biochim .
acta _ 1056 ( 1991 ) 301 .
marcus and r. almeida _
_ 94 ( 1990 ) 2973 ; r. almeida and r.a .
marcus , _ ibid .
_ 94 ( 1990 ) 2978 .
y. hu and s. mukamel , _ j. chem .
_ 91 ( 1989 ) 6973 .
joseph and w. bialek , _
_ 97 ( 1993 ) 3255 ; _ biophys . j. _ 63 ( 1992 ) 397 .
r. egger , c.h .
mak and u. weiss , _ phys .
50 ( 1994 ) r655 .
martin , j. breton , a.j .
hoff , a. migus and a. antonetti , _ proc .
usa _ 83 ( 1986 ) 957 ; j. breton , j.l .
martin , g.r .
fleming and j.c .
lambry , _ biochemistry _ 27 ( 1988 ) 8276 . c. kirmaier and d. holten , _ proc .
usa _ 87 ( 1990 ) 3552 ; _ biochemistry _ 30 ( 1991 ) 611 .
w. holzapfel , u. finkele , w. kaiser , d. oesterhelt , h. scheer , h.u .
stilz and w. zinth , _ chem .
lett _ 160 ( 1989 ) 1 ; _ proc .
usa _ 87 ( 1990 ) 5168
. k. dressler , e. umlauf , s. schmidt , p. hamm , w. zinth , s. buchanan and h. michel , _ chem .
_ 183 ( 1991 ) 270
. m. marchi , j.n .
gehlen , d. chandler and m. newton , _ j. am . chem . soc .
_ 115 ( 1993 ) 4178 .
s.creighton , j.k .
hwang , a. warshel , w.w .
parson and j. norris , _ biochemistry _ 27 ( 1988 ) 774 ; a. warshel and w.w .
parson , _ ann .
_ 42 ( 1991 ) 279 .
k. schulten and m. tesch , _ chem . phys .
_ 158 ( 1991 ) 421 .
r. egger and c.h .
mak , _ j. phys
_ 98 ( 1994 ) 9903 .
the qmc algorithm is described in more detail in : r. egger and c.h .
mak , _ phys .
b _ ( in press ) .
a.j . leggett , s. chakravarty , a.t .
dorsey , m.p.a .
fisher , a. garg and w. zwerger , _ rev .
_ 59 ( 1987 ) 1 ; u. weiss , _ quantum dissipative systems _ ( world scientific , singapore , 1993 ) .
m. du , s.j .
rosenthal , x. xie , t.j .
dimagno , m. schmidt , d.k .
hanson , m. schiffer , j.r .
norris , and g.r . fleming , _ proc .
usa _ 89 ( 1992 ) 8517 ; y. jai , t.j .
dimagno , c.k .
chan , z. wang , m. du , d.k .
hanson , m. schiffer , j.r .
norris , g.r .
fleming and m.s .
_ 97 ( 1993 ) 13180 .
gehlen , m. marchi and d. chandler , _ science _ 263 ( 1994 ) 499 .
chan , t.j .
dimagno , l.x.q .
chen , j.r .
norris and g.r .
fleming , _ proc .
usa _ 88 ( 1991 ) 11202 . | we present a detailed analysis of the mechanism of the primary charge separation process in bacterial photosynthesis using real - time path integrals . direct computer simulations as well as
an approximate analytical theory have been employed to map out the dynamics of the charge separation process in many regions of the parameter space relevant to bacterial photosynthesis .
two distinct parameter regions , one characteristic of sequential transfer and the other characteristic of superexchange , have been found to yield charge separation dynamics in agreement with experiments .
nonadiabatic theory provides accurate rate estimates for low - lying and very high - lying bacteriochlorophyll state energies , but it breaks down in between these two regimes . # 1 # 1#2#3#4 = # 2 # 1#2#3#4 = # 2 = to # 2 rotate |
during the last decades our picture of galaxies being welteninseln ( e.g. kant , 1755 ) , i.e. island universes , has completely changed . at the beginning of extragalactic research in the 1920s galaxies
were thought to be in general islands , i.e. isolated stellar systems , with just a few obvious exceptions like the m51 system .
the success of the hubble classification could have been a reason to discard systems not fitting into it .
thus less regular systems have been neglected by definition as peculiar .
the interest in these systems raised strongly when new catalogues by vorontsov
velyaminov ( 1959 ) and arp ( 1966 ) became available .
they demonstrated that even the peculiar objects have a lot of common features like bridges , tails , or rings .
additionally , better observational instruments allowing for deeper images and new wavelength ranges revealed more complex structures . thus , a lot of galaxies formerly classified as non interacting may have had some gravitational interaction in the past which is still reflected in e.g. warps , flares , or thick discs .
especially useful are hi observations which can cover a much larger radial extension than the optical images .
therefore , hi is a very good tracer for tidal interactions . the theoretical understanding of interacting galaxies suffered for a long time from the lack of computational power allowing for a numerical solution of the gravitational n - body problem .
after a remarkable treatment by holmberg ( 1941 ) who built an analogue computer ( consisting of light bulbs and photo cells ) for determining the gravitational force , it took 20 years until n - body simulations were performed on a general purpose computer ( pfleiderer & siedentopf , 1961 ) : the basic idea of these _ restricted n - body _
simulations was the assumption that the potential in interacting galaxies can adequately be modeled by two particles which represent the galactic masses and move under their mutual gravitation , i.e. on keplerian orbits . with these assumptions
all the other particles are just test particles , and the complete n - body problem has been reduced to @xmath3 single body problems for a time
dependent potential . in a remarkable series of simulations
toomre & toomre ( 1972 ) applied this technique in order to determine the parameters of a couple of well studied interacting systems like arp 295 , m51 + ngc5195 , or ngc4038/39 .
motivated by these simulations and by a remarkable coincidence of the estimated fraction of strong interactions with the fraction of elliptical and s0 galaxies toomre ( 1977 ) put forward the idea of a merger origin of elliptical galaxies .
hence , a whole class of galaxies would have been formed mainly by strong interactions . however , restricted n - body simulations were not able to check this idea , because they do not include processes like dynamical friction or transfer of orbital angular momentum of the galaxies into galactic spin in a self consistent manner .
moreover , the assumption of unaffected point like galaxies becomes rather poor for merging systems .
thus , computationally much more expensive self consistent calculations were required .
already first models using @xmath4 particles proved the idea of merging ( white , 1978 ; gerhard , 1981 ) . with increasing computational power new n - body techniques
have been developped which increased the accessible particle number by many orders of magnitude .
especially the tree method developped by barnes & hut ( 1986 ) and hernquist ( 1987 , 1990 ) was perfectly suited for interacting systems , because the organization of the force calculation the most time consuming part of n - body calculations
adapts in the tree scheme easily and automatically to clumpy mass distributions for a moderate computational price ( @xmath5 ) compared to direct simulations ( @xmath6 ) . by this ,
barnes ( 1988 ) was able to simulate encounters of disc galaxies including all dynamical components , i.e. the disc , the bulge , and the halo as n - body systems .
compared to faster grid based methods ( e.g. sellwood , 1980 ) or expansion methods ( e.g. hernquist & ostriker , 1992 ) direct n - body simulations ( or semi direct methods like trees ) seem to be more flexible with respect to strongly varying geometries and scalelengths .
an alternative to these sophisticated numerical techniques is the use of special purpose computers like the machines of the grape project ( sugimoto et al . ,
they implement newton s law of gravity ( modified for gravitational softening ) in the hardware which allows for a very fast direct determination of the gravitational forces , though there still exists the @xmath7 bottleneck .
e.g. for simulations with @xmath8 particles a grape3af ( with 8 grape processors ) is competitive with a tree
code running on a cray t90 .
simultaneous to the development of efficient stellar n - body integrators mainly two methods ( based on particles ) for the treatment of a dissipative , star forming gas component have been applied : the _ smoothed particle hydrodynamics _ scheme ( sph ) solves the gasdynamical equations , i.e. the diffuse nature of the interstellar medium ( ism ) is emphasized ( e.g. hernquist & katz , 1989 ) .
an alternative approach focussed on the clumpiness of the clouds in the ism .
these clouds were treated as _ sticky particles _ : without physical collisions with other clouds or close encounters they move on ballistic orbits like stars . however ,
in case of a collision the clouds might merge or loose kinetic energy depending on the adopted microphysics ( e.g. casoli & combes , 1982 ; palou et al .
, 1993 ; theis & hensler , 1993 ) . by these technical and numerical improvements of the last 10 years
reliable n - body simulations even for the infall of small satellites into larger galaxies have become possible ( e.g. quinn et al .
, 1993 ; walker et al . , 1996 ; huang & carlberg , 1997 ; velzquez & white , 1998 ) .
the interest in these minor mergers raised especially after the discovery of the sagittarius dwarf galaxy sgri which seems to be an ongoing merger with the milky way ( ibata et al .
, 1994 ) , i.e. minor mergers seem to be a common phenomenon .
this is in agreement with estimates based on the cosmological cdm model which predicts that 85% of the finally existing galaxies have undergone a merger with a companion of at least 20% of its own mass ( frenk et al . , 1988 ) .
despite some debate on this high merging rate ( tth & ostriker , 1992 ) tidal interaction or at least minor merging seem to be a normal evolutionary process for a large fraction of giant galaxies . though there is a large number of papers dealing with principle effects of interactions like tidal tails or induced star formation , there exist considerably fewer papers about the modeling of special objects ( e.g. salo & laurikainen ( 1993 ) on ngc7752/53 or thomasson & donner ( 1993 ) on m81 and ngc3077 ) .
this deficiency is mainly caused by two reasons : first , for a detailed confrontation of theory with observation high resolution data in configuration and velocity space are required .
these should cover a large fraction of the space between the interacting galaxies . in principle
, hi would be a good tracer for the global dynamics , however , there are just a few observational sites which give data of sufficient quality .
the second problem is the large parameter space resulting in two connected difficulties : finding a good fit at all and determining its uniqueness ( or other acceptable parameter sets ) .
observationally only three kinematical quantities the projected position on the sky and the line - of - sight velocity can be measured .
so , for each galaxy the line - of - sight position and the two components of the velocity in the plane of the sky are unknown .
another parameter , the galactic mass , depends on the availability of velocity data , the determination of the distance , and the reliability of the conversion from velocity to masses .
these 14 parameters just fix the orbit in case of a two body interaction .
moreover , one has to specify the parameters that characterize both stellar systems , e.g. characteristic scales , orientation , or rotation of a disc . by
this one ends up with a high dimensional parameter space which is even including some prejudices about the system in general too large for a standard search method . for instance , let us assume that the orientation , rotation , and scale length of the discs , the galactic masses and all the 6 available kinematic data are known .
the remaining parameter space has still 6 dimensions .
a regular grid with a poor coverage of 10 grid points per dimension demands @xmath9 models or 340 years grape simulation time ( assuming 3 cpu
hours for a complete calculation ) or still a month with a restricted n - body program ( assuming 3 cpu - seconds per simulation ) .
an alternative approach to investigate large parameter spaces are genetic algorithms ( holland , 1975 ; goldberg , 1989 ) .
the basic idea is to apply an evolutionary mechanism including sexual reproduction operating on a set of parameters which are selected for parenthood according to their fitness. the fitness is determined by the ability of the parameters to reproduce the observations , whereas the reproduction is done by cross
over of the parental parameters and applying a mutation operator .
though genetic algorithms ( ga ) have been used in many branches of science , there are just a few applications in astrophysics , e.g. for fitting rotation curves or analysis of doppler velocities in @xmath10 scuti stars ( for a review see charbonneau , 1995 ) .
recently , wahde ( 1998 ) demonstrated the ability of genetic algorithms to recover orbital parameters for artificial data generated by n - body simulations of strongly interacting galaxies .
he stressed that in these systems the positional information can be already sufficient to determine the orbital elements . in this paper
i want to demonstrate how different n - body techniques can be combined in order to create a model for the dynamics of interacting galaxies provided the intensity distribution is available .
the scheme consists basically of two steps . in the first step
restricted n - body simulations will be performed to constrain the model parameters : if a sufficient accurate data set is available , the genetic algorithm will be used to find the orbital parameters and to check the uniqueness of the result .
additionally , the sensitivity of the solution can be checked by an extended parameter study .
alternatively , in case of insufficient data , one could at least use the restricted n - body models to get a first guess of the parameters , by this creating an artificial intensity map and then applying the genetic algorithm to check its uniqueness . in the second step self
consistent models are necessary in order to tune the parameters and to check the applicability of the restricted n - body method .
these checks have to address two questions : is the neglect of self gravity acceptable ? does gas dynamics alter the results significantly ? technically this means that self consistent simulations with and without gas have to be performed .
the numerical models of this paper are performed with respect to ngc4449 which is a prototype of an actively star forming galaxy similar in size and mass to the large magellanic cloud .
ngc4449 is surrounded by an extended hi halo which shows some distortion ( van woerden et al . , 1975 ; bajaja et al . ,
hunter et al .
( 1998 ) published recently high resolution hi images taken at the vla showing a large scale lopsided structure .
though there seems to be a close dwarf companion , ddo 125 , it is not clear whether the observed streamers are tidal tails caused by an interaction with ddo125 ( hunter et al . , 1998 ) .
for example , tidal distortions are missing in the optical image of both galaxies and no clear bridge has been detected .
however , assuming a face on encounter theis & kohle ( 1998 ) demonstrated qualitatively by n - body simulations that the main hi features could stem from such an encounter provided the mass of ddo125 exceeds 10% of the mass of ngc4449 .
an alternative scenario assumes that the hi structure is an ongoing infall of gas onto ngc4449 .
this might be caused by the encounter with ddo125 followed by a compression and cooling of the gas resulting in an infall of the gas ( silk et al . , 1987 ) .
however , it seems to be difficult that the observed large
scale , lopsided and regular structure in ngc4449 can be maintained over a typical timescale like the orbital period which is longer than 1 gyr for the outer streamer . already
the high internal velocity dispersion of 10 kms@xmath11 would destroy the streamers on a timescale of @xmath12 yr ( hunter et al . , 1998 ) . in this paper
i investigate the interaction scenario addressing the question whether the observed structures can be formed by tidal interaction at all and what the constraints on the mass distribution or the orbits of both galaxies are .
[ observations ] summarizes the observational data of ngc4449 and ddo125 relevant for the models of this paper .
the different modeling techniques and the numerical results are described in sect .
[ numerics ] : first , the restricted n - body method is introduced ( sect .
[ restricted ] ) and a set of first results is shown in sect .
[ firstguess ] and [ parameterstudy ] .
afterwards the genetic algorithm approach is explained in sect .
[ geneticalgorithm ] and its results are shown in sect .
[ unique ] . finally , the restricted n - body simulations are compared with self - consistent models with ( sect . [ stickyparticles ] ) and without gas ( sect . [ scnogas ] ) .
the results of these different approaches are summarized and discussed in sect .
[ discussion ] .
when the optical images of ngc4449 and the large magellanic cloud are compared , many similarities like the presence of a bar , the size or the total luminosity are found
. however , different to the lmc , which has the milky way in its neighbourhood , ngc4449 seems to be a perfect candidate to study the properties of a genuine magellanic type irregular unaffected by the tidal field of a large galaxy .
its distance of 2.9 5.6 mpc for the hubble constant . independently of @xmath13 karachentsev & drozdovsky ( 1998 ) derived from the photometry of the brightest blue stars a distance of 2.9 mpc .
however , the error of this measurement is unclear . in accordance with hunter et al .
( 1998 ) a distance of 3.9 mpc will be assumed in this paper .
thus , @xmath14 corresponds to 1.1 kpc .
] allows also for detailed investigations which were performed in the visual ( e.g. crillon & monnet , 1969 ; hunter & gallagher , 1997 ) , in the radio band ( e.g. van woerden et al . , 1975 ; klein et al . , 1996 )
, in the hi emission line ( e.g. bajaja et al . , 1994 ; hunter & gallagher , 1997 ; hunter et al . ,
1998 ) , in the co emission line ( e.g. sasaki et al . ,
1990 ; hunter & thronson , 1996 ; kohle et al . , 1998 ) , in the infrared ( e.g. hunter et al . , 1986 ) , in the fuv ( e.g. hill et al . ,
1994 ) and in x rays ( e.g. della ceca et al . , 1997 ; bomans & chu , 1997 ) . inside the holmberg diameter of 11
kpc ngc4449 shows ongoing strong star formation along a bar - like structure ( hill et al . , 1994 ) .
this activity seems to induce a variety of morphological features like filaments , arcs , or loops which have a size of several kpc ( hunter & gallagher , 1997 ) .
already the early radio observations of van woerden et al .
( 1975 ) exhibited a large hi
structure exceeding the holmberg diameter of 11 kpc by a factor of 7.5 .
recent observations by bajaja et al .
( 1994 ) with the effelsberg telescope and more detailed results by hunter et al .
( 1998 ) with the vla revealed a complex structure of several components ( fig .
[ hunterobservation ] ) : an elongated ellipse of hi gas centered on ngc4449 has a total mass of @xmath15 .
this gas rotates rigidly inside 11 kpc reaching a level of 97 kms@xmath11 in the deprojected velocity .
outside 11 kpc the rotation curve is flat .
bajaja et al .
( 1994 ) derived a dynamical mass of @xmath16 inside 32 kpc and a hi gas mass fraction of 3.3% . within 11
kpc the dynamical mass is @xmath17 and the mass inside the optical radius of about 5.6 kpc is @xmath18 .
applying single dish data for a @xmath19 beamwidth hunter et al .
( 1998 ) derive about @xmath20 .
kinematically interesting is the counterrotation of the hi gas inside the optical part of ngc4449 with respect to the outer regions . for the central ellipse hunter et al .
( 1998 ) determined a position angle of @xmath21 and an inclination of @xmath22 . south of the galactic center a streamlike structure of 25 kpc emanates which abruptly splits into two parts : a small spur ( 5 kpc ) which points towards ddo125 , a close irregular galaxy , and a long extended stream .
the latter first points 24 kpc to the north ( in the following : _
vertical streamer _ ) , then 49 kpc north east and , finally , 27 kpc to the south east ( together called the _
northeastern streamer _ ) , by this covering 180@xmath23 around ngc4449 .
remarkable are the long straight sections and the abrupt changes of direction .
the total mass of the streamers measured by the vla observations is about @xmath24 .
hunter et al .
( 1998 ) pointed out that by comparison with effelsberg data there is an additional diffuse hi mass not detected in the vla observations which amounts to two thirds of the hi in the extended structures .
ddo125 is located in the south east of ngc4449 at a projected distance of 41 kpc .
the line - of - sight velocity difference between ddo125 and ngc4449 is only about 10 kms@xmath11 .
tully et al .
( 1978 ) found rigid rotation inside the optical radius of ddo125 and an inclination of the adopted disc of @xmath25 . from this
they derived a total mass of @xmath26 ( corrected for a distance of 3.9 mpc ) which is in agreement with single dish observations ( @xmath19 beamwidth ) of fisher & tully ( 1981 ) .
however , ebneter et al .
( 1987 ) reported vla
observations which show a linear increase of the rotation curve out to the largest radii where hi has been detected , i.e. out to twice the optical radius .
this already allows for an estimate of a lower mass limit for ddo125 which is a factor 8 larger than the value of tully et al .
( 1978 ) , i.e. @xmath27 . since neither flattening of the rotation curve nor
any decline has been observed , this value is only a lower limit . if one compares that mass with the _ total _ mass of ngc4449 derived by bajaja et al .
( 1994 ) one gets a mass ratio of about @xmath28 which is , however , just a lower limit . comparing the masses inside the fiducial range of rigid rotation gives a higher mass ratio of @xmath29 .
for the simulations described in this paper three different n - body approaches have been applied : the method of restricted n - body simulations ( pfleiderer & siedentopf , 1961 ; toomre & toomre , 1972 ) is described in sect . [
restricted ] .
the self gravitating models are split into calculations without gas ( sect .
[ scnogas ] ) and simulations which include gas by means of sticky particles ( sect .
[ stickyparticles ] ) .
the advantage of the first method is the reduction of the full n - body problem to @xmath3 single body problems and the use of particles as test particles which gives high spatial resolution at the places of interest for almost no computational cost .
furthermore , the cpu time scales linearily with particle number which is superior to almost all n - body codes . because of this the calculation is 1000 times faster than self
consistent simulations .
hence , the restricted n - body calculation is at the moment the only method which allows about @xmath30 simulations in a couple of cpu - hours .
such a fast computation is a prerequisite for all systematic searches in parameter space like the genetic algorithm described in sect .
[ geneticalgorithm ] .
on the other hand , the simplified treatment of the galactic potential in restricted n - body simulations prohibits any effect of the tidal interaction on the orbits of both galaxies , e.g. no transfer of orbital angular momentum into galactic spin is possible ( i.e. no merging ) . in order to overcome this problem , the results of restricted n - body models and genetic algorithms must be compared with detailed self consistent simulations . in a first step the stellar dynamical applicability of the restricted models
is checked by a comparison with self gravitating systems calculated with a grape3af special purpose computer ( sugimoto et al .
1990 ; see sect .
[ scnogas ] ) .
so far the dynamics of the system has been investigated in terms of stellar dynamics , though in many cases ( and also in the case of ngc4449 ) hi _ gas _ is observed . in general , the effects of neglecting gasdynamics are not clear .
e.g. a purely stellar dynamical ansatz was successfully applied for the system m81 and ngc3077 ( thomasson & donner , 1993 ) , whereas in other systems the dynamics of the gas can deviate strongly from the behaviour of the stars ( e.g. noguchi , 1988 ; barnes & hernquist , 1996 ) .
therefore , one has to compare stellar and gas dynamical results which is done in this paper by a sticky particle method ( sect .
[ stickyparticles ] ) .
+ the units are chosen to 1 kpc , @xmath31 ( the fiducial mass of ngc4449 ) and g=1 , which result in a time unit of 1.78 myr and a velocity unit of 549 kms@xmath11 .
the basic idea of restricted n - body simulations is to reduce the n - body problem to @xmath3 1body problems by assuming that the gravitational potential is given by a simple relation , e.g. a two or a few body problem which has a known analytical solution or can be solved by fast standard methods . here
i assume that the gravitational forces on each particle are given by the superposition of the forces exerted by two point
like objects ( the galaxies ) moving on keplerian orbits .
thus , for the orbits one has to specify 12 parameters which reduce to 6 in the center - of - mass frame : the orbital plane is fixed by the inclination angle and the argument of the ascending node .
the orbit itself is characterized by its eccentricity and the location of the pericenter , i.e. the pericentric distance and the argument of the pericenter .
finally , the initial location of the reduced particle ( or the time of pericentric passage ) has to be specified to fix the phase . for a comparison with the observations the masses of the galaxies and the phase of the observation have to be supplied , too .
three of the six orbital parameters can be fixed by the observed location of both galaxies on the plane of sky ( @xmath32-@xmath33 plane ) and the line - of - sight velocity .
the masses can be estimated by the velocity profiles of the galaxies provided the distance to them is known . in this paper
the time of pericentric passage is set to zero . as input the orbital eccentricity and inclination and the minimum distance
are given .
furthermore , the directly observable parameters , i.e. the masses of the galaxies , the projection of the relative position of both galaxies on the plane of sky , and their line - of - sight velocity difference are supplied .
they are used to determine the argument of the pericenter , the phase ( or time since pericentric passage ) of the observed system and the argument of the ascending node of the orbital plane .
additionally , the initial distance or phase has to be specified : in case of parabolic and hyperbolic encounters the distance was set to 100 kpc , which guarantees a sufficient low tidal influence at the start of the simulation . in case of an elliptical encounter , the situation is more difficult , because the apocentric distance might be small , especially if the eccentricity is low .
thus , the system might be affected permanently by tidal forces or even a series of repeated encounters which make the usage of the restricted n - body method doubtful . for these elliptical orbits the initial azimuthal angle of the orbit
was set to 120@xmath23 ( 0@xmath23 corresponds to pericenter ) .
the test particles are arranged in a flat disc moving on circular orbits .
the disc itself is characterized by its orientation , i.e.its inclination and position angle , the scalelength and the outer edge . additionally , an inner edge can be specified in order to avoid a waste of computational time by integrating tidally unaffected orbits well inside the tidal radius of the galaxy . in
most of the restricted n - body simulations the test particles are distributed in 30 to 60 equidistant rings of 125 particles , i.e. a total of @xmath34 to @xmath35 particles .
the typical spacing of the rings is 0.7 to 1.5 kpc depending on the outer edge of the disc .
the time integration of the equations of motion is performed by a cash
karp runge kutta integrator of fifth order with adaptive timestep control ( press et al . , 1992 ) . in order
to prevent a vanishingly small timestep by an accidental infall of a test particle to the galactic center , the gravitational force is softened on a length scale of 1 kpc . in order to speed - up the calculation
the galactic positions are tabulated by cubic splines and the integrator uses these look - up tables . on a sparc10 with a 150 mhz cpu a typical 7500-particle encounter takes 23 cpu seconds , which can be reduced by a factor of 2 - 3 by applying an inner edge of 10 kpc . for a typical genetic algorithm simulation with 900 particles and no offset for an inner edge 2 cpu seconds are required per simulation . in a series of simulations the orbital parameters ( minimum distance @xmath36 , the eccentricity @xmath37 , the mass ratio @xmath38 and the orientation of orbital plane ) as well as the disc parameters ( size @xmath39 of the disc , its position angle @xmath40 and inclination @xmath41 ) are varied ( see table [ modeltable ] ) .
the temporal evolution of the primary galaxy ( ngc4449 ) in the reference model a is displayed in fig .
[ resevol ] . at the beginning
both galaxies have a distance of 100 kpc corresponding to 995 myr before closest approach .
the particles inside 10 kpc are omitted for an easier identification of the evolving structure .
this neglect does not alter the results , since the related particles are only weakly affected by the interaction , as the persistence of the elliptical shape of the central hole demontrates .
at the moment of closest approach ( @xmath42 ) , only the outer region of ngc4449 s disc reacted on the intruder .
71 myr later ( i.e. system time 40 ) , a dense outer rim starting at the projected position of ddo125 and pointing to the north begins to form .
secondly , a weak bridge - like feature seems to connect both galaxies .
both features are more pronounced after 178 myr . at that time
an additional tidal feature commences to evolve north of the center .
when ddo125 has a projected distance of 41 kpc which occurs 362 myr after closest approach , three features are discernible : the remnant of the former bridge starts south of the center of ngc4449 pointing to the south - west .
the vertical feature starts at the end of the first structure and points to the north .
the third structure is a long tidal arm which starts west of ngc4449 pointing to the north - east before turning again back to the south - east . a comparison of the intensity map of model a ( lower right diagram in fig . [ resevol ] ) with the observations ( fig .
[ hunterobservation ] ) demonstrates that the main observational features can be reproduced qualitatively by the reference model a , i.e. a parabolic orbit , a mass ratio @xmath43 and a minimum distance @xmath44 kpc .
the parameters for the disc orientation are chosen in agreement with hunter et al.s suggestion of @xmath45 , @xmath46 , whereas the disc radius was set to @xmath47 kpc .
though this numerical model is not a detailed fit of the data , the characteristic sizes and locations of the streamers as well as the unaffected disc can be found .
the material located between the streamer and the disc , which is not seen in hunter s data , could correspond to the extended hi - mass detected in the single dish observations by van woerden et al .
in the following sections i will discuss the influence of the individual parameters . the global uniqueness of the reference model is investigated in sect .
[ unique ] . in this section
the influence of the basic parameters , i.e. the orientation of the orbital plane , the orbital eccentricity and the minimum distance , the mass ratio of both galaxies , the orientation of the hi disc and its extent are discussed ( s. table [ modeltable ] ) . in the following simulations all parameters except the one of interest are kept constant with respect to model a. thus , the vicinity of the reference model in parameter space is studied .
+ [ modeltable ] _ model parameters of the restricted n - body simulations _ ._parameters of the different restricted n - body models : model name , orbital eccentricity @xmath37 ( col . 2 ) , mass ratio @xmath38 of both galaxies ( col . 3 ) , minimum separation @xmath36 of galactic centers ( in kpc , col . 4 ) , inclination @xmath41 of orbital plane ( in degrees , col . 5 ) , orientation of ngc4449 disc ( in degrees ) : inclination @xmath48 ( col . 6 ) and
position angle @xmath49 ( col .
7 ) , maximum radius @xmath50 of gaseous disc . _ [ cols="<,^,^,^,^,^,^,^,<",options="header " , ] * eccentricity .
* the eccentricity mainly influences the interaction timescale which is defined here as the time between maximum approach and a projected distance of 41 kpc , i.e. the moment of observation .
an eccentricity of 5.0 ( corresponding to a hyperbolic encounter ) decreases the interaction time by a factor 2.6 resulting in less damage of the primary galaxy ( fig .
[ resecc ] ) : only the remnant of the bridge feature can be discerned , whereas the other two streamers found in the reference model are missing completely . with decreasing eccentricity
the north - eastern streamer becomes pronounced before the vertical structure arises .
the latter is weakly expressed for @xmath51 . reducing @xmath37
further to bound orbits makes the features even more prominent . in case of @xmath52 the north - eastern and the vertical streamer form even a single structure clearly detached from the center
however , due to the strength of the interaction a second broad tidal arm is ejected from ngc4449 leaving a low - density region south to the northern tip of the north - eastern streamer
. + * mass ratio . *
the variation of the mass ratio gives a lower limit of about 10% in order to create a sufficient tidal response ( fig .
[ resmass ] ) .
if the mass ratio exceeds 0.3 , the vertical feature becomes less vertical and the north - eastern streamer is more diffuse , both due to the enhanced tidal pull by the intruder .
+ * minimum distance .
* though the interaction time is not strongly influenced by the minimum distance @xmath36 , it determines the maximum strength of the tidal field ( fig .
[ resrmin ] ) : for small separations the structures become more diffuse and the primary is much stronger affected .
the vertical feature is much more extended creating an additional tidal arm ( similar to the model with an orbital eccentricity @xmath52 ) .
if the distance of closest approach exceeds 30 kpc , almost no streamers are formed .
+ * orbital plane . *
the orientation of the orbital plane characterizes basically the observer s location or the line - of - sight which is fixed by two parameters . for the models here
only the orbital inclination is a free parameter , since the second parameter , the position angle of the orbital plane , is fixed by the observed relative line - of - sight velocity of both galaxies .
the models show that the inclination mainly affects the orientation of the vertical streamer ( fig .
[ resiorb ] ) : an inclination outside the range @xmath53 $ ] gives too much deviation from the south - north direction of the vertical feature .
+ * hi distribution in ngc4449 . *
a specification of the mass- and especially the hi - distribution is a more difficult task , because contrary to the well - known two - body problem the number of free parameters characterizing the structure and internal dynamics of the interacting galaxies is unknown and a matter of investigation itself .
for example , it is not a priori clear whether the hi is distributed like a disc , a spheroid , or something else .
in order to reduce the number of free parameters i will assume here that the observed inner elongated ellipse is the remnant of an original exponential hi disc , which can be specified by its orientation ( which is observed ! ) , its scale length and radial extent . since the particles are treated as test particles , a variation of the scale length does not change the final location of a particle , but its mass which is required for the calculation of the intensity map . a small scale length ( e.g. 5 kpc ) would confine too much hi in the center that is unaffected by the encounter .
since the observations show that a large fraction of hi is outside 20 kpc , i have chosen a large scale length of 30 kpc which gives only a weak radial decline of the hi surface density and , by this , a sufficient amount of hi in the outer regions . disc size .
the influence of the maximum radius @xmath39 of the disc is shown in fig .
[ resrhig ] .
if @xmath39 is smaller than 35 kpc , the vertical streamer has not been formed at all indicating that this feature stems from the hi gas far outside the center . the north - eastern arm and the bridge ,
however , are already built up for @xmath54 kpc . with increasing disc size
the vertical streamer becomes more pronounced and longer which excludes disc sizes larger than 45 kpc .
disc orientation .
the orientation of the hi disc also strongly affects the appearance of the primary galaxy .
a variation of the position angle @xmath40 by 20@xmath23 gives already hi distributions which are incompatible with the observations ( fig .
[ resdisc ] ) . if @xmath40 is decreased , the vertical streamer is more extended and diffuse . on the other hand ,
if @xmath40 is increased , the north - eastern arm becomes too long and the vertical feature vanishes almost completely .
similarly , the results depend strongly on the inclination @xmath41 : inclinations below @xmath25 give only weak bridges and weak vertical streamers , whereas the orientation of the vertical streamers deviates strongly from the northern direction for inclinations exceeding 70@xmath23 .
+ the performed parameter study demonstrates that even close to the parameters of model a , the final particle distribution shows significant variations which are not in agreement with the observations .
thus , the reference model might be a good and locally unique candidate describing the dynamics of ngc4449 and ddo125 .
the confidence in this encounter scenario is especially enhanced by the strong dependence on the orientation of the hi disc of ngc4449 . qualitatively good models are only found , if the disc orientation agrees with the values determined independently by observations . on the other hand
, the former parameter study only investigates a small fraction of parameter space and , thus , the uniqueness of the solution is not confirmed .
moreover , in general it is very tedious to find a good model by hand , i.e. by an unsystematic non automated search in parameter space .
thus , the probability to accept the first solution found seems to be very large ( especially , if cpu intensive simulations are performed ) , by this neglecting other regions in parameter space which might also give acceptable fits . in the next section an automatic search in parameter space by means of a genetic algorithm
is introduced .
this allows in principle for an _ ab initio _ determination of acceptable parameter sets ( if data sets of sufficient quality are available ) or at least for a uniqueness test of a favoured scenario like the reference model described above .
the idea of applying models of organic evolution for optimization problems in engineering dates back to the 1960s and 1970s ( e.g. rechenberg , 1965 ; schwefel , 1977 ) .
different to standard gradient techniques for optimization which work deterministically ( e.g. the downhill simplex method ( press et al . , 1992 ) ) rechenberg s concept of _ evolutionsstrategie _ is a probabilistic one : starting with a more or less random individual ( _ parent _ ) , i.e. a single point in search space , a _ child _ is generated by a random mutation of the parameter set characterizing the parent .
the quality of both individuals with respect to the optimization problem ( i.e. their _ fitness _ ) determines now the parent of the next generation .
repeating this process of mutation and selection improves the quality of the individual monotonously .
this simple evolutionary strategy , which is just a special implementation of evolutionary algorithms ( eas ) in general , demonstrates the advantages of these methods : compared to complete grids in the parameter space the probabilistic , but oriented nature of the evolutionary search strategy allows for an efficient check of a high dimensional parameter space .
compared to gradient methods which are very fast near the optimum , eas do not need any gradient information which might be computationally expensive .
additionally , they depend only weakly on the starting point and - most important - they are able to leave local optima .
the price for these features is a large number of fitness evaluations or test points in search space before converging to a good solution .
holland ( 1975 ) extended the _ evolutionsstrategie _ of rechenberg by abandoning asexual reproduction : he used a population of individuals instead of a single individual ( and its offsprings ) . from the population two individuals were selected according to their fitness to become parents .
these parents represent two points ( _ chromosomes _ ) in search space and the corresponding coordinates are treated like genes on a chromosome .
these chromosomes are subject to a _ cross over _ and
a _ mutation _ operator resulting in a new individual which is a member of the next generation .
this breeding is repeated until the next generation has been formed .
finally , the whole process of sexual reproduction is repeated iteratively until the individuals representing a set of points in parameter space confine sufficiently one or several regions of high fitness . despite many differences in the detailed implementations of genetic algorithms ( gas ) ,
the basic concept of an iterated application of randomized processes like recombination , mutation , and selection on a population of individuals remains in all gas ( goldberg , 1989 ; bck , 1996 ) . for the optimization problem of fitting an n - body simulation to a given observation
, i will briefly describe the main ingredients of my fitting ga - program in the following sections .
the ga used here is a slightly modified version of p. charbonneau s code pikaia ( for details see charbonneau ( 1995 ) ) . in order to determine the fitness of an individual ( i.e. a special parameter set )
, the n - body simulation has to be performed and compared with the observation .
this is done by mapping the particles on a quadratic grid of size 60 kpc centered on ngc4449 .
the resolution of the grid was set to 7x7 in order to allow for a better resolution of the tidal features of an encounter . for each grid cell
the intensity was calculated by summing up the masses of the individual particles in that cell .
the conversion to observed intensities can be performed by assuming a mass to light ratio for the particles ( or treating it as another free parameter ) or by normalizing the intensities . here
the intensities are normalized to the total intensity of the whole grid .
the quality can be measured now by the relative deviation @xmath10 of the intensities in both maps : @xmath55 the fitness @xmath56 is then estimated by @xmath57 in case of an impossible parameter configuration ( e.g. a circular orbit with a distance less than the observed projected distance ) or any convergence problems the fitness of that point in search space is set to zero prohibiting any offsprings .
typically only a few models in @xmath30 parameter sets fail , most of them in the first few generations before the system starts to approach the search space region of interest . in order to determine the parents of the next generation the members of the actual generation
are ranked by their fitness .
then the parents are selected randomly , whereas the probability to be chosen is proportional to the rank ( _ roulette wheel selection _ ) . a typical example for the status of the ga after a couple of generations is shown in fig . [ fig3 ] .
in the left column the positions of the original data are discretized on the grid to yield the intensities displayed on the lower left .
it should be noted that the resolution of the grid is about 8.6 kpc which rather exceeds the thickness of the tidal structures . in the middle column
the best fit of the actual ( here 7th ) generation agrees already qualitatively with the original . the quality of the fit is calculated only by a comparison of the lower two intensity diagrams .
the fitness distribution in the population resembles a gaussian at the beginning of the fitting procedure . in organic systems
the information is coded in genes in form of sequences of the four nucleotide bases where a triplet of them ( a codon ) encodes one amino acid , the building block of proteins .
the genes themselves are arranged in one or several chromosomes which contain all the genetic information ( or the genotype ) .
the phenotype which is subject to the selection process is the result of the genetic information and the interaction with its environment
. translated to our fitting problem , the phenotype is the final intensity map created by a special choice of parameters or initial conditions .
the latter correspond to the proteins which express the genetic information physically .
the mapping between the parameters and the genes or the number of chromosomes is not fixed in gas .
many gas apply a binary alphabet for encoding the parameters ( e.g. holland , 1975 ; goldberg , 1989 ) . in charbonneau s program pikaia a decimal encoding is implemented . for my calculations i combined four decimal digits giving a number between zero and one .
this gene has been mapped to a real parameter by a linear or logarithmic transformation to the investigated parameter range , e.g. the orbital inclination is mapped linearily to the range [ @xmath58 , 180@xmath23 ] ( fig .
[ coding ] ) .
all the genes are assumed to be located in linear manner on a single chromosome .
different to organic evolution the phenotype is completely determined by the genotype , no environmental influence on the individual is acting after its genotype is fixed .
the main difference between the evolutionary strategies and genetic algorithms is the sexual reproduction of gas .
this means that the chromosomes of the parents are combined in a new manner by means of a cross over mechanism in addition to mutation .
the principle idea is illustrated in fig .
[ crossover ] : after determining the two parents a random position on the chromosomes is selected . at this position the chromosomes are split into two fragments which are exchanged among the chromosomes resulting in two new ones . in order to introduce completely new information into a genepool a mutation process , i.e. a random change of the information in the chromosomes
is introduced ( fig .
[ mutation ] ) . for each position on a chromosome mutation
is applied with a small probability ( typ .
@xmath60 ) . after selecting a position for mutation
, an integer random number in the range [ 0,9 ] replaces the original value at that position .
thus , a real change in the genetic information takes place with a probability of 90% , even if that position of a gene is subject to mutation .
the overall probability for at least one real mutation on a chromosome coding 6 parameters by 4 digits is about 11% .
thus , cross over is the most important process introducing diversity into the population , whereas mutation affects only each 9th member of the population . in principle
, the mutation rate can be increased , but the price would be a more or less random search in parameter space which is already after a few generations inferior to the oriented search of a ga method ( wahde , 1998 ) .
a principle danger of gas is inbreeding , i.e.the homogenization of the genepool by one or a few dominant genes .
if they successfully replace other genes , cross over does not introduce sufficient ( or if inbreeding is complete any ) diversity into the population and thus any further improvement of the chromosomes is prohibited . the only way to overcome this situation is to introduce more diversity by mutation .
therefore , pikaia allows for a variable mutation rate which increases the mutation probability @xmath61 whenever the fitness distribution becomes narrow , i.e. inbreeding is suspected to operate . on the other hand @xmath61
is decreased , if the fitness distribution becomes very broad , i.e. the randomization is supposed to be dominant .
after @xmath62 children have been created , the old generation is completely replaced ( _ full generational replacement _ ) by the next generation .
the only exception is the survival of the fittest member of the parent generation , if it is fitter than the child which would replace it ( _ elitism _ ) .
the latter mechanism prevents the system from forgetting successful solutions .
the initial population is randomly chosen in the accessible parameter space .
a typical population size is about @xmath63 members .
the ga converges typically after about @xmath64 generations .
thus , @xmath30 or a few @xmath30 n - body simulations are required for a single ga fit .
consistent n - body models like the mentioned 3cpu - h grape models would still need 3.4 years , whereas the restricted n - body models reduce the cpu requirement to 5.6 hours on a 150mhz sparc10 . in order to check the uniqueness of the reference model a
, several runs with different parameter sets encoded in genes were performed .
[ gaevol ] shows a typical run of the ga operating on a population of 100 members .
the reference model displayed in the upper left diagram as well as the ga models have been calculated with 900 test particles .
this makes the streamers less prominent , but still discernible .
the next plots show the ga status at different generations : at generation 1 , the best model is just created out of the randomly chosen initial parameter sets .
however , already at the second generation , i.e. the first application of the reproduction operator , the best model clearly exhibits the north - eastern streamer . at generation 11 ,
even the weaker vertical feature and the bridge feature show up . at the end of this ga
run the particle distribution of the original model a and the best fit of the ga are almost identical , at least when compared by eye .
the fitness of the best fit has strongly increased from 0.04 to 0.14 where values above 0.1 indicate already a very good fit .
most of the improvements were found during the first 50 generations , i.e. analyzing 5000 models .
after that generation the homogeneity of the population led to inbreeding which strongly prohibits the arise and spread of new better parameters .
however , this is not a problem for the models here , because a very good fit to all parameters is already found after 40 generations as the direct comparison with the parameters of the reference model demonstrates ( fig .
[ gaparams ] ) .
the relative deviation of the derived parameters from the original values is less than 15% in all cases , and for many of them even better than 5% . in a series of different ga runs
i varied the population size , the number of generations or particles , the parameter combination encoded on chromosomes , the fitness function , and the random initial population . in almost all of them the best fit gave an acceptable fit of the reference model which demonstrates the capability of gas to recover the parameters even for weak encounters .
the models which failed basically suffered from inbreeding or from a fitness function which did not discriminate sufficiently between high and low quality fits . moreover , in none of the ga runs a different region in parameter space producing a good fit ( @xmath65 ) has been found .
_ thus , the model a or more exact the corresponding intensity map seems to result from a unique parameter set_. however , this is only a motivated guess , not a mathematical proof !
another interesting question is how strongly do the ga results depend on the completeness or accuracy of the data .
i studied this in two ways : first , i used a coarser grid ( 4x4 ) in order to mimic a lower resolution of observational data ( left column in fig .
[ gatest ] ) . then the ga is only qualitatively able to reproduce the reference model .
obviously , too much information is lost , when the resolution becomes too small . on the other hand , an increase of the number of cells to 10x10
did not improve the fit found on a 7x7 grid , but introduced more noise into the intensity maps . in the second test i discarded one or several grid cells from the fitness calculation ( middle and right column in fig . [ gatest ] ) .
omitting one cell does not affect the final fit , independently of the location of the cell ( here the center ) .
however , if a whole feature like the vertical streamer is neglected , the ga is unable to recover the original data .
the self consistent models in this paper were performed by a direct n - body integration using the grape3af board in kiel for the calculation of the gravitational forces and potentials .
the time integration was done by a leap frog scheme with a fixed timestep of 1.78 myr .
the grape board uses intrinsically a plummer softening , the softening length was chosen to be 0.2 kpc .
these choices give a total energy conservation of better than 1% during the whole simulation .
different to the restricted n - body models the set - up of the initial particle configuration for self - gravitating systems is far from being trivial . in principle
, a stable stationary stellar system fulfills the collisionless boltzmann equation .
thus , a solution of the latter could be used as a trial galactic system . from observations of the milky way it is known , however , that even if the galaxy is axisymmetric , three integrals of motion are required to construct a stationary model .
unfortunately , only two are known analytically ( the energy and the @xmath66-component of the angular momentum ) . additionally ,
even if a particle configuration is generated from a distribution function depending on the integrals of motion , it is unclear whether it is stable against perturbations which might lead to e.g. the toomre instability or the bar instability .
thus , one has to check the stability on the timescale of interest numerically . in this paper
i use the models of kuijken & dubinski ( 1995 ) consisting of three components which are determined self - consistently from the boltzmann equation : the bulge follows a king model and the halo a lowered evans model . for the disc
the third integral is approximated by the @xmath66-energy , i.e. the energy in vertical oscillations .
the advantage of kuijken & dubinski s method is the possibility to specify parameters like scale length or scale height of the disc directly , whereas many other methods ( e.g. barnes , 1988 ) that start not in virial equilibrium , readjust the mass distribution on a dynamical timescale , which means less control on the structural parameters .
the long term stability has been checked by simulations of an isolated disc with @xmath67 particles ( 8000 for the disc , 4000 for the bulge and 6000 for the halo ) . within 1.5 gyr
the lagrange radii at 10% and 90% mass vary only by 2 - 4% .
they show no systematic trend except for a small expansion during the first 100 myr which is probably caused by gravitational softening . due to two body relaxation the scale height of the disc increases by almost a factor of 2 within 1.5 gyr .
however , this heating is not expected to alter the result of the dominant interaction significantly .
the fourier amplitudes @xmath68 , defined by @xmath69 ( with the surface density @xmath70 ) , are always less than 5% , in agreement with no significant bar or spiral pattern built up during the simulation .
[ resscon ] shows a comparison of the restricted n - body simulations with the self - consistent models . except for the omitted central region in the former and the more diffuse structure in the latter ,
the large scale features are almost identical in both simulations .
the differences close to the central 10 kpc ( especially the phase of the bridge feature in the @xmath52 model ) are due to the extended dark halo in the self - consistent models .
it is remarkable that the vertical and the north - eastern streamer are not affected by the halo .
this confirms the applicability of the restricted n - body method here . in agreement with these simpler models
the intensity maps of both self - consistent models ( @xmath71 ) show clearly the streamer structure ( fig . [ resscon ] ) .
another interesting aspect concerns the response of the halo to the intruder .
different to the disc , the halo remains almost unchanged by the encounter ( fig .
[ scondiskhalo ] ) .
this constancy of the halo potential would in principle allow to include also a rigid halo in the restricted n - body models and , hence , also in the ga fitting procedure , by this extracting information about the dark halo . for the gasdynamical models i use a sticky particle scheme which treats each particle of the hi disc as a gaseous cloud .
the radius of these clouds is assumed to follow @xmath72 which is derived from the mass radius relation of rivolo & solomon ( 1987 ) corrected for glancing collisions ( @xmath73 is the mass of a gas cloud ) .
when the distance of two clouds falls below the sum of their radii , merging of the clouds is allowed , if their relative orbital angular momentum is smaller than the maximum angular momentum a merged cloud can acquire before break - up . if this angular momentum criterion prevents a merger , the clouds keep their kinematical data .
star formation is only qualitatively included by adopting an upper mass limit of @xmath74 for clouds .
if they exceed this limit no further merging is allowed , i.e. they are treated like stars . however , the star formation criterion is not important for the models of this paper .
( for details of the cloud model or the merging mechanism see theis & hensler , 1993 ) .
this dissipative scheme has been successfully applied to collapsing systems reproducing many observed properties of spiral and elliptical galaxies . for the long term evolution of barred disc galaxies a comparison of this scheme with that of palou et al .
( 1993 ) for @xmath75 gives basically the same results ( jungwiert , 1997 ) .
thus , the chosen model should be applicable for a reliable determination of the difference between gas dynamical and purely stellar dynamical models .
a comparison of gaseous models with dissipationless models gives no significant difference ( fig .
[ scondissnodiss ] ) . due to the low gas densities only about 20 - 30% of the clouds undergo inelastic collisions which are mainly occuring in the denser central regions .
thus , the outer hi halo is only very weakly influenced by ( sticky particle ) gas dynamics and the applicability of purely stellar simulations for the hi seems to be possible . it would be interesting to compare these models with an sph approach emphasizing the diffuse nature of the interstellar medium .
the previous simulations demonstrated that models based on the encounter scenario are able to reproduce the morphology of the three main streamers .
furthermore , the interaction puts gas to the south - west of ngc4449 , where a clumpy hi distribution is observed ( hunter et al . , 1998 ) .
though its detailed patchy structure has not been recovered by the models , an initial gas distribution less regular than the smooth disc - like one used for the numerical models might be responsible for this deviation .
the parameter study showed that already small variations in the parameters give hi distributions which are not compatible with the observations .
especially , the strong dependence of the final structure on the initial orientation of the hi disc confines its inclination and position angle within 10@xmath23 to the values derived _
independently _ from the observations .
this remarkable coincidence would occur necessarily in the encounter scenario , whereas in the infall scenario it happens just by chance .
a second critical feature is the abrupt direction change of the streamers . in the infall scenario
they mark the path of the perturbed gas which collapses to the galactic center .
however , it is hard to imagine a realistic static potential that could create orbits with such strong direction changes as e.g. from the vertical to the bridge feature .
moreover , due to the large distance of these features from the galactic center the potential should be more or less spheroidal or already spherical , which makes it even more difficult to produce such orbits .
the most critical parameter for the interaction scenario is the mass ratio @xmath38 of ddo125 and ngc4449 .
the simulations demonstrated that a mass ratio smaller than 10% can not reproduce the streamer structure . in order to determine @xmath38
, both galactic masses have to be known . in bajaja
et al.s ( 1994 ) estimate for the mass of ngc4449 they assumed that the hi is in rotational equilibrium even at a distance of 32 kpc .
however , in case of an interaction with ddo125 the velocities at large distances might be strongly affected by the interaction and definitely not being in rotational equilibrium as the ejection of the tidal arms in a continuation of the simulations demonstrates .
hence the derived mass of @xmath31 is at best an upper limit to the mass of ngc4449 .
a better guess might be the dynamical mass inside the central gaseous ellipse which seems to be unaffected by interaction . e.g. the dynamical mass inside 11 kpc would be a factor 3 smaller than bajaja s value . even more difficult than for ngc4449 is the mass determination of ddo125 .
observations by tully et al .
( 1978 ) gave a dynamical mass of @xmath26 inside the optical radius of ddo125 .
however , ebneter et al . ( 1987 ) reported a linear increase of the rotation curve exceeding the range observed by tully et al . by a factor of two
moreover , they found no hint for a flattening or turn - over in ddo125 s rotation curve .
this gives a lower mass limit of @xmath27 .
combining the mass estimates yield a lower limit of the mass ratio of 6% when using bajaja s mass determination of ngc4449 .
a more realistic mass ratio might be derived from the ranges of rigid rotation resulting in @xmath76 .
thus , observationally it can not be excluded that the mass ratio is below the critical one for the interaction scenario , however , the more realistic estimates seem to be in agreement with the values found in the simulations .
the situation could be clarified when a reliable mass determination of ddo125 is available .
another important question is related to the initial conditions of the previous simulations , i.e. the origin of the extended hi distribution . in principle
, there are several possibilities : _ idea 1 .
the hi comes from ddo125 .
_ in a previous encounter ( @xmath77 ) the gas is stripped off from ddo125 .
due to the long orbital period of e.g. 3.6 gyr for @xmath52 the gas has enough time to settle down in a regular disc - like structure .
however , the amount of hi gas in ddo125 is much smaller than the gas seen around ngc4449 .
it seems to be unlikely that ddo125 could have survived the loss of such a big fraction of its gas content without disruption .
the hi comes from ngc4449 .
_ though a star forming dwarf galaxy drives a galactic wind e.g. by means of type ii supernovae , a detailed fine tuning of the energy injection would be required to prevent a blow out of the hi
gas , especially if it got enough energy to form such an extended system .
moreover , the counter - rotation of the hi outside and inside the optical radius of ngc4449 makes the scenario not very appealing .
the hi gas stems from a previous minor merger with ngc4449_. though a satellite galaxy can be disrupted by the tidal field of ngc4449 , it is very unlikely that it is disrupted already far outside the galactic center where the hi gas is detected .
the hi originates from the collapse of a dwarf companion .
_ a dwarf galaxy forms in the vicinity of ngc4449 .
by this it undergoes a strong collapse which leads to mass ejection due to violent relaxation . in principle
, a mass loss of up to 30% of the companion could occur providing enough gas to explain the hi .
however , where is the dwarf today ? excluding an unlikely radial orbit , the dwarf s orbit could in principle decay by dynamical friction .
however , due to the large distance to ngc4449 the decay time should be at least several orbital periods which exceeds the hubble time .
another major encounter .
_ though ngc4449 is a member of the loose canes venatici group , there seem to be no close galaxy which has a radial velocity in agreement with the rotational sense of the gaseous disc in ngc4449 .
so , who is the partner ?
several n - body methods are combined in order to develop a method for the determination of the parameters of interacting galaxies .
this method is applied to the hi distribution of ngc4449 . in a first step
the fast restricted n - body method is used to confine a region in parameter space which reproduces the main observed features . in a second step a genetic algorithm ( also using restricted n - body calculations ) is employed which allows for both , an automatic fit of observational data even in a high - dimensional parameter space and/or a uniqueness test of a favoured parameter combination . for a genetic algorithm one typically has to follow a population of ( at least ) 100 members for 100 generations in order to get a good fit provided the data are sufficiently accurate .
missing single pixels do not inhibit the parameter determination , as long as the key features are included in the data . since a typical restricted n - body simulation takes a few cpu - seconds on a sparc workstation , a whole fitting procedure is finished after 36 cpu hours . in the third step the results of the previous steps
are compared with detailed self
consistent n - body simulations . in the case of ngc4449
they show that the restricted n - body calculations are reliable models for this encounter .
the comparison with the sticky particle models demonstrates that the hi gas can be modeled without any restriction by a purely stellar dynamical approach provided the encounter is as weak as in the case of ngc4449 and ddo125 . from the previous simulations i conclude that the extended hi features observed in ngc4449 are created by an encounter with ddo125 .
prior to the encounter the hi gas formed an extended , almost homogenous disc with a large scale length of about 30 kpc and a radial extension of 40 ( @xmath78 ) kpc .
the orientation of the disc is in agreement with the orientation of the inner ellipsoidal hi distribution .
the orbital plane has an inclination angle of about 40@xmath23 ( @xmath79 ) .
the eccentricity is close to that of a parabolic encounter ( @xmath80 ) and the apocenter distance of the galaxies is about 25 ( @xmath81 ) kpc .
the closest approach happened @xmath82 yr ago .
the mass ratio of both galaxies must be approximately 0.2 ( and can not be smaller than 0.1 ) .
* acknowledgements . *
the simulations were partly performed with the grape3af special purpose computer in kiel ( dfg sp345/5 ) .
the author is grateful to deidre hunter , jay gallagher , uli klein , sven kohle and hugo van woerden for stimulating discussions about ngc4449 .
i am also greatly indebted to konrad kuijken and john dubinski , who made their program for the generation of initial particle distributions available to the public , and paul charbonneau and barry knapp for providing their pikaia code .
last , but not least , i want to thank joachim kppen and tim freyer for a careful reading of the manuscript .
-.8ex kant i. , 1755 , in _ allgemeine naturgeschichte und theorie des himmels oder versuch von der verfassung und dem mechanischen ursprunge des ganzen weltgebudes nach newtonischen grundstzen abgehandelt _ ,
h. ebert , 1890 , verlag wilhelm engelmann , leipzig kohle s. , klein u. , henkel c. , hunter d.a .
, 1998 , in proc . of _ the magellanic clouds and other dwarf galaxies
t. richtler & j.m .
braun , bad honnef , see also http://www.astro.uni-bonn.de/ webgk / ws98/skohle_p.html | several n - body methods were combined in order to develop a method for the determination of the parameters of interacting galaxies .
this method has been applied to the hi distribution of ngc4449 . in a first step
the fast restricted n - body models were used to confine a region in parameter space reproducing the main observational features . in a second step a genetic algorithm is introduced which allows for an automated search in the high - dimensional parameter space as well as for a uniqueness test of special parameter sets . using artificial data
i show that the genetic algorithm reliably recovers orbital parameters , provided that the data are sufficiently accurate , i.e. all the key features are included . in the third step the results of the restricted n - body models
are compared with detailed self
consistent n - body simulations . in the case of ngc4449 the applicability of the simple
restricted n - body calculations is demonstrated .
additionally , it is shown that the hi gas can be modeled here without any restrictions by a purely stellar dynamical approach .
ngc4449 is an active star
forming dwarf galaxy of magellanic type . from radio observations
van woerden et al . (
1975 ) found an extended hi
halo around ngc4449 which is at least a factor of 10 larger than the optical diameter @xmath0 kpc .
more recently , bajaja et al . ( 1994 ) and hunter et al .
( 1998 ) discerned details in the hi halo : a disc like feature around the center of ngc4449 and a lopsided arm structure . in a series of simulations
it is demonstrated that the main features can be obtained by a gravitational interaction between ngc4449 and another nearby dwarf galaxy , ddo125 . according to these calculations the closest approach between both galaxies happened @xmath1 yr ago at a minimum distance of @xmath2 kpc on a nearly parabolic orbit . in case of an encounter scenario ,
the dynamical mass of ddo125 should not be smaller than 10% of ngc4449 s mass . before the encounter
the observed hi was arranged in a disc with a radius of 3540 kpc around the center of ngc4449 which had the same orientation as the central ellipsoidal hi structure .
the origin of this disc is still unclear , but it might be caused by a previous interaction . * modeling encounters of galaxies : * + * the case of ngc4449 * + christian theis _ institut fr theoretische physik und astrophysik , universitt kiel , 24098 kiel , germany , email : [email protected]_ 1.0 cm |
let @xmath2 be a non - orientable closed surface of genus @xmath0 , i.e. @xmath2 is a connected sum of @xmath0 real projective planes . the group @xmath3 of isotopy classes of diffeomorphisms over @xmath2 is called _ the mapping class group _ of @xmath2 .
let @xmath4 be the mod @xmath5 intersection form on @xmath6 , and @xmath7 the group of automorphisms of @xmath6 preserving the mod @xmath5 intersection form .
mccarthy and pinkall @xcite showed that the homomorphism from @xmath3 to @xmath7 defined by the action of @xmath3 on @xmath6 is surjective .
the kernel @xmath8 of this surjection is called _ the level @xmath5 mapping class group of @xmath3_. in this paper , we construct a minimal generating set for @xmath8 , and determine its abelianization .
lickorish @xcite showed that @xmath3 is generated by dehn twists and @xmath9-homeomorphisms , and korkmaz @xcite determined its first homology group .
furthermore , chillingworth @xcite found a finite system of generators for @xmath3 .
birman and chillingworth @xcite obtained the finite system of generators by using the argument on the orientable two fold covering of @xmath2 .
the group @xmath3 is not generated by dehn twists , namely , lickorish @xcite showed that the subgroup of @xmath3 generated by dehn twists is an index @xmath5 subgroup of @xmath3 . on the other hand ,
a @xmath9-homeomorphism acts on @xmath10 trivially , hence , the group @xmath3 is not generated by @xmath9-homeomorphisms . in @xcite , szepietowski proved that ( theorem 5.5 ) the level @xmath5 mapping class group @xmath8 is generated by @xmath9-homeomorphisms , and that ( corollary 5.6 ) @xmath8 is generated by involutions .
therefore , @xmath11 is a @xmath12-module .
@xmath9-homeomorphisms are defined as follows . a simple closed curve @xmath13 ( resp .
@xmath14 ) in @xmath2 is _ two - sided _
_ one - sided _ ) if the regular neighborhood of @xmath13 ( resp .
@xmath14 ) is an annulus ( resp .
mbius band ) . for a one - sided simple
closed curve @xmath15 and a two - sided simple closed curve @xmath16 which intersect transversely in one point , let @xmath17 be a regular neighborhood of @xmath18 , which is a union of the tubular neighborhood of @xmath15 and that of @xmath16 and then is homeomorphic to the klein bottle with a hole .
let @xmath19 be a regular neighborhood of @xmath15 .
we denote by @xmath20 a homeomorphism over @xmath2 which is described as the result of pushing @xmath19 once along @xmath16 keeping the boundary of @xmath21 fixed ( see figure [ fig : y - homeo ] )
. with circle indicates a place where to attach a mbius band.,height=151 ] we call @xmath20 a @xmath9-homeomorphism .
for a two sided simple closed curve @xmath22 on @xmath2 , we denote by @xmath23 a dehn twist about @xmath22
. then @xmath24 . for @xmath25 , we define an oriented simple closed curve @xmath26 as in figure [ fig : alpha ] . for short ,
we define @xmath27 , @xmath28 . for @xmath29 .
the small arrow beside @xmath26 indicates the direction of the dehn twist @xmath30 .
, height=102 ] szepietowski showed the following .
* theorem 3.2 , remark 3.10)[theorem : szepietowski ] when @xmath1 , @xmath8 is generated by the following two types of elements , 1 .
@xmath31 for @xmath32 , @xmath33 and @xmath34 , 2 .
@xmath35 for @xmath36 . in this paper ,
we reduce the number of generators of @xmath8 , and show that it is minimal .
[ proposition : generators ] when @xmath1 , @xmath8 is generated by the following two types of elements , 1 .
@xmath31 for @xmath32 , @xmath33 and @xmath34 , 2 .
@xmath37 for @xmath38 . [ theorem : main ] when @xmath39 , the set given in the above theorem [ proposition : generators ] is a minimal generating set for @xmath8 . in section 2 , we prove theorem [ proposition : generators ] and obtain an upper bound for the dimension of the @xmath12-module @xmath11 .
broaddus - farb - putman @xcite , kawazumi @xcite ( see also ( * ? ? ?
* part v ) ) , and perron @xcite constructed mod @xmath40 johnson homomorphisms in different manners on the level @xmath40 mapping class groups of orientable surfaces for a positive integer @xmath40 , independently . in section 3 , we consider _ the mod @xmath5 johnson homomorphism _ on the level 2 mapping class group of a non - orientable surface to determine its abelianization . in appendix
a , we give another definition of @xmath41 in the same manner as johnson s original homomorphism in @xcite .
in section 4 , by using this homomorphism , we give a lower bound for the dimension of the @xmath12-module @xmath11 . since our lower and upper bound are equal ( see lemma [ lemma : upper bound ] and lemma [ lemma : lower bound ] ) , we obtain the following . [
theorem : abel ] when @xmath42 , @xmath43 since the dimension of @xmath11 is equal to the number of our generators given in theorem [ proposition : generators ] , we obtain corollary [ theorem : main ] .
in this section , we prove theorem [ proposition : generators ] .
we first fix some notation . for @xmath44
$ ] and @xmath45 $ ] @xmath46 , we define a product @xmath47 as the isotopy class @xmath48 $ ] . for @xmath49 , [ \beta ] \in \pi_1(n_g)$ ] , the product @xmath49 \cdot [ \beta]$ ]
is represented by the closed curve following @xmath16 at first and then following @xmath50 .
let @xmath51 be the subgroup of @xmath8 generated by the elements ( i ) ( ii ) in the statement of theorem [ proposition : generators ] .
we will show that @xmath52 for any @xmath53 .
.,height=151 ] let @xmath54 be a regular neighborhood of @xmath55 in @xmath2 .
then @xmath56 .
we define @xmath57 .
let @xmath58 be a regular neighborhood of @xmath59 in @xmath54 .
then @xmath60 and one component of @xmath61 is isotopic to @xmath62 and the other component @xmath63 of @xmath61 bounds a disk in @xmath54 . for simplicity
, we set the notations @xmath64 , @xmath65 , @xmath66 , @xmath67 , @xmath68 .
then , by chain relation , we see @xmath69 , where the last equation is valid in @xmath3 because @xmath63 bounds a disk in @xmath2 . by braid relations
, we see , @xmath70 where @xmath71 means @xmath72 . since @xmath73 , we see @xmath74 . by drawing some figures ,
we observe that @xmath75 , @xmath76 , @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath82 , @xmath83 . by the above observations and the equation
@xmath84 shown by szepietowski ( * ? ? ?
* lemma 3.3 ) , we see @xmath85 @xmath86 . on the other hand , for each @xmath87 , there are two - sided curves @xmath88 intersecting @xmath89 transversely in one point and satisfy @xmath90 szepietowski showed @xmath91 in ( * ? ? ?
* lemma 3.5 ) .
therefore , we see @xmath52 . by the above argument , we can show the following two results .
[ prop : bcc map ] for any separating simple closed curve @xmath62 on @xmath2 , @xmath92 = 0 $ ] as an element of @xmath11 .
at least one component , say @xmath58 , of @xmath93 is non - orientable . since this surface @xmath58 is constructed from a disk by removing several disks and attaching several mbius bands , we can show that @xmath94 is the square of the product of @xmath9-homeomorphisms , by using the same argument of the proof for theorem [ proposition : generators ] . since @xmath11 is a @xmath12-module , we see @xmath92 = 0 $ ] .
[ lemma : upper bound ] when @xmath1 , @xmath95 since @xmath8 is generated by involutions , @xmath11 is a @xmath12-module .
as we saw in theorem [ proposition : generators ] , @xmath8 is generated by @xmath96 elements .
hence , the abelianization is also generated by @xmath96 elements as a @xmath12-module .
let @xmath98 be a closed disk in the surface @xmath2 , and pick a point @xmath99 in @xmath100 .
we denote by @xmath101 the group of isotopy classes of diffeomorphisms over @xmath2 fixing @xmath99 , and by @xmath102 the level 2 mapping class group @xmath103 . for @xmath104 ,
let @xmath105 denote the element represented by a loop as in figure [ fig : circle ] .
the fundamental group @xmath106 is a free group of rank @xmath0 generated by @xmath107 .
, height=102 ] we denote by @xmath108 the first homology @xmath6 , and denote by @xmath109 the homology class represented by @xmath110 .
the group @xmath108 is a free @xmath12-module of rank @xmath0 , generated by @xmath111 .
let us denote @xmath112 .
in this section , we construct an @xmath101-equivariant homomorphism @xmath113 , where @xmath114 is the submodule generated by @xmath115 .
we also lift " it to a homomorphism @xmath116 .
first , we review the magnus expansion .
see , for example , @xcite and @xcite .
let @xmath117 denote the completed tensor algebra generated by @xmath108 , and let @xmath118 denote the subalgebra @xmath119 .
note that the subset @xmath120 is a subgroup of the multiplicative group @xmath121 .
define a homomorphism @xmath122 by @xmath123 , which is called the standard magnus expansion .
denote by @xmath124 the composition map of @xmath125 and the projection @xmath126 . in the following ,
we denote by @xmath127 and @xmath128 . the map @xmath129 induces a map @xmath130 .
the boundary curve @xmath131 is represented by @xmath132 .
since @xmath133 is the fundamental group of the surface @xmath2 , @xmath134 is generated by @xmath135 . for @xmath136 , @xmath137 hence , for any @xmath138 , we have @xmath139 .
thus , @xmath140 induces a map @xmath130 . as a corollary of (
* lemma 2.1 ) , we have : [ lem : mod2 johnson ] let @xmath141 .
the map @xmath142 defined by @xmath143 is a crossed homomorphism .
by poincar duality , we have @xmath144 since the level 2 mapping class group @xmath102 acts on @xmath108 trivially , the restriction @xmath145 is an @xmath101-equivariant homomorphism . for @xmath146 ,
let us denote @xmath147 recall the forgetful exact sequence @xmath148 where the map @xmath149 is so - called the pushing map .
note that @xmath150 since @xmath151 acts trivially on @xmath152 .
we can compute values of @xmath41 as follows .
for an integer @xmath153 , let @xmath154 denote the @xmath155-invariant part of @xmath156 .
[ lem : image of tau1 ] for @xmath157 and @xmath34 , @xmath158 in particular , we have @xmath159 where @xmath160 is the composition map of the inclusion @xmath161 and the projection @xmath162 . when @xmath163 , the diffeomorphism @xmath164 acts on @xmath133 as follows .
@xmath165 thus , we have @xmath166 therefore , we obtain @xmath167 when @xmath163 .
when @xmath168 , the diffeomorphism @xmath164 acts on @xmath133 as follows .
@xmath169 in the same way , we also obtain @xmath167 when @xmath168 .
next , by ( * ? ? ?
* lemma 3.3 ) , we have @xmath170 for @xmath34 .
hence , we have @xmath171 finally , we consider @xmath172 . for @xmath34 , @xmath173
thus , we have @xmath174 [ lem : intersection of modules ] when @xmath141 , @xmath175 let @xmath176 denote the first stiefel - whitney class .
define a @xmath3-equivariant map @xmath177 by @xmath178 .
then , we have @xmath179 therefore , we see that @xmath180 , and the restriction map @xmath181 is injective .
they imply that @xmath182 . by lemma [ lem :
image of tau1 ] and lemma [ lem : intersection of modules ] , we have : the map @xmath183 lifts to the map @xmath116 with respect to the injective homomorphism @xmath184 .
the goal of this section is to prove the following lemma , which gives a lower bound of the dimension of @xmath11 as a @xmath12-module .
[ lemma : lower bound ] when @xmath1 , @xmath185 by lemma [ lemma : upper bound ] and lemma [ lemma : lower bound ] , we can determine the dimension of the @xmath12-module @xmath11 , and complete the proof of theorem [ theorem : abel ] .
let us denote @xmath186 .
define a homomorphism @xmath187 by @xmath188 when @xmath189 , @xmath190 is exact .
the set @xmath191 is a basis of @xmath192 .
since @xmath193 we have @xmath194 .
next , we consider @xmath195 .
it is clear that @xmath196 . for @xmath197 ,
let us denote @xmath198 .
since @xmath199 is a free @xmath12-module generated by @xmath200 , the set @xmath201 is a basis of @xmath202 . in particular , @xmath203 in the same way , we have @xmath204 and @xmath205 thus , we have @xmath206 since the dimensions of @xmath195 and @xmath202 are equal , we have @xmath207 . [ lemma : image of johnson ]
when @xmath1 , @xmath208 in the proof of lemma [ lem : image of tau1 ] , we have @xmath209 .
since @xmath3 acts on @xmath210 transitively and @xmath41 is @xmath101-equivariant , we have @xmath211 for any @xmath212 . for @xmath213
, we have @xmath214 thus , we obtain @xmath215 .
it implies @xmath216 on the other hand , we have @xmath217 it shows @xmath218 thus , we obtain @xmath219 the equation @xmath220 implies what we want .
we can now prove lemma [ lemma : lower bound ] . by lemma [ lemma
: image of johnson ] , we have @xmath221 the forgetful exact sequence @xmath222 induces the exact sequence @xmath223 they imply @xmath224 in particular , by theorem [ theorem : abel ] and lemma [ lemma : image of johnson ] , we obtain the following corollary . when @xmath42 , the induced homomorphism @xmath225 is injective .
we give another definition of the mod @xmath5 johnson homomorphism in the same manner as original one given by johnson @xcite .
the exact sequence @xmath228 induces the five term exact sequence between their homology groups with @xmath12-coefficients .
thus , we have @xmath229 since we have @xmath230}\otimes\mathbb{z}/2\mathbb{z},\ ] ] it suffices to show that @xmath231 by knneth formula , we have @xmath232 .
actually , this is an @xmath3-module isomorphism .
let @xmath51 be an arbitrary group defined by the exact sequence @xmath233 where @xmath234 and @xmath235 are the sets of generators and relators , respectively .
if we consider the five term exact sequence between their homology groups with @xmath12-coefficients , we have @xmath236r^2}.\ ] ] this can be considered as the hopf formula with @xmath12-coefficients .
recall that @xmath237 is a free group , and @xmath134 is normally generated by the only one relator @xmath238 .
if we put @xmath239 and @xmath240 , the cycle @xmath241 + \sum_{i=1}^{g-1}[\gamma_1 ^ 2\cdots\gamma_i^2|\gamma_{i+1}^2]\ ] ] represents a generator of @xmath242 corresponds to @xmath243 . under the homomorphism @xmath244
, it maps to @xmath245 .
thus , we have obtained the isomorphism as stated .
the isomorphism @xmath246)\otimes\mathbb{z}/2\mathbb{z}\to ( h^{\otimes2})^{s_2}/\braket{\omega}\ ] ] maps @xmath247 $ ] to @xmath248^{\otimes 2}$ ] for @xmath249 , where @xmath247 $ ] is the element in @xmath250\otimes\mathbb{z}/2\mathbb{z}$ ] represented by @xmath251 , and @xmath248\in h$ ] is the first homology class represented by @xmath249 . for @xmath252 ,
define a map @xmath253)\otimes\mathbb{z}/2\mathbb{z}$ ] by @xmath254 $ ] .
the map @xmath255 is a homomorphism , because for @xmath256 , we have @xmath257 = [ \varphi(\alpha)\alpha^{-1}]+[\alpha\varphi(\beta)\beta^{-1}\alpha^{-1 } ] = \tau_1(\varphi)(\alpha)+\tau_1(\varphi)(\beta).\ ] ] the proof is the same as that of ( * ? ? ?
* lemma 2c and 2d ) . for @xmath259 , @xmath260+[\varphi(\alpha)\alpha^{-1}].\ ] ] since the isomorphism in lemma [ lemma : mod2coinv ] is @xmath3-equivariant , @xmath102 acts on @xmath261)\otimes\mathbb{z}/2\mathbb{z}$ ] trivially .
thus , we have @xmath262=[\psi(\alpha)\alpha^{-1}]$ ] .
hence we obtain @xmath263 99 , on the homeotopy group of a non - orientable surface , math .
. camb .
soc . 71 ( 1972 ) , 437448 .
erratum : math .
soc . 136 ( 2004 ) , 441441 . ,
groupes et algbres de lie , ch .
2 , hermann , paris , 1972 . , irreducible sp - representations and subgroup distortion in the mapping class groups , comm . math .
86 no . 3 ( 2011 ) , 537556 . , a finite set of generators for the homeotopy group of a non - orientable surface , math .
soc . 65 ( 1969 ) , 409430 . , an abelian quotient of the mapping class group @xmath264 , math . ann . 249 ( 1980 ) , 225242 . , cohomological aspects of magnus expansions , preprint , math.gt/0505497 ( 2005 )
, first homology group of mapping class groups of nonorientable surfaces , math .
123 ( 1998 ) , 487 - 499 . ,
homeomorphisms of non - orientable two - manifolds , math .
59 ( 1963 ) , 307317 .
, on the homeomorphisms of a non - orientable surface , math .
61 ( 1965 ) , 6164 . , representing homology automorphisms of nonorientable surfaces , max planck inst .
preprint mpi / sfb 85 - 11 , revised version written on 26 feb 2004 available from ` http://www.math.msu.edu/~mccarthy/publications/selected.papers.html ` , filtration de johnson et groupe de torelli modulo @xmath265 , @xmath265 premier , c.r .
paris 346 no .
11 - 12 ( 2008 ) , 667670 .
, the abelianization of the level @xmath40 mapping class group , j. topology 3 no.4 ( 2010 ) , 847882 . ,
crosscap slides and the level 2 mapping class group of a nonorientable surface , geom .
dedicata 160 ( 2012 ) , 169183 , a finite generating set for the level 2 mapping class group of a nonorientable surface , kodai math .
j. 36 ( 2013 ) , 114 . | we construct a minimal generating set of the level 2 mapping class group of a nonorientable surface of genus @xmath0 , and determine its abelianization for @xmath1 . |
advances in nano - fabrication are increasingly enabling us to create nano - scale devices that exhibit non - classical or quantum - mechanical behaviour .
such quantum devices are of great interest as they may pave the way for a new generation of quantum technology with various applications from quantum metrology to quantum information processing .
however , to create quantum devices that perform useful functions , we must be able to understand their behaviour , and have effective means to controllably manipulate it .
analysis of system dynamics and the design of effective control strategies is almost impossible without the availability of sufficiently accurate mathematical models of the device . while these models should capture the essential features of the device , to be useful , they must also be computationally tractable , and preferably as simple as possible .
there are different approaches to deal with this problem .
one , which we shall refer to as a first - principles approach , involves constructing model hamiltonians based on reasonable assumptions about the relevant physical processes governing the behaviour of the system , making various simplifications , and solving the schrodinger equation in some form , usually using numerical techniques such as finite element or functional expansion methods .
the empirical approach , on the other hand , starts with experimental data and observations to construct a model of the system . in practice
both approaches are needed to deal with complex systems .
first - principle models are crucial to elucidate the fundamental physics that governs a system or device , but experimental data is crucial to account for the many unknowns that result in variability of the characteristics of the system , which is often difficult to predict or explain theoretically . for instance , ` manufactured ' systems such as artificial quantum dot ` atoms ' and ` molecules ' vary considerably in size , geometry , internal structure , etc , and sometimes even small variations can significantly alter their behaviour . in this paper we will focus on constructing mathematical models solely from experimental data without any prior knowledge of the specifics of the system , an approach one might call black - box system identification .
we will restrict ourselves to quantum systems whose essential features can be captured by low - dimensional hilbert space models .
for instance , a quantum dot molecule such as the silicon double quantum dot system in fig . [ fig : charge - qubit ] may consist of millions of individual atoms with many degrees of freedom , but the observable dynamics may be well described by overall charge distribution states in the quantum dot molecule . taking these charge distribution states as basis states for a hilbert space @xmath0 , the observable dynamics of the system is governed by a hamiltonian operator acting on @xmath0 , and possibly additional non - hermitian operators to account for dissipative effects if the system is not completely isolated from its environment . if the dimension of the hilbert space @xmath0 is huge then complete characterization of the hamiltonian and dissipation operators may be a hopeless task , but in certain cases the dynamics of interest takes place in low - dimensional hilbert space @xmath0 , or we may in fact wish to _ design _ a device whose dynamics can be described by a low - dimensional hilbert space model , as is the case in quantum information processing @xcite , where we desire to create systems that act as quantum bits , for example . the paper is organized as follows .
we start with a general description of the basic assumptions we make about the system to be modelled and the type of measurements and experiments upon which our characterization protocols are based . in section
[ sec : confinement ] we discuss the problem of subspace confinement , i.e. , how to experimentally characterize how well the dynamics of system is confined to a low - dimensional subspace such as a two - dimensional ( qubit ) hilbert space . in section [ sec : qubit_ident ] we present general protocols for identifying hamiltonian and decoherence parameters for controlled qubit - like systems .
we conclude with a discussion of generalizations and open problems .
we start with a generic model assuming only that the state of the system can be represented by a density operator ( i.e. , a positive operator of trace one ) @xmath1 acting on a hilbert space @xmath0 of dimension @xmath2 , and that its evolution is governed by the quantum liouville equation ( in units such that @xmath3 ) : @xmath4 + \l_d[\op{\rho}(t)],\ ] ] where @xmath5 is an effective hamiltonian and @xmath6 a super - operator that accounts for dissipative effects due to environmental influences , etc .
initially , all that is known about @xmath5 and @xmath6 is that @xmath5 is a hermitian operator on @xmath0 , and @xmath6 a completely positive super - operator acting on density matrices @xmath7 , although we may make some additional assumptions about the structure of @xmath5 or @xmath6 .
for instance , we shall generally assumes @xmath6 to be of kossakowski - sudarshan - lindblad form @xcite : @xmath8 = \sum_{k=1}^{n^2 - 1 } \gamma_k \d[\op{v}_k ] \op{\rho}(t)\ ] ] where the super - operators @xmath9 $ ] are defined by @xmath10b = a b a^\dagger - ( a^\dagger a b + b a^\dagger a)/2,\ ] ] @xmath11 are ( generally non - hermitian ) operators on @xmath0 and @xmath12 are positive real numbers .
if the system is subject to dynamic control , then the hamiltonian ( and sometimes the relaxation operators @xmath11 ) are not fixed but dependent on external control fields , which we shall denote by @xmath13 . in the simplest case
, we may assume a linear dependence of the hamiltonian on the controls @xmath14 = \op{h}_0 + \sum_{m=1}^m f_m(t ) \op{h}_m,\ ] ] where the @xmath15 are fixed hermitian operators on @xmath0 and the reservoir operators @xmath11 are constant .
the ultimate objective of characterization is to identify the operators @xmath16 $ ] ( or @xmath15 ) and @xmath6 by performing suitable experiments on the system .
this task would be simplified if we could assume that the system can be prepared in an arbitrary initial state @xmath17 , and that we can perform generalized measurements or projective measurements in arbitrary bases at any time but these requirements are generally unrealistic .
for example , our ability to perform measurements on the system prior to characterization is limited by the direct readout processes available , and often we only have a single , basic sensor such as a single electron transistor ( set ) @xcite providing relatively limited information about the charge distribution in a quantum dot molecule , or detection may be accomplished via a readout transition that involves coupling one state of the system to a fluorescent read - out state , e.g. , via a laser , etc .
furthermore , preparation of non - trivial states generally depends on knowledge of the operators @xmath16 $ ] and @xmath6 , the very information about the system we are trying to obtain .
these practical restrictions rule out conventional quantum state or process tomography techniques , which presume the ability to measure the system in different measurement bases and prepare it in different initial states to obtain sufficient information to reconstruct the quantum state or process @xcite . in this paper we consider a rather typical experimental scenario , where we are limited to measurements of a fixed observable and evolution under a hamiltonian that can be modified by varying certain control settings . for the most part ,
we restrict ourselves further to piecewise constant controls .
the only assumption on the measurement process we make is that it can be formally represented by some hermitian operator @xmath18 with @xmath2 eigenvalues corresponding to measurement outcomes @xmath19 , i.e. , that it has a spectral decomposition of the form @xmath20 this measurement also serves as initialization of the system as outcome @xmath19 means that the system will be left in an eigenstate @xmath21 associated with the eigenvalue @xmath19 .
if @xmath18 has @xmath2 unique eigenvalues , i.e. , all eigenvalues occur with multiplicity one , then the measurement is sufficient to initialize the system in a unique state ; if it has degenerate eigenvalues associated with eigenspaces of dimension greater @xmath22 then some measurement outcomes will not determine the state uniquely .
following the idea of intrinsic characterization , our objective is to extract information about the system without recourse to any external resources , i.e. , using no information from measurements other than the information provided by the sensors built into the device , and no external control fields except the ability to change the settings of the built - in actuators ( such as variable gate voltages ) subject to constraints .
although this requirement of characterization relying only on the built - in sensors and actuators may seem excessively restrictive , excluding many forms of spectroscopy , for example , it has the advantage of simplicity ( no external resources required ) .
moreover , external sensors and actuators may disturb the system , and thus characterization of the system in their presence may not yield an accurate picture of the dynamics in the absence of the additional apparatus .
we restrict ourselves here to systems sufficiently weakly coupled to a sufficiently large reservoir , whose dynamics can be described by adding a dissipation super - operator of the form ( [ eq : diss ] ) to the hamiltonian dynamics of the subsystem of interest . although many systems can be modelled this way , it should be noted that this approach has limitations .
for example , the dynamics of a subsystem @xmath23 that is strongly is coupled to a finite reservoir @xmath24 , such as single spin coupled to several nearby spins , can be very complicated and non - markovian .
although non - markovian dynamics can in principle be dealt with by allowing time - dependent relaxation operators @xmath11 , in such cases it is not always possible to describe the dynamics of the subsystem of interest in terms of the hamiltonian dynamics on the subspace and a set of simple relaxation or decoherence operators .
rather , it may become necessary to consider the system @xmath25 and characterize its dynamics , described by a schrodinger equation with a hamiltonian @xmath26 , instead to obtain an accurate picture of the subsystem dynamics .
the basic ideas of intrinsic characterization can be applied to this larger system , and the protocols we will describe can in principle be extended to such higher - dimensional systems , although full characterization of the system plus reservoir hamiltonian @xmath26 using only the built - in sensors and actuators may not be possible .
the degree of characterization possible will depend on the size of the hilbert space and the capabilities of the built - in sensors and actuators , e.g. , to discriminate and manipulate different states of the larger system . before attempting to identify the system hamiltonian ( and decoherence operators ) , an important first task is therefore estimating the dimension of the hilbert space in which the dynamics takes place .
a fundamental prerequisite for constructing a hilbert space model is knowledge of the underlying hilbert space .
this is a nontrivial problem as most systems have many degrees of freedom , and thus a potentially huge hilbert space , but effective characterization of the system often depends on finding a low dimensional hilbert space model that captures the essential features of the system .
furthermore , in applications such as quantum information processing the elementary building blocks are required to have a certain hilbert space dimension .
for example , for a quantum - dot molecule to qualify as a qubit , we must be able to isolate a two - dimensional subspace of the total hilbert space , and be able to coherently manipulate states within this 2d subspace without coupling to states outside the subspace ( leakage ) .
this requires several characterization steps : 1 .
isolation of a 2d subspace and characterization of subspace confinement ; 2 .
characterization of the hamiltonian dynamics including effect of the actuators ; and 3 .
characterization of ( non - controllable ) environmental effects ( dissipation ) .
the choice of a suitable subspace depends also on the measurement device as the measurement must be able to distinguish the basis states .
thus for a potential charge qubit device , for example , only a subspace spanned by charge states that can be reliably distinguished by the set is a suitable candidate , and ideally there should be two ( or in general @xmath2 ) orthogonal states that can be perfectly distinguished by the measurement , so that we have a well - defined reference frame ( basis ) @xmath27 , and the measurement can be represented by an observable of the form ( [ eq : a ] ) with @xmath2 distinct eigenvalues @xmath19 .
furthermore , we would like the readout process to act as a projective measurement , so that each measurement projects the state of the system onto one of the measurement basis states @xmath21 .
the ability to perform projective measurements is not a trivial requirement , especially for solid state systems , and is not an absolute necessity as characterization protocols can be adapted to weak measurements , but projective measurements can in principle be achieved with various sensors such as rf - sets , and we shall assume here that we are working in this regime .
after a possible subspace has been identified , it is crucial to check that the subspace is sufficiently isolated , i.e. , that we can reliably initialize the system in a state in this subspace and that it remains in this subspace under both free and controlled evolution .
this characterization of subspace confinement is very important . if the device is a candidate for a qubit , for example , it is essential that subspace leakage be much less than the rate of bit or phase flip errors due to imperfect control or decoherence as leakage or loss error correction protocols are considerably more demanding in terms of complexity and resources required .
characterization of subspace confinement depends on the characteristics of the sensors , i.e. , the type of measurements we can perform . for our charge qubit with set readout example , if the set can be calibrated to be sufficiently sensitive to enable detection of states outside the chosen subspace in addition to being able to discriminate the subspace basis states @xmath21 , i.e. , if the true measurement has ( at least ) @xmath28 mutually exclusive outcomes @xmath19 , @xmath29 and @xmath30 if the system state is outside the subspace , then the characterization of subspace confinement is relatively easy . if we perform @xmath31 experiments of the form 1 . * initialize : * measure and record outcome @xmath32 2 .
* evolve : * let the system evolve for time @xmath33 under some fixed hamiltonian @xmath34 3 .
* measure : * repeat measurement and record outcome @xmath35 and denote by @xmath36 the number of times the measurement results for the initial and final measurements were @xmath32 and @xmath35 , respectively , then , for @xmath31 sufficiently large , the leakage out of the subspace is approximately equal to the fraction of experiments for which the first measurement was @xmath19 with @xmath37 and the second measurement was @xmath28 , i.e. , @xmath38 repeating the experiments for different evolution times @xmath33 and averaging then gives an indication of the rate of subspace leakage , e.g. , if we estimate @xmath39 for @xmath40 , @xmath41 , then @xmath42 if the set ( or other measurement process ) is not sufficiently sensitive to reliably distinguish at least @xmath2 subspace basis states @xmath21 as well as states outside the subspace , then subspace characterization is more challenging . for instance , suppose we have a 2d subspace and a sensor that can reliably detect only one state , say @xmath43 , a common scenario for many systems where readout transitions are used that detect only a single state . in this case
the measurement outcomes are @xmath44 or @xmath45 .
if the measurement is projective and the dynamics confined to a 2d subspace , we can identify @xmath46 with outcome @xmath47 but this identification will lead to errors if the dynamics is not really confined to a 2d subspace .
however , even in this case we can still estimate the level of confinement to a 2d subspace under the evolution of a hamiltonian @xmath48 from observable coherent oscillations .
for example , let @xmath49 be the probability of obtaining outcome @xmath44 when measuring the time - evolved state @xmath50 .
if the system is initialized in the state @xmath43 and the dynamics of the system under @xmath48 is perfectly confined to a two - level subspace then conservation of probability implies that the heights @xmath51 and @xmath52 of the 0th and 1st order terms in the fourier spectrum of @xmath53 must satisfy @xmath54 .
we can use the deviation from this equality to bound the subspace leakage @xmath55 @xcite : @xmath56 as the upper and lower bounds depend only on the 0th and 1st order fourier peaks , they can usually be easily determined from experimental data although finite resolution due to discretization and projection noise will reduce the accuracy of the estimates @xcite .
moreover , in this case @xmath57 only indicates the confinement of the system to _ some _ two - level subspace , and we must be careful as for different controls @xmath58 the dynamics under @xmath48 may be confined to _
different _ two - level subspaces , in which case the system can not be considered a proper qubit .
for instance ,
for the 10-level test hamiltonian @xmath59 used for the simulated experiments in figs [ fig : leakage1 ] and [ fig : leakage2 ] , the first procedure measures the projection onto the subspace @xmath60 spanned by the measurement basis states @xmath61 and @xmath62 , while the second procedure measures the confinement of the dynamics to the best - fitting 2d subspace @xmath63 , which for the given hamiltonian is spanned by @xmath64 and comparison of the results confirms that the average confinement @xmath65 for the subspace @xmath60 as obtained by the first procedure is less ( @xmath66% ) than the confinement for subspace @xmath63 ( @xmath67% ) .
once a suitable subspace has been chosen , we can proceed to the second stage of the characterization , identification of the hamiltonian and decoherence operators .
the simplest type of system we can consider here is a qubit with hilbert space dimension @xmath68 . in this case ,
any non - trivial projective measurement , i.e. , any measurement with two distinguishable outcomes @xmath44 and @xmath47 can be represented by an observable @xmath69 .
the eigenstates @xmath43 and @xmath70 of @xmath18 define a basis for the hilbert space , and we can define the pauli operators @xmath71 ,
@xmath72 , @xmath73 and @xmath74 with respect to this basis @xmath75 or in the usual matrix notation @xmath76 taking ( without loss of generality ) the measurement outcomes to be @xmath77 and @xmath78 , we have @xmath79 in this basis . furthermore , we can expand any hamiltonian of the system as [ eq : h2 ] @xmath80 the @xmath81 term can generally be ignored as the identity @xmath71 commutes with all other pauli matrices and @xmath82 corresponds to multiplication by a global phase factor , i.e. , trivial dynamics .
hence , it suffices to determine the real vector @xmath83 , or in polar form , the angles @xmath84 to determine the hamiltonian . for a single hamiltonian
, we can furthermore choose the coordinate system such that @xmath85 .
however , since @xmath5 depends on control inputs @xmath58 , the parameters @xmath86 , @xmath87 and @xmath88 also depend on the controls @xmath58 , and we usually need to determine @xmath48 for many different @xmath58 . in this case
we can choose a reference hamiltonian , e.g. , @xmath89 , with @xmath85 , but we must determine the relative angles @xmath88 with respect to the reference hamiltonian for all other control settings . to relate the hamiltonian parameters to observable dynamics , it is instructive to visualize qubit states and their evolution on the bloch sphere .
if we define @xmath90 with @xmath91 for @xmath92 then it is easy to check that there is a one - to - one correspondence between density operators @xmath1 and points @xmath93 inside the closed unit ball in @xmath94 .
furthermore , the evolution of @xmath95 under the ( constant ) hamiltonian ( [ eq : h2 ] ) corresponds to a rotation of @xmath96 about the ( unit ) axis @xmath97 with angular velocity @xmath86 as illustrated in fig .
[ fig : bloch ] .
the figure also shows that we can determine the angle @xmath87 and angular frequency @xmath98 from the projection of @xmath96 under the rotation about the axis @xmath99 onto the @xmath100-axis .
as the figure shows , given @xmath101 , we can in principle extract @xmath86 and @xmath87 from the first minimum @xmath102 of @xmath101 , although in practice it is usually preferable to use fourier analysis or harmonic inversion techniques .
@xmath101 , being the expectation value @xmath103 $ ] of the observable @xmath79 in the state @xmath7 , can be obtained experimentally as follows @xcite : 1 .
* initialize : * measure and record outcome @xmath104 @xmath105 system in state @xmath106 .
evolve : * let the system evolve for time @xmath33 under the control settings @xmath58 @xmath105 system now in state @xmath107 with @xmath108 3 .
* measure : * repeat measurement and record outcome @xmath109 @xmath105 system now in state @xmath110 .
* repeat : * steps ( i)(iii ) @xmath31 times .
let @xmath36 be the number of times the the initial measurement produced outcome @xmath32 and the final measurement produced outcome @xmath35 .
then there are four possible combinations of outcomes @xmath111 , @xmath112 , @xmath113 and @xmath114 , which must add to the total number of experiments @xmath31 . the number of experiments that started in the state @xmath43 [ or @xmath115 is @xmath116 , and the fraction of experiments for which the second measurement yields @xmath44 conditioned on the system starting in the state @xmath43 is @xmath117 .
thus the ensemble average of @xmath79 at time @xmath33 , assuming we started in the state @xmath43 , is @xmath118 and for @xmath119 this relative frequency should approach the true expectation value of @xmath101 assuming @xmath120 .
thus , we can in principle determine @xmath101 to arbitrary accuracy by choosing @xmath31 large enough . by repeating the experiments for different evolution times @xmath121 , e.g. , using a stroboscopic mapping approach ( see fig . [
fig : stobo ] ) , we can determine @xmath101 as a function of @xmath33 , from which we can extract @xmath122 and @xmath123 as discussed .
for instance , noting that @xmath124 , fig . [ fig : leakage2 ] shows that the 1st order fourier peak @xmath125 of @xmath101 in this example is @xmath126 for @xmath127 and thus comparison with eq .
( [ eq : z ] ) shows that @xmath128 $ ] , or equivalently , @xmath129 .
these values are reasonably good estimates for the real values @xmath130 and @xmath131 , except for the sign of @xmath87 , or the orientation of the rotation axis , which we can not determine from the given data , as these rotations have identical projections onto the @xmath132-axis .
this is reflected in eq .
( [ eq : z ] ) by the fact that both coefficients @xmath133 and @xmath134 contain squares . in principle , we can determine the parameters @xmath86 and @xmath87 to arbitrary accuracy using this approach based on regular sampling and fourier analysis @xcite , although the total number of experiments can become prohibitively large .
an alternative that merits further investigation is the use of adaptive sampling techniques to reduce the total number of experiments necessary .
having determined the parameters @xmath135 and @xmath136 for a particular control setting @xmath58 and chosen a suitable reference hamiltonian @xmath89 with @xmath137 , to complete the characterization of @xmath138 , we must determine the horizontal angle @xmath139 .
the reference hamiltonian @xmath89 must not commute with the measured observable @xmath18 , or equivalently the angle @xmath140 between the rotation axis @xmath141 and the @xmath100-axis must be nonzero
. ideally @xmath141 should be as close to orthogonal to the @xmath100-axis as possible . assuming @xmath142 $ ] , and @xmath135 and @xmath136 are known
, we can determine the angles @xmath139 by performing the following experiments @xcite : 1 .
* initialize : * measure and record outcome @xmath104 @xmath105 system in state @xmath106 .
* prepare : * rotate around reference axis @xmath143 by angle @xmath144 , @xmath145 @xmath105 system in new state @xmath146 with @xmath147 .
evolve : * let the system evolve for time @xmath33 under the control settings @xmath58 @xmath105 system in new state @xmath148 with + [ eq : z2 ] @xmath149+d\sin(\omega_{\vec{f}}t)\\ c & = \sin\theta_{\vec{f } } \cos\theta_{\vec{f } } \cos(\phi_{\vec{f}}-\beta)\\ d & = \sin\theta_{\vec{f } } \sin(\phi_{\vec{f}}-\beta ) . \end{aligned}\ ] ] 4 .
* measure : * repeat measurement and record outcome @xmath109 @xmath105 system in new state @xmath150 . 5 .
* repeat : * steps ( i)(iv ) @xmath31 times . as before ,
if @xmath36 is the number of times the initial measurement produced outcome @xmath32 and the final measurement produced outcome @xmath35 then we have @xmath151 for sufficiently large @xmath31 , thus allowing us to determine @xmath101 experimentally .
since @xmath136 , @xmath135 and @xmath152 are known from the @xmath153 characterization step , eq .
( [ eq : z2 ] ) allows us to determine @xmath139 via the coefficients @xmath154 and @xmath155 , which can be determined in principle either through curve fitting or by taking the fourier transform of @xmath101 and noting that @xmath154 and @xmath155 correspond to the real and imaginary part of the 0th and 1st order fourier peaks . as a specific example , assume we have two hamiltonians @xmath156 and @xmath157 and we have already determined @xmath158 , @xmath159 , and @xmath160 , @xmath161 .
we choose @xmath156 as reference hamiltonian and note that a rotation about @xmath156 by @xmath162 maps the state @xmath163 to @xmath164 .
we then let state @xmath165 evolve for different length of time @xmath33 under hamiltonian @xmath157 , and determine the projection @xmath101 .
this will produce a measured @xmath132-trace as shown in fig .
[ fig : phi ] . eq .
( [ eq : z2 ] ) shows that in theory the height of the real part of the 0th and 1st order fourier peaks should be the same except for the sign , but the graph clearly shows that this is not the case for our noisy ( simulated ) experimental data .
in fact , the 1st order peaks around @xmath166 are significantly broadened , while the @xmath167-peak is very narrow .
this suggests that we are likely to obtain the best estimate for @xmath88 by setting @xmath168(0)$ ] and @xmath169 + \beta$ ] , where @xmath170 in our case .
indeed , this yields an estimate of @xmath171 , which is very close to actual value @xmath172 used in the simulation .
note that using the 1st order fourier peaks gives far less accurate results due to the significant broadening of the peaks .
this can be reduced using phase matching conditions and other tricks @xcite , but in this case simply using only the 0-peak is sufficient to get a good estimate . in some settings
it is sufficient to characterize the hamiltonians @xmath48 for a discrete , finite set of controls @xmath173
. this would be the case for bang - bang - type control schemes that require only switching between a finite set of control hamiltonians .
in other cases , in particular when continuously varying fields are to be applied , we would also like to know the functional dependence of the hamiltonian on the control fields , e.g. , whether we can assume a control - linear model as in ( [ eq : hf ] ) , or if there are non - linear or crosstalk effects , etc .
although characterizing the control dependence of the hamiltonian is a non - trivial problem , given sufficiently many @xmath48 data points , it is possible to test the quality of fit of a particular model using statistical means . for instance , suppose we have characterized the hamiltonians @xmath48 for @xmath174 .
let @xmath175 be the corresponding vectors in @xmath94 as defined above , and let @xmath176 be the @xmath177th component of the control vector @xmath178 .
we can find the best - fitting linear model by finding vectors @xmath179 , @xmath180 , where @xmath181 is the number of independent control variables , that minimize the norm of the residuals @xmath182 where @xmath183 the norm of the residuals @xmath184 then gives an indication of the quality of fit of a linear dependence model , and allows us to compare the quality of fit for different models .
as an example , consider a hamiltonian @xmath48 depending on a one - dimensional control input @xmath185 . to characterize the functional dependence of @xmath186 on @xmath187 , we choose a finite set of control settings @xmath188 for @xmath189 .
we characterize first the rotation frequencies @xmath190 and declination angles @xmath191 for each @xmath192 , and then the relative angles @xmath193 as discussed above . from these , we calculate the components @xmath194 , @xmath195 and @xmath196 of the hamiltonian according to eq .
( [ eq : h2 ] ) , and plot them versus @xmath187 as shown in fig .
[ fig : fdepend ] .
specifically , in our case we find @xmath197 and hence @xmath198 , @xmath199 appears to be constant and @xmath200 exhibits a quadratic dependence on the control @xmath187 .
a least square fit of the data using a linear and quadratic fitting function , respectively , yields @xmath201 and @xmath202 , which is in very good agreement with the actual functional dependence in the model @xmath203 and @xmath204 despite the fact that the simulated data appears very noisy and the @xmath87 and @xmath86 estimates did not employ tuning such as phase matching to further improve the accuracy .
the protocols for hamiltonian identification above are sufficient if dissipative effects are small on the timescales for coherent control , i.e. , if the system and environment coupling is much weaker than the system - control interaction .
if the coupling between the system and its environment is stronger then dissipative effects must be directly incorporated into the characterization protocols .
assuming markovian decoherence of the form ( [ eq : diss ] ) , it is easy to show that for @xmath205 , we have @xmath206 , where @xmath207 are the elementary relaxation operators @xmath208 defined in terms of the pauli matrices @xmath209 , and @xmath210 is some unitary operator in @xmath211 that defines a preferred decoherence basis . the operator @xmath74 is usually associated with pure decoherence ( pure dephasing ) while the raising and lowering operators @xmath212 are associated with population relaxation from @xmath70 to @xmath43 and vice versa . very often it is furthermore assumed that @xmath213 is the identity , i.e. , that the preferred basis for relaxation is the same as the measurement basis .
although this need not be the case , it is often a good approximation .
the main effect of relaxation processes is to dampen the observed coherent oscillations @xmath101 that result from the hamiltonian dynamics , and try to force the system ` asymptotically ' into a ` steady ' state .
this effect manifests itself in lorentzian broadening of the peaks in the oscillation spectrum , which can be used for characterization purposes .
for example , fig .
[ fig : decoherence ] shows significantly damped oscillations as a result of relaxation .
the fact that the oscillations decay to @xmath214 suggests either pure dephasing , or a symmetric relaxation process with @xmath215 , i.e. , equal probability of relaxation from state @xmath43 to @xmath70 and vice versa , or a combination of both . since symmetric relaxation and pure dephasing
have very similar signatures , it can be difficult to distinguish these processes from the oscillation trace alone . however , if there exists a control @xmath58 such that @xmath216 , which corresponds to a rotation about the @xmath100 ( or measurement ) axis , then we can in principle distinguish the two processes by initializing the system repeatedly in either state @xmath43 or @xmath70 and letting the system evolve under @xmath48 for various periods of time before measuring again .
if the decoherence process is pure dephasing then the system will remain in the initial state , while we expect to observe random jumps ( with equal probability ) for symmetric relaxation . assuming this preliminary characterization step suggests that the damping in the figure is due to pure dephasing
, we can then extract information about the rotation frequency @xmath217 and dephasing rate @xmath218 by fitting a lorentzian envelope function @xmath219 to the first - order peak in the fourier spectrum as shown in the figure .
in our example , using a simple least - squares minimization yields estimated values for both @xmath217 and @xmath218 that are close to the actual values used in the simulated experiment .
this approach can be generalized for more complicated relaxation processes @xcite . alternatively ,
if we can characterize the hamiltonian dynamics sufficiently to be able to initialize the system in non - measurement basis states and simulate readout in different bases , then generalized lindblad operators can be estimated using repeated process tomography @xcite .
we have described some basic tools and techniques for characterizing the extent to which the dynamics of a certain system is confined to a low - dimensional subspace of a potentially large hilbert space , as well as for identifying both hamiltonian and decoherence parameters experimentally , with emphasis on protocols that are realistic even in the presence of limited direct readout and control capabilities .
although the explicit schemes presented focused on qubit - like systems , the same basic ideas can be applied to higher - dimensional systems .
however , the number of hamiltonian and decoherence parameters to be determined for higher - dimensional system makes finding explicit protocols for complete characterization quite challenging .
the qubit characterization protocols show that identifying the relative angles between several control hamiltonians requires a two - stage process and more complicated two - step experiments ( initialization in non - measurement basis state followed by controlled evolution ) than identifying the rotation frequency and declination angle for a single hamiltonian . as the dimension of
the system increases more parameters are necessary to fully characterize the hamiltonians , leading to more complex multi - stage characterization protocols .
another problem is that the number of experiments required to implement such schemes may become prohibitively large for parameter estimation based on spectroscopic analysis .
one of the most promising strategies to avoid this problem is the development of efficient adaptive characterization schemes as opposed to protocols based on regular sampling to minimize the number of measurements / experiments required . for systems composed of separate smaller units such as multi - qubit systems , boot - strapping approaches to characterization may also provide a fruitful alternative .
for example , assuming we have separately characterized individual qubits using qubit identification protocols and therefore have full local control , we can reduce the problem of two - qubit identification in principle to identifying the interaction hamiltonian . for two - qubit systems with local control and a fully non - local interaction hamiltonian
this can be achieved using entanglement mapping or concurrence spectroscopy @xcite . | we consider the task of intrinsic control system identification for quantum devices .
the problem of experimental determination of subspace confinement is considered , and simple general strategies for full hamiltonian identification and decoherence characterization of a controlled two - level system are presented . |
the critical point terminating a first order phase transition line in the phase diagram of qcd at high temperature and density is one of the most characteristic features that may be discovered in heavy - ion collision experiments . to understand the phase structure , first principle calculations of qcd by numerical simulations are very important .
one of the interesting approaches to find a first order phase transition is to construct the canonical partition function @xmath0 by fixing the total quark number @xmath3 or quark number density @xmath4 . from the canonical partition function , one can estimate the quark number giving the largest contribution to the grand partition function @xmath5 . because two different states coexist at a first order transition point , two different quark numbers give equally large contributions simultaneously if the transition is of first order .
the canonical partition function is defined by a fugacity expansion of @xmath5 , @xmath6 where @xmath7 , @xmath8 and @xmath9 are the quark determinant , the gauge action and the number of flavor , respectively .
the term @xmath10 can be regarded as the probability distribution of the quark number @xmath11 .
moreover , it is worth introducing an effective potential @xmath12 as a function of @xmath11 , @xmath13 where @xmath14 is the helmholtz free energy . in a first order phase transition region , this effective potential has minima at more than one value of @xmath11 . at the minima ,
the derivative of @xmath12 vanishes : @xmath15 hence , in the first order transition region , we expect @xmath16 takes the same value at different @xmath11 . here , @xmath17 is the chemical potential which gives the effective potential at a minimum point @xmath18 and becomes @xmath19 in the thermodynamic limit .
plane and the behavior of @xmath20 as a function of @xmath21 . , title="fig:",width=355 ] -0.2 cm -0.3 cm the phase structure in the @xmath22 plane is sketched in the left panel of fig .
[ fig : cano ] .
the thick line is the phase transition line .
we expect that the transition is crossover at low density and becomes of first order at high density . in the first order transition region , the two coexisting states are mixed .
the region between the two thick lines is the mixed state .
the expected behavior of @xmath23 along the lines a and b are shown in the right figure .
when the temperature is higher than the temperature at the critical point @xmath24 ( line a ) , @xmath23 increases monotonically as the density increases
. however , for the case below @xmath25 ( line b ) , this line crosses the mixed state .
corresponding to the double - well potential in a finite volume , @xmath23 is expected to be an s - shaped function . in the infinite volume limit
, @xmath23 does not increase in the region between @xmath26 and @xmath27 , since the surface energy between the two states can be ignored .
the glasgow method @xcite has been a well - known method to compute the canonical partition function .
a few years ago , the above mentioned behavior at a first order phase transition was observed in 4 flavor qcd with staggered fermions calculating the quark determinant by the glasgow algorithm on a small lattice @xcite .
also , simulations with a fixed quark number , i.e. in the canonical ensemble , have been tried @xcite
. however , the calculations by the glasgow algorithm and the simulations for a canonical ensemble require large computational cost and are difficult except on a small lattice with present day computer resources . in this report
, we propose a method based on a saddle point approximation @xcite . by this approximation ,
the computational cost is drastically reduced and the first order like behavior was observed for 2 flavor qcd .
from eq . ( [ eq : cpartition ] ) , the canonical partition function can be obtained by an inverse laplace transformation , @xmath28 where @xmath29 is an appropriate real constant and @xmath30 is a real variable .
we have used the fact that @xmath31 for any complex @xmath19 .
the grand partition function can be evaluated by the following expectation value at @xmath32 .
@xmath33 we apply a saddle point approximation to evaluate eq .
( [ eq : canonicalp ] ) , which reduces the computational cost . if one selects the @xmath29 at a saddle point in eq .
( [ eq : canonicalp ] ) , the necessary information is limited that around the saddle point when the volume is sufficiently large . moreover , if we restrict ourselves to study the low density region , the value of @xmath34 near the saddle point can be estimated by a taylor expansion around @xmath32 .
the calculation by the taylor expansion is much cheaper than the exact calculation and the study using a large lattice is possible .
also , the truncation error can be systematically controlled by increasing the number of the expansion coefficients .
we assume the existence of a saddle point @xmath35 in the complex @xmath36 plane for each configuration , which satisfies @xmath37 . here , the quark number density in a lattice unit and a physical unit are @xmath38 and @xmath39 , respectively , @xmath40 $ ] and @xmath41 .
we then perform a taylor expansion around the saddle point and obtain the canonical partition function , @xmath42 dx \right\rangle_{(t , \mu_q=0 ) } \nonumber \\ & \approx & \frac{3}{\sqrt{2 \pi } } { \cal z}_{\rm gc } ( t,0 ) \left\langle \exp \left [ v \left ( d(z_0 ) - \bar{\rho } z_0 \right ) \right ] e^{-i \alpha/2 } \sqrt { \frac{1}{v |d''(z_0)| } } \right\rangle_{(t , \mu_q=0 ) } , \label{eq : zcspa}\end{aligned}\ ] ] where @xmath43 @xmath44 and @xmath45 .
higher order terms in the expansion of @xmath46 are negligible when the volume @xmath47 is sufficiently large . within the framework of the saddle point approximation ,
the derivative of @xmath48 with respect to @xmath11 or @xmath21 can be evaluated by @xmath49 e^{-i \alpha /2 } \sqrt { \frac{1}{v |d''(z_0)| } } \right\rangle_{(t , \mu_q=0 ) } } { \left\langle \exp \left [ v \left ( d(z_0 ) - \bar{\rho } z_0 \right ) \right ] e^{-i \alpha /2 } \sqrt { \frac{1}{v |d''(z_0)| } } \right\rangle_{(t , \mu_q=0)}}. \label{eq : chemap}\end{aligned}\ ] ] this formula is similar to that of the reweighting method for finite @xmath19 .
the operator in the denominator corresponds to a reweighting factor , and @xmath50 is an expectation value of the saddle point calculated with this modification factor . with a saddle point approximation .
, title="fig:",width=259 ] -0.2 cm -0.3 cm
we compute the derivative of @xmath48 using the data obtained in @xcite with the 2 flavor p4-improved staggered quark action , @xmath1 . because the operators in eq .
( [ eq : chemap ] ) are complex , this calculation suffers from the sign problem . to eliminate the sign problem ,
the approximation proposed in @xcite is used : if one assumes that the distribution of the complex phase is well - approximated by a gaussian function , the complex phase factor @xmath51 can be replaced by @xmath52 $ ] .
we estimate the quark determinant by the taylor expansion up to @xmath53 . because the calculation of eq .
( [ eq : chemap ] ) is similar to the calculation by the reweighting method , the configurations which give important contribution are changed by the modification ( reweighting ) factor . to avoid this problem , we use the multi-@xmath54 reweighting method . by this method ,
the important configurations are automatically selected among all configurations generated at many simulation points of @xmath55 .
the details are given in @xcite .
the result of @xmath20 is shown in fig .
[ fig : chem ] as a function of @xmath56 for each temperature @xmath57 .
@xmath58 is the pseudo - critical temperature at @xmath32 .
the dot - dashed line is the value of the free quark - gluon gas in the continuum theory , @xmath59 $ ] . from this figure
, we find that a qualitative feature of @xmath20 changes around @xmath60 , i.e. @xmath20 increases monotonically as @xmath21 increases above 0.8 , whereas it shows an s - shape below 0.8 .
this s - shape is a signature of a first order phase transition . with some approximations ,
the critical value of @xmath20 is is estimated to be about @xmath61 , which is roughly consistent with the critical point estimated in @xcite by calculating the effective potential of the plaquette using the same configurations , @xmath62 .
the difference between these two results may include a systematic error .
our result strongly suggests the existence of the critical point terminating the first order phase transition line in the @xmath2 phase diagram of qcd .
further studies are necessary to predict the critical point quantitatively .
gibbs , _ phys .
b _ * 172 * ( 1986 ) 53 , i.m .
barbour , s.e .
morrison , e.g. klepfish , j.b .
kogut , and m.p .
lombardo , _ phys .
d _ * 56 * ( 1997 ) 7063 ; _ nucl . phys
. b(proc . suppl . )
_ * 60a * ( 1998 ) 220 ; a. hasenfratz and d. toussaint , _ nucl .
b _ * 371 * ( 1992 ) 539 .
a. alexandru , m. faber , i. horvath , and k .- f .
liu , _ phys .
* 72 * ( 2005 ) 114513 , a. li , x. meng , a. alexandru , and k .- f .
pos_(lattice 2008 ) ( 2008 ) 178 .
s. ejiri , _ phys .
* 78 * ( 2008 ) 074507 . | we propose a method to find the qcd critical point at finite density calculating the canonical partition function @xmath0 by monte - carlo simulations of lattice qcd , and analyze data obtained by a simulation with two - flavor p4-improved staggered quarks with pion mass @xmath1 .
it is found that the shape of an effective potential changes gradually as the temperature decreases and a first order phase transition appears in the low temperature and high density region .
this result strongly suggests the existence of the critical point in the @xmath2 phase diagram . |
the @xmath0 decay into @xmath5 is an example of isospin violation .
the @xmath0 has isospin @xmath6 , spin @xmath7 , and hence it does not couple to the @xmath5 system in the isospin limit , which implies the rule @xmath8 .
the experimental situation on this decay is rather confusing .
there are two older results whose central values are very different but their quoted errors are so big that both were still compatible : the first one from @xcite gives @xmath9 .
the second one from @xcite provides @xmath10 . very recently
two new , more precise , but conflicting results have been reported from the two experiments at the vepp-2 m in novosibirsk : the cmd-2 collaboration reports a value @xmath11 @xcite whereas the snd collaboration @xcite obtains @xmath12 .
isospin violation has become a fashionable topic in chiral perturbation theory ( @xmath13 ) @xcite but the @xmath14 decay is however unreachable with plain @xmath13 , since it involves the propagation of the pair of pions around 1 gev , far away from the @xmath13 applicability range .
nevertheless , new nonperturbative schemes imposing unitarity and still using the chiral lagrangians have emerged enlarging the convergence of the chiral expansion . in @xcite
the inverse amplitude method ( iam ) is used in one channel and good results are obtained for the @xmath15 , @xmath16 and @xmath17 regions , amongst others , in @xmath18 and @xmath19 scattering . in @xcite the method is generalized to include coupled channels and one is able to describe very well the meson - meson scattering and all the associated resonances up to about 1.2 gev .
a more general approach is used in @xcite by means of the n / d method , in order to include the exchange of some preexisting resonances explicitly , which are then responsible for the values of the parameters of the fourth order chiral lagrangian . here
we shall follow the work @xcite since it provides the most complete study of the different meson - meson scattering channels , including the mesonic resonances and their properties up to 1.2 gev .
in particular , this method yields a resonance in the @xmath20 channel , the @xmath21 resonance , related to the @xmath0 , and this allows us to obtain the strong contribution to the @xmath14 decay .
we also consider electromagnetic contributions at tree level which turn out to be dominant and were already considered in @xcite . in order to calculate the contribution of an intermediate photon to the @xmath14 decay ,
let us consider the effective lagrangian for vector mesons presented in @xcite , which is written in terms of the su(3 ) pseudoscalar meson matrix @xmath22 and the antisymmetric vector tensor field @xmath23 defined in @xcite @xmath24=\frac{f_v}{2\sqrt{2 } } \langle v_{\mu\nu}f_+^{\mu\nu}\rangle+ \frac{i\,g_v}{\sqrt{2}}\langle v_{\mu\nu}u^\mu u^\nu\rangle , \label{lag}\ ] ] where
@xmath25 " indicates the su(3 ) trace . in order to introduce the physical states
@xmath0 and @xmath26 , we assume ideal mixing between the @xmath27 and @xmath21 vector resonances and hence taking into account that the @xmath27 does not couple to pairs of mesons at the order of eq . ( 1 ) , the coupling of the @xmath0 is easily deduced from that of the @xmath21 by simply multiplying the results of the @xmath21 by the factor @xmath28 . with these ingredients and the standard @xmath29 coupling we can write the contribution of a feynman diagram with the @xmath0 going to a photon which then couples to a pair of pions , and which is given by @xmath30 where @xmath31 and @xmath32 are , respectively
, the momenta of positive and negative pions and @xmath33 is the pion electromagnetic form factor , which at the @xmath0 mass is given by @xmath34 .
this can be compared with the coupling of the @xmath0 to @xmath35 , or @xmath36 , which can be obtained from the @xmath37 term in eq .
( [ lag ] ) and reads @xmath38 which provides the right @xmath0 decay width with a value of @xmath39 mev . by analogy to eq .
( [ lphikk ] ) , eq .
( [ lphipipi ] ) gives a @xmath0 coupling to @xmath1 @xmath40 which provides the @xmath41 decay width with the tree level photon mechanism . with a value of @xmath42 mev from the @xmath43 decay @xcite and using the coupling of eq .
( [ g8phikk ] ) one obtains a branching ratio to the total @xmath0 width of @xmath44 . in order to evaluate the strong contribution to the process we consider the @xmath45 amplitude corrected from isospin violation effects due to quark mass differences .
the method used is based on the chiral unitary approach to the meson - meson interaction followed in @xcite .
the technique starts from the @xmath46 and @xmath47 @xmath13 lagrangian and uses the iam in coupled channels , generalizing the one channel version of the iam developed in @xcite . within the coupled channel formalism ,
the partial wave amplitude is given in the iam by the matrix equation @xmath48^{-1}\,t_2 , \label{iam}\ ] ] where @xmath49 and @xmath50 are @xmath46 and @xmath47 @xmath13 partial waves , respectively . in principle
@xmath50 would require a full one - loop calculation , but it was shown in @xcite that it can be very well approximated by @xmath51 where @xmath52 is the tree level polynomial contribution coming from the @xmath53 chiral lagrangian and @xmath54 is a diagonal matrix for the loop function of the intermediate two meson propagators which are regularized in @xcite by means of a momentum cut - off . in the present case , in which isospin is broken explicitly and @xmath7
, we are dealing with three two - meson states : @xmath55 , @xmath56 and @xmath5 , that we will call 1 , 2 and 3 , respectively .
the amplitude is a @xmath57 matrix whose elements are denoted as @xmath58 .
the @xmath49 and @xmath52 amplitudes used in the present work and calculated in the isospin breaking case , are collected in the appendix of @xcite .
the fit of the phase shifts and inelasticities is carried out here in the isospin limit , as done in @xcite and there are several sets of @xmath59 coefficients which give rise to equally acceptable fits .
we write in table [ tabla2 ] the values of the coefficients of the different sets of chiral parameters .
the corresponding results for the phase shifts and inelasticities can be seen in @xcite where it is shown that the small differences in the results appear basically only in the @xmath60 and @xmath61 resonance regions , where data have also larger errors or are very scarce . in order to evaluate the contribution to the @xmath62 coupling from the strongly interacting sector
we evaluate the @xmath63 amplitude ( @xmath64 ) and the @xmath65 amplitude ( @xmath66 ) near the pole of the @xmath21 resonance which in our case appears around @xmath67 mev .
close to the @xmath21 pole the amplitudes obtained numerically are then driven by the exchange of an @xmath21 . by assuming a coupling of the type of eq .
( [ lphikk ] ) for the @xmath21 to @xmath55 and @xmath5 , these two amplitudes , close to the @xmath21 pole , are given by @xmath68 where @xmath69 is the three - momentum of the @xmath70 particle in the cm frame . by looking at the residues of the amplitudes @xmath64 , @xmath66 in the @xmath21 pole
we can get the products @xmath71 and @xmath72 .
thus , defining @xmath73 we obtain the ratio of the @xmath74 to @xmath74 by means of the ratio of @xmath75 to @xmath76 , and hence taking @xmath74 from eq .
( [ lphikk ] ) , we get the value for @xmath77 .
then , by adding the above contribution with that of eq .
( [ g8phikk ] ) we can obtain the @xmath78 decay width .
we have taken @xmath79 , as demanded by vector meson dominance @xcite . &
@xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 & @xmath85 & br@xmath86 + set 1&0.91&1.61&-3.65&-0.25&1.07&0.58&-0.4&666 mev & 1.3@xmath87 + set 2&0.91&1.61&-3.65&-0.25&1.07&0.58&0.05&751 mev & 1.0@xmath87 + set 3&0.88&1.54&-3.66&-0.27&1.09&0.68&0.10&673 mev & 1.3@xmath87 + & @xmath88 & @xmath89 & @xmath90 & @xmath91 & @xmath92 & @xmath93 & @xmath94 & @xmath95 & + chpt&0.4&1.4&-3.5&-0.3 & 1.4&0.5&-0.4&@xmath96 & ' '' '' + ref.@xcite&@xmath970.3&@xmath970.3&@xmath971.1&@xmath970.5&@xmath970.5&@xmath970.3 & @xmath970.2 & & + _ _ each set of chiral parameters has then been used in the isospin - breaking amplitudes given in the appendix of @xcite , obtaining a value of @xmath98 given in table [ tabla2 ] .
the dispersion of the results provides an estimate of the systematic theoretical uncertainties . from table
[ tabla2 ] , we obtain , after taking into account the strong contributions @xmath99 on the other hand , explicit calculations of the absorptive part of the @xmath100 intermediate channel @xcite give a contribution of about 1/4 of the kaon loops . in order to estimate the uncertainties from neglecting the photonic loops we take a conservative estimate and consider them of the same magnitude as the strong interaction correction , and , hence , add an extra @xmath101 uncertainty . adding in quadrature the errors from the different sources ,
our final result is the band of values : @xmath102 which is compatible with the present pdg average within errors and lies just between the results of the two recent experiments , which are much more precise , but mutually incompatible .
the @xmath0 meson can not decay into two pions or @xmath103 in the isospin limit .
the decay into two neutral pions is more strictly forbidden by symmetry and the identity of the two pions . as a consequence
the decay of the @xmath0 into @xmath2 and @xmath3 is forbidden at tree level .
however , the @xmath0 decays into two kaons and the processes described can proceed via the loop diagrams depicted in fig . 1 where the intermediate states in the loops stand for @xmath104 .
the evaluation of the diagrams of fig . 1 is done in @xcite .
the terms with @xmath37 of eq .
( 2 ) contribute to all the diagrams in the figure .
however , the @xmath105 term of eq .
( 2 ) only contributes to the diagrams containing the contact vertex @xmath106 , like diagrams ( a ) , ( e ) .
the idea follows closely the work of @xcite but for the treatment of the final state interaction of the mesons one uses here the norperturbative chiral techniques . in this case for l=0 , which is the only partial wave needed
, one can use the results of @xcite , where it is proved that the use of the bethe salpeter equation in connection with the lowest order chiral lagrangian and a suitable cut off in the loops gave a good description of the meson meson scalar sector .
furthermore , in @xcite it was proved that the meson meson amplitude in those diagrams factorized on shell .
the loops of type ( a ) , ( b ) and ( c ) can be summed up using arguments of gauge invariance following the techniques of @xcite and lead to a finite amplitude . on the other hand , the terms involving @xmath105 and a remnant momentum dependent term from the @xmath37 lagrangian in eq .
( 2 ) only appear in the contact vertex @xmath106 , and the diagrams of type ( b ) , ( c ) are now not present .
hence , in this case the only loop function involved is the one of two mesons which is regularized as in @xcite for the problem of the meson meson scattering .
the average over polarization of the @xmath0 for the modulus square of @xmath107 matrix is then easily written and for the case of @xmath2 decay one finds @xmath108 for the @xmath109 case we have @xmath110 where @xmath111 and @xmath112 are the loop functions mentioned above . we have evaluated the invariant mass distribution for these decay channels and in fig . 2 we plot the distribution @xmath113 for @xmath114 which allows us to see the @xmath115 contribution since the @xmath116 is the important scalar resonance appearing in the @xmath117 amplitude @xcite .
the results are obtained using @xmath37=55 mev and @xmath105=165 mev , which are suited to describe the @xmath104 and @xmath118 decay of the @xmath0 .
the solid curve shows our prediction , with @xmath119 , the sign predicted by vector meson dominance , as we quoted above .
the dashed curve is obtained considering @xmath120 .
in addition we show also the results of the intermediate dot - dashed curve which correspond to taking for @xmath37 and @xmath105 the parameters of the @xmath16 decay , @xmath37=69 mev and @xmath105=154 mev .
we compare our results with the recent ones of the novosibirsk experiment @xcite .
we can see that the shape of the spectrum is relatively well reproduced considering statistical and systematic errors ( the latter ones not shown in the figure ) .
the results considering @xmath120 are in complete disagreement with the data .
the finite total branching ratio which we find for the @xmath114 decay is @xmath121 , which is slightly smaller than the result given in @xcite , @xmath122 , where the first error is statistical and the second one systematic .
the result given in @xcite is @xmath123 , compatible with our prediction .
should we use the values for @xmath105 and @xmath37 of the @xmath16 decay we would obtain @xmath44 . the branching ratio obtained for the case
@xmath109 is @xmath124 .
the results obtained at novosibirsk are @xcite @xmath125 and @xcite @xmath126 .
should we use the values for @xmath105 and @xmath37 of the @xmath16 decay we would obtain @xmath127 .
the spectrum , not shown , is dominated by the @xmath128 contribution .
the results reported here are two examples of the successful application of the chiral unitary techniques .
a recent review of multiple applications of these methods can be seen in @xcite .
we would also like to acknowledge financial support from the dgicyt under contracts pb96 - 0753 and aen97 - 1693 and from the eu tmr network eurodaphne , contract no .
erbfmrx - ct98 - 0169 .
i. b. vasserman et al .
b * 99 * , 62 ( 1981 ) . v. b. golubev et al .
phys . * 44 * , 409 ( 1986 ) .
g. v. fedotovich for the cmd-2 collaboration .
talk given at the 8th international conference on hadron spectroscopy .
hadron 99 , 24 - 28 august 1999 , beijing china .
s. burdin for the snd collaboration .
talk given at the 8th international conference on hadron spectroscopy .
hadron 99 , 24 - 28 august 1999 , beijing china .
s. weinberg , physica a * 96 * , 327 ( 1979 ) .
j. gasser annd h. leutwyler , ann .
phys . * 158 * , 142 ( 1984 ) ; j. gasser and h. leutwyler , nucl .
b * 250 * , 465 , 517 , 539 ( 1985 ) .
t. n. truong , phys . rev .
661 , 2526 ( 1988 ) ; phys . rev
. lett .
67 , 2260 ( 1991 ) ; a. dobado , m. j. herrero and t. n. truong , phys .
b * 235 * , 134 ( 1990 ) ; a. dobado and j. r. pelez , phys .
d * 47 * , 4883 ( 1993 ) ; phys .
d * 56 * , 3057 ( 1997 ) .
j. a. oller , e. oset and j. r. pelez , phys .
, 3452 ( 1998 ) ; phys .
d * 59 * , 074001 ( 1999 ) ; erratum - ibid .
d * 60 * , 099906 ( 1999 ) .
f. guerrero and j. a. oller , nucl .
b * 537 * , 459 ( 1999 ) .
j. a. oller and e. oset , phys .
d * 60 * , 074023 ( 1999 ) .
a. bramon and a. varias , phys .
d * 20 * , 2262 ( 1979 ) .
h. genz and s. tatur , phys .
d * 50 * , 32563 ( 1994 ) .
g. ecker , j. gasser , a. pich and e. de rafael , nucl .
b * 321 * , 311 ( 1989 ) ; g. gasser , h. leutwyler , a. pich and e. de rafael , phys .
b * 223 * , 425 ( 1989 ) .
j. oller , e. oset and j. r. pelez , submitted to phys . rev .
d , hep - ph/9911297 .
j. bijnens , g. ecker and j. gasser , _ chiral perturbation theory in the da@xmath22ne physics handbook _
( second edition ) , eds .
l. maiani , g. panchieri and n. paver ( frascati ) .
e. marco , s. hirenzaki , e. oset and h. toki , to appear in phys .
b , hep - ph/9903217 .
a. bramon , a. grau and g. pancheri , phys .
b * 289 * , 97 ( 1992 ) .
j. a. oller and e. oset , nucl .
phys . a * 620 * , 438 ( 1997 ) ; erratum nucl . phys .
a * 624 * , 407 ( 1999 ) .
j. a. oller , phys .
b * 426 * , 7 ( 1998 ) . f. e. close , n. isgur and s. kumano , nucl .
b * 389 * , 513 ( 1993 ) .
m. n. achasov et al . , phys .
b * 440 * , 442 ( 1998 ) .
r. r. akhmetshin et al .
, hep - ph/9907006 . m. n. achasov et al . , phys .
b * 438 * , 441 ( 1998 ) .
r. r. akhmetshin et al .
b * 415 * , 452 ( 1997 ) .
j. a. oller , e. oset and a. ramos , to be published in prog . part .
* 45 * , issue 1 . | we report on recent results on the decay of the @xmath0 into @xmath1 and @xmath0 radiative decays into @xmath2 and @xmath3 , which require the consideration of the final state interaction of a pair of mesons in a region inaccessible to chiral perturbation theory . by using nonperturbative chiral unitary methods for the meson meson interaction
we can obtain the corresponding decay widths and the results are compared with recent experimental data .
= 11.6pt |
warm inflation is known as the scenario of dissipative inflaton motion@xcite . in this paper
we address the generic situation where the inflaton field , which couples to a massive trigger field , dissipates its energy into light fields during inflation * without * introducing additional interactions .
namely , a trigger field @xmath1 coupled to the inflaton @xmath2 with the interaction @xmath3 obtains large mass @xmath4 during inflation , and decays ( at least gravitationally ) into light matter with a decay rate @xmath5 the condition for fast decay ( @xmath6 ) is trivial .
we define the critical value @xmath7 by @xmath8 and find that the decay into light matter ( @xmath9 ) is efficient for @xmath10 . for a more concrete argument , we consider the hybrid - type potential with the superpotential @xmath11 where @xmath12 is the inflaton , and @xmath13 are a pair of charged fields ( trigger fields ) that are responsible for reheating after inflation . for the region @xmath14
, we consider the typical flat potential @xmath15 where @xmath16 and @xmath17 are dimensionless parameters and @xmath18 is an integer greater than 4 . the first term @xmath19 denotes the inflaton potential caused by the trigger field , which is displaced from the true minimum to the origin .
the terms proportional to @xmath17 originate from the typical supergravity action . for simplicity
, we consider the trigger field @xmath1 , the typical interaction given by @xmath20 and the decay rate @xmath21 .
since the decay into light matter ( @xmath9 ) is effective during @xmath10 , the dissipative coefficient of the inflaton , which is mediated by the heavy trigger field @xmath1 , is given by@xcite @xmath22 where @xmath23 is assumed .
, the dissipative coefficient is given by @xmath24@xcite . ] the strength of the damping is measured by @xmath25 , given by the dissipative coefficient and the hubble parameter : @xmath26 which can be used to rewrite the field equation of the dissipating field as @xmath27 for the dissipation caused by gravitational decay , the ratio is given by @xmath28 where @xmath29 is a dimensionless constant .. namely , we do not consider a situation with many light matter fields , which may enhance the decay rate by a factor of @xmath30 , where @xmath18 is the number of light fields
. the planck - scale suppression can be relaxed if the cut - off scale appears below @xmath31 .
these effects may lead to a very large @xmath32 . ]
note that for the modest values of @xmath33 and @xmath34 , strong dissipation is not possible if the hubble parameter is large ( @xmath35 ) during inflation .
we thus implicitly assume inflation of the scale @xmath36 .
from the above equations , we find that strong dissipation defined by @xmath37 is possible due to gravitational decay @xmath38 . for @xmath34
, the condition @xmath37 leads to a strong dissipative regime given by @xmath39 thermalization of the decay product , which is very important in warm inflation , is not always important in solving the @xmath0 problem .
namely , if the dissipative coefficient is calculated in the zero - temperature limit , the dissipation may occur without thermalization , although inflation is not warm in that case .
if the dissipation is strong for the gravitational decay , we do not need to introduce either additional symmetry that forbids a @xmath40 mass or an artificial interaction that leads to very efficient decay of the trigger field . to examine more explicitly the slow - roll conditions of dissipative inflaton
, we introduce new slow - roll parameters for the motion , different from the conventional slow - roll parameters .
they are given by @xmath41 where the usual slow - roll parameters ( @xmath42 and @xmath0 ) are defined by @xmath43 here the subscript denotes the derivative with respect to the field . for warm evolution
the third slow - roll parameter can be defined as @xmath44 which sometimes raises a trivial condition , depending on the model .
calculating the slow - roll parameter is straightforward . if the inflaton is far from the origin , terms proportional to @xmath45 dominate the potential . then , using @xmath46 during inflation , the slow - roll parameters are given by @xmath47 the condition @xmath48 leads to @xmath49 and the second condition @xmath50 leads to @xmath51 the simplest example is @xmath52 and @xmath53 , which leads to dissipative slow - roll during @xmath54 .
we show a schematic picture in figure 1 . introducing @xmath55 defined by @xmath56
we find that the above conditions are not applicable for @xmath57 , since in this region the quadratic term is more important than the higher dimensional terms . for @xmath58 ,
the slow - roll conditions are given by @xmath59 which lead to @xmath60 we thus find that strong dissipation is indeed possible for @xmath57 .
however , in contrast to standard hybrid inflation , the slow - roll does not end with the waterfall at @xmath61 .
we thus find that slow - roll inflation is possible for gravitational dissipation , but the parameter space for the slow - rolling is not trivial , depending on the model .
we show a schematic picture for a successful scenario in figure 1 .
the distinguishing property of this scenario is the appearance of @xmath62 that defines the end of dissipative slow - roll inflation .
note that in the usual hybrid - type inflation model , the end of inflation is defined by @xmath63 , where @xmath64 becomes unstable . in the present model , in contrast to non - dissipative inflation ,
the dissipative slow - roll ends at @xmath65 , where the dissipative slow - roll parameter reaches unity .
the inflaton motion is fast during @xmath66 , which occurs just before reheating at @xmath61 .
( 200,210)(110,330 ) [ fig : perspective ]
in the above argument , we considered the @xmath0-problem in supersymmetric inflation with the typical hybrid - type potential , and showed that dissipative slow - roll is possible .
this solves the traditional @xmath0 problem , though it is still not certain if the usual curvature perturbations created at the horizon exit meet the requirements of the primordial cosmological perturbations .
in fact , the cobe normalization leads to@xcite @xmath67 where @xmath68 denotes the possible correction from the background radiation .
of course , @xmath69 if there is no significant radiation background coupled to the inflaton . here
we have introduced a new parameter defined by @xmath70 .
the source of the radiation is not specified here . from the langevin equation ,
the root - mean square fluctuation amplitude of the inflaton field @xmath71 after the freeze out is obtained to be @xmath72 in the above argument , we assumed @xmath73 .
for @xmath74 , if the slow - roll conditions are satisfied , the cobe normalization leads to @xmath75 these results depend crucially on the model .
notably , we can not specify the temperature of the background radiation for the supersymmetric inflation model , since there are many directions in supersymmetric theory , which may dissipate and source the background@xcite , even though they are not the inflaton .
the above curvature perturbations , which are calculated using the standard perturbations created at the horizon crossing , may not meet the cosmological requirements .
however , the mismatch does not mean that there is a crucial difficulty in the scenario .
namely , alternative scenarios for the curvature perturbations may be important for the model , as there are typically many flat directions in the supersymmetric inflation model .
we thus consider in the next subsection an alternative for the generation of the curvature perturbations , which works specifically in the dissipative scenario .
other possibilities , such as inhomogeneous scenarios of phase transition@xcite , ( p)reheating@xcite or curvatons@xcite are not discussed in this paper , because they are not characteristic of dissipative inflation .
see ref.@xcite for more applications of the warm - flat directions .
the inflaton potential and the interactions during inflation may be inhomogeneous due to entropy perturbations related to light fields such as moduli or flat directions in the supersymmetric model@xcite .
the inhomogeneities in the potential or in the interactions may lead to inhomogeneous couplings or masses that may cause fluctuations of the inflaton velocity .
this is the basic idea of modulated inflation@xcite .
a similar situation may lead to the inhomogeneous end of hybrid inflation , as has been discussed in ref.@xcite .
however , in the present scenario the inflaton velocity at the end of inflation is rapid ( i.e. , @xmath76 or @xmath77 at the end of inflation ) and the perturbations caused by the end - boundary are less significant than the typical cosmological perturbations . on the other hand and contrary to the conventional ( cold ) hybrid model
, the perturbations caused by the modulated inflaton velocity ( @xmath78 ) can be significant in the dissipative inflation model . in this section
, we examine the scenario of modulated inflation in the dissipative inflation model . for a typical example
, we consider the @xmath79-dependence of the dissipative coefficient @xmath80 ^ 3 $ ] and the corrections suppressed by the cut - off scale : @xmath81 where @xmath79 is a light field that may be warm during inflation@xcite .
it is possible to assume a hierarchy between the cut - off scale @xmath82 and the planck scale ( @xmath83 ) .
note that in the usual cold model , in which the dissipation is negligible and @xmath64 during inflation , @xmath84 can not cause fluctuations of the inflaton velocity .
in contrast to the cold model , in the dissipative inflation model @xmath85 caused by @xmath84 can lead to significant inhomogeneities in the inflaton velocity .
inhomogeneous dissipation ( @xmath85 ) creates curvature perturbations given by@xcite @xmath86 here the flat direction @xmath79 may obtain @xmath40 mass from the supergravity , but it still can move slowly if the dissipation is significant for the flat direction@xcite .
also , for the warm - flat direction @xmath79 , the amplitude of the perturbation is given by @xmath87 which shows that the amplitude is characterized by its interaction .
here we considered strong dissipation @xmath88 for the light field @xmath79 .
the amplitude of the warm direction is always larger than the amplitude of the cold - flat direction .
the source of the background radiation is not specified , since many directions in the supersymmetric theory may dissipate during inflation and contribute to the background radiation .
the spectral index of the curvature perturbation is usually determined by the potential and the interaction of the light field @xmath79 , but for dissipative inflation it may also depend on the background radiation which may be sourced from the other flat directions .
in this paper we considered a dissipative inflation model for conventional supergravity using the typical hybrid - type potential .
if there is no symmetry that forbids hubble - induced mass , the inflaton mass will be as large as the hubble scale during inflation .
this is the famous @xmath0 problem of the supergravity inflation model .
the purpose of this paper is to find a natural solution to the problem without introducing symmetries or artificial interactions that may ruin the simplicity of the original hybrid inflation . in order to solve the problem within the naive setup
, we considered only the gravitational decay mode ( the gravitational decay of the trigger field into light matter ) and find that the dissipation of the inflaton may be large and can dominate the friction of the inflaton motion .
then we showed that dissipative damping can provide a solution to the traditional @xmath0 problem .
the dissipative inflation scenario is of course not the same as the cold scenario .
a significant discrepancy that characterizes the dissipative scenario is the way the inflation ends . in the usual cold model , inflation ends with the so - called `` waterfall '' of the trigger field at @xmath61 , while in the dissipative model slow - roll ends before the waterfall .
the differences in the dynamics of the inflaton motion lead to differences in the inflationary scenario , especially for the generation of curvature perturbations .
considering the perturbations of the dissipative coefficient , we found that modulated inflation can create significant curvature perturbations .
we wish to thank k.shima for encouragement , and our colleagues at tokyo university for their kind hospitality .
1 a. berera , `` warm inflation , '' phys .
lett . * 75 * , 3218 ( 1995 ) [ arxiv : astro - ph/9509049 ] .
a. berera , `` the warm inflationary universe , '' contemp .
phys . * 47 * , 33 ( 2006 ) [ arxiv:0809.4198 [ hep - ph ] ] .
m. bastero - gil and a. berera , `` determining the regimes of cold and warm inflation in the susy hybrid model , '' phys .
d * 71 * , 063515 ( 2005 ) [ arxiv : hep - ph/0411144 ] .
a. r. liddle and d. h. lyth , `` cosmological inflation and large - scale structure , '' _ cambridge , uk : univ . pr .
( 2000 ) 400 p_. t. matsuda , `` aspects of warm - flat directions , '' arxiv:0908.3059 [ hep - ph ] ; t. matsuda , `` remote inflation as hybrid - like sneutrino / mssm inflation , '' arxiv:0905.4328 [ hep - ph ] ; t. matsuda , `` remote inflation : hybrid - like inflation without hybrid - type potential , '' jcap * 0907 * , 003 ( 2009 ) [ arxiv:0904.2821 [ astro-ph.co ] ] ; t. matsuda , `` evolution of the curvature perturbations during warm inflation , '' jcap * 0906 * , 002 ( 2009 ) [ arxiv:0905.0308 [ astro-ph.co ] ] .
t. matsuda , `` cosmological perturbations from an inhomogeneous phase transition , '' class .
* 26 * , 145011 ( 2009 ) [ arxiv:0902.4283 [ hep - ph ] ] ; t. matsuda , `` cosmological perturbations from inhomogeneous preheating and multi - field trapping , '' jhep * 0707 * , 035 ( 2007 ) [ arxiv:0707.0543 [ hep - th ] ] .
e. w. kolb , a. riotto and a. vallinotto , `` curvature perturbations from broken symmetries , '' phys .
d * 71 * , 043513 ( 2005 ) [ arxiv : astro - ph/0410546 ] ; t. matsuda , `` generating the curvature perturbation with instant preheating , '' jcap * 0703 * , 003 ( 2007 ) [ arxiv : hep - th/0610232 ] ; t. matsuda , `` generating curvature perturbations with mssm flat directions , '' jcap * 0706 * , 029 ( 2007 ) [ arxiv : hep - ph/0701024 ] . t. matsuda , `` hybrid curvatons from broken symmetry , '' jhep * 0709 * , 027 ( 2007 ) [ arxiv:0708.4098 [ hep - ph ] ] ; t. matsuda , `` no curvatons or hybrid quintessential inflation , '' jcap * 0708 * , 003 ( 2007 ) [ arxiv:0707.1948 [ hep - ph ] ] .
a. d. linde and v. f. mukhanov , `` nongaussian isocurvature perturbations from inflation , '' phys .
d * 56 * , 535 ( 1997 ) [ arxiv : astro - ph/9610219 ] ; d. h. lyth and d. wands , `` generating the curvature perturbation without an inflaton , '' phys .
b * 524 * , 5 ( 2002 ) [ arxiv : hep - ph/0110002 ] ; t. moroi and t. takahashi , `` effects of cosmological moduli fields on cosmic microwave background , '' phys .
b * 522 * , 215 ( 2001 ) [ erratum - ibid .
b * 539 * , 303 ( 2002 ) ] [ arxiv : hep - ph/0110096 ] .
t. matsuda , `` curvaton paradigm can accommodate multiple low inflation scales , '' class .
. grav .
* 21 * , l11 ( 2004 ) [ arxiv : hep - ph/0312058 ] ; t. matsuda , `` topological curvatons , '' phys .
d * 72 * , 123508 ( 2005 ) [ arxiv : hep - ph/0509063 ] ; t. matsuda , `` hilltop curvatons , '' phys . lett .
b * 659 * , 783 ( 2008 ) [ arxiv:0712.2103 [ hep - ph ] ] .
l. kofman , `` probing string theory with modulated cosmological fluctuations , '' arxiv : astro - ph/0303614 .
t. matsuda , `` evolution of curvature perturbation in generalized gravity theories , '' class .
* 26 * , 145016 ( 2009 ) [ arxiv:0906.0643 [ hep - th ] ] ; t. matsuda , `` delta - n formalism for the evolution of the curvature perturbations in generalized multi - field inflation , '' arxiv:0906.2525 [ hep - th ] ; t. matsuda , `` modulated inflation , '' phys .
b * 665 * , 338 ( 2008 ) [ arxiv:0801.2648 [ hep - ph ] ] ; t. matsuda , `` successful d - term inflation with moduli , '' phys .
b * 423 * , 35 ( 1998 ) [ arxiv : hep - ph/9705448 ] .
f. bernardeau , l. kofman and j. p. uzan , `` modulated fluctuations from hybrid inflation , '' phys .
d * 70 * , 083004 ( 2004 ) [ arxiv : astro - ph/0403315 ] ; d. h. lyth , `` generating the curvature perturbation at the end of inflation , '' jcap * 0511 * , 006 ( 2005 ) [ arxiv : astro - ph/0510443 ] ; t. matsuda , `` elliptic inflation : generating the curvature perturbation without slow - roll , '' jcap * 0609 * , 003 ( 2006 ) [ arxiv : hep - ph/0606137 ] . | we consider supersymmetric inflation with the hybrid - type potential . in the absence of the symmetry that forbids hubble - induced mass terms
, the inflaton mass will be as large as the hubble scale during inflation .
we consider gravitational decay of the trigger field as the least decay mode and find that the damping caused by the dissipation can dominate the friction of the inflaton when the heavy trigger field is coupled to the inflaton .
the dissipative damping provides a solution to the traditional @xmath0 problem without introducing additional symmetry and interactions .
considering the spatial inhomogeneities of the dissipative coefficient , we find that modulated inflation ( modulation of the inflaton velocity ) can create significant curvature perturbations . sit - hep / tm-58 1.5 truecm .75 truecm * tomohiro matsuda * .4 truecm _ laboratory of physics , saitama institute of technology , _
_ fusaiji , okabe - machi , saitama 369 - 0293 , japan _
\1 .
truecm addtoresetequationsection 1 . |
the small handful of unusual objects known as anomalous x - ray pulsars ( axps ) are characterized by their slow spin periods ( 512 s ) , large inferred magnetic fields ( @xmath110@xmath11 g ) , and persistent x - ray luminosities in great excess of available spin - down power . these young , isolated pulsars have properties common to another class of objects , the soft gamma repeaters ( sgrs ) . for recent reviews , see @xcite and @xcite .
potentially the most intriguing and most revealing property shared by axps and sgrs is their highly volatile nature : both are known to exhibit occasional and sudden dramatic flux and spin variability in the form of x - ray bursts , flares , and glitches .
the magnetar model @xcite was instrumental in explaining both the `` anomalous '' x - ray emission , as well the episodes of burst activity .
it identified the sources as highly magnetized isolated neutron stars , with field strengths on the order of @xmath12 g. the magnetar model uniquely , and correctly , predicted that sgr - like bursts would be observed from axps .
a major proposed effect of this enormous magnetization is the `` twisting '' of the magnetosphere , with associated magnetospheric currents and heating of the crust at the twisted field - line footpoints @xcite .
this gives rise to the unusual x - ray spectrum , which has generally been empirically well characterized by a two - component model consisting of a power law plus blackbody .
bursts of high - energy emission occur presumably when the crust succumbs to magnetic stresses and deforms , leading to a rapid rearrangement of the external magnetic field .
recently , @xcite proposed the existence of a plasma corona contained within the closed magnetosphere to explain the broad band spectrum of magnetars that extends from the infrared ( ir ) to hard x - rays beyond 100 kev @xcite .
1e 1048.1@xmath05937 has had an unusual history , even by axp standards .
prior to 2002 , monitoring of the pulsed flux with the _ rossi x - ray timing explorer _ ( _ rxte _ ) showed that its spin down was so unstable that phase coherence could be maintained for periods of only months at a time @xcite , behaviour which echoed earlier reports of @xmath13 , flux and spectral variability @xcite .
two small bursts , the first ever found in an axp , were discovered in 2001 from 1e 1048.1@xmath05937 @xcite , clinching the suspected association between axps and sgrs . since 2002 , tri - weekly _ rxte _ observations of 1e 1048.1@xmath05937 allowed the changing pulsed flux and spin - down rate to be monitored on relatively short timescales .
* hereafter gk04 ) reported order - of - magnitude torque variability on timescales of weeks to months that only marginally correlated with large luminosity variations . the unusually slow and long - lived pulsed flux `` flares '' ( not to be confused with the sgr giant flares ; see @xcite ) that began in late 2001 and lasted into 2004 are in some ways unlike activity seen in any other axp or sgr thus far .
the 2@xmath010 kev spectrum and pulsed fraction are also variable on long timescales , the latter being anti - correlated with luminosity @xcite . at near - infrared ( near - ir ) wavelengths , the flux varied dramatically during 2002@xmath02003 @xcite ; unfortunately , the sparsity of near - ir observations prohibits measurements of the variability timescale , although a near - ir / x - ray anti - correlation has been suggested @xcite .
further activity in the form of another burst and small pulsed - flux increase was seen in 2004 @xcite . here , we report on a series of simultaneous _ cxo _ and _ hst_observations obtained during the course of 2006 , as well as on archival _ cxo _ , _ xmm _ and _ vlt _ observations , in [ sec : observations ] .
we also present _ rxte _ timing results from the past three years . in 2007 march ,
1e 1048.1@xmath05937 unexpectedly entered yet another new phase of activity , and we report on target - of - opportunity observations with _ cxo _ and _ swift _ taken after the 2007 march event . in [ sec : quiescent discussion ] we discuss the 2004 - 2006 period of radiative and spin quiescence in the context of both magentar and accretion models , and in [ sec : flare discussion ] , we describe and interpret the events following the 2007 march flare .
as part of a long - term project , we regularly monitor axps using the _ rossi x - ray timing explorer _ ( _ rxte _ ) . 1e 1048.1@xmath05937 is by far the most frequently observed axp due to its relatively poor rotational stability .
the _ rxte _
observations ( see [ sec : rxte ] ) are crucial in measuring the spin evolution of the source and its pulsed flux . however , the high background and large unimaged field - of - view of _ rxte _ make spectral measurements of 1e 1048.1@xmath05937 difficult and total flux ( and consequentially pulsed fraction ) measurements impossible . to make these measurements
, we obtained five _ chandra x - ray observatory _
observations ( _ cxo _ ; see [ sec : chandra ] ) approximately equispaced throughout 2006 . in order to probe the origin of the near - ir emission of this source , we observed with the _ hubble space telescope _ ( _ hst _ ; see
[ sec : hst ] ) simultaneously with _
cxo_. the motivation for simultaneous _
cxo _ and _ hst _
observations was to search for correlated variability in the different bands , as that provides insight into the physical mechanisms generating the radiation .
we also examined archival @xmath14-band observations with the _ very large telescope _
( _ vlt _ ; see [ sec : vlt ] ) , which contribute to our monitoring of its near - ir brightness .
following the 2007 march glitch detected in our _ rxte _ monitoring observations of 1e 1048.1@xmath05937 @xcite , we initiated three _
cxo _ and two _ swift gamma - ray burst explorer _ ( _ swift _ ;
[ sec : swift ] ) target - of - opportunity ( too ) observations to follow the x - ray flux and spectrum of the source . in order to characterize the long - term evolution of 1e 1048.1@xmath05937 we also incorporate and reanalyze archival _ xmm - newton _ and _ cxo_observations ( see * ? ? ?
* for details on how these data were processed ) .
separate post - outburst optical and near - ir observations of 1e 1048.1@xmath05937 are reported in detail elsewhere @xcite .
we have observed 1e 1048.1@xmath05937 regularly since 1997 with _ rxte_(@xcite ; ) .
our data were obtained using the proportional counter array ( pca ) on board _ rxte _ which consists of five identical and independent xenon / methane proportional counter units ( pcus ) .
we use our _ rxte _ observations of 1e 1048.1@xmath05937 to look for the presence of bursts ( see @xcite for details ) , to look for pulse profile changes , to monitor its pulsed flux , and to monitor its frequency evolution using phase - coherent timing when possible . for the timing analysis , we created barycentered lightcurves in the [email protected] kev band with 31.25 ms time resolution .
as past monitoring has shown that it can be difficult to maintain pulse phase coherence over timescales longer than a few weeks , observations of 1e 1048.1@xmath05937 since 2002 are done three times per week , with the three observations carefully spaced so as to allow a phase - coherent analysis and a precise frequency measurement .
thus , for each observation , we fold at the pulse period determined via periodogram , cross - correlate the folded profiles with a high signal - to - noise template , and fit the resulting phases with a linear function whose slope provides the average frequency . frequencies determined in this way are shown in figure [ fig : timing ] . as can be seen in this figure , the 2004 - 2006 frequencies were actually much more stable than in the past ; we therefore attempted a fully phase - coherent analysis as well
. this will be described elsewhere ( dib et al .
, in preparation ) .
frequency derivatives , which are also displayed in figure [ fig : timing ] and are from and dib et al . , ( in preparation ) , are clearly much more stable in the 2004 - 2006 interval than previously .
this period of rotational stability was accompanied by x - ray pulsed flux stability and relative quiescence : the pulsed flux time series of 1e 1048.1@xmath05937 in the 2@xmath010 kev band is presented in figure [ fig : timing ] .
the pulsed flux was calculated using a method similar to that described in @xcite and is based on the rms of the folded profile , but without variance subtraction .
the long period of rotational stability ended in 2007 march , when a large glitch was observed @xcite .
the glitch also signaled the end of the period of pulsed flux quiescence .
the pulsed flux suddenly increased by a factor of @xmath13 in the energy range 2@xmath010 kev .
the upper limit on the rise time of the pulsed flux for this event is approximately one week .
the peak pulsed flux reached by the source was @xmath110% larger than the peak flux reached during the largest of the two previously observed flares , and the rise time was at least 4 times smaller . on 2007 may 17 , the date of the last x - ray imaging observation included in this paper , the pulsed flux had decreased by @xmath110% .
it is as yet difficult to compare the decay timescale of the pulsed flux of this new event to that of the previous flares however it will be possible in the near future . [
sec : swift ] x - ray imaging observations were carried out with the _ cxo _ and _ swift_telescopes .
the date , total exposure time and resulting count rate for each region - filtered background - subtracted observation are listed in table [ tab : obs ] and a detailed description of our spectral fitting is given in
[ sec : spectral fit ] .
1e 1048.1@xmath05937 was observed five times in 2006 with _
cxo _ in quasi - equispaced intervals , and three times in 2007 april .
the observations were made with the advanced ccd imaging spectrometer ( acis ) camera using the s3 chip in continuous clocking ( cc ) mode .
this mode generates a 1@xmath31024 pixel image , which is read out every 2.8 ms ; pile - up in this mode is not an issue .
the subsequent analysis was based on the `` level 2 '' events files for which the event times are photon arrival times ( as opposed to readout times for the `` level 1 '' events ) . following the standard threads we then corrected the `` level 2 '' events for certain caveats related to acis cc - mode data .
we extracted a rectangular region centered on the source with a width typically of @xmath130 pixels .
we estimated the background using rectangular regions on either side of the source which extended from @xmath110 pixels beyond the edge of the source region to @xmath110 pixels from the edge of the image . following the standard threads , the source and background spectra
were then extracted using ` ciao ` v3.4 .
we grouped the source spectrum such that there were no fewer than 20 counts per bin after background subtraction .
the response matrix function ( rmf ) and area response function ( arf ) for each observation were also generated using ` ciao ` and ` caldb ` .
we used timed exposure ( te ) mode response matrices because of the absence of available spectral calibration for cc mode .
1e 1048.1@xmath05937 was observed three times in 2007 with _ swift _ using the x - ray telescope ( xrt ) .
xrt has two observing modes : photon counting ( pc ) and windowed - timing ( wt ) .
pc mode provides a @xmath1600@xmath3600 pixel@xmath15 image with low ( @xmath12.5 s ) time resolution , whereas wt mode provides a one - dimensional @xmath1200-pixel wide image with high ( @xmath10.74 ms ) time resolution .
xrt automatically switches from wt to pc - mode when the count rate exceeds @xmath12 counts s@xmath8 . in our analysis
we ignored the wt mode data as they were too short to be of use .
we used the pc - mode event files from the standard pipeline ; these were cleaned and barycentered . using ` ftool xrtcentroid ` , we determined the centroid of the image , and input the cleaned , barycentered events into the command line interface ` xselect ` .
we selected a circular source region with a 25 pixel radius centered on the centroid of the image . since the source was slightly piled up , we excluded a 4 pixel radius circular region centered on the centroid of the image .
we used 4 pixels because that is the recommended exclusion radius for a source with 1e 1048.1@xmath05937 s spectrum ( l. angelini , private communication ) .
for the background , we selected events from an annulus with an inner radius of 50 pixels and outer radius of 130 pixels .
we further filtered our events by selecting only those that had grades 0 to 12 . with ` xselect `
we then created source and background pha files .
the ` ftool xrtmkarf ` was used to create the arfs , and we input the source pha file with the excluded center so that it would be corrected for pile - up .
we used the file provided by ` caldb ` appropriate for pc mode data events of grades 0 to 12 to make the rmf .
lccccccccc + 2000 dec 28 & 51906 & 0112780401 & 4447 & 1.461(18 ) & 3.31(22 ) & 0.656(17 ) & 4.8(4 ) & 0.6(2 ) & 0.76(2 ) + 2003 jul 16 & 52836 & 0147860101 & 41145 & 3.429(9 ) & 2.93(4 ) & 0.599(6 ) & 11.5(1 ) & 1.21(5 ) & 0.44(4 ) + 2004 jul 9 & 53195 & 0164570301 & 32679 & 0.985(6 ) & 2.75(5 ) & 0.542(15 ) & 9.6(1 ) & 2.1(2 ) & 0.561(7 ) + + 2004 jul 10 & 53196 & 4653 & 28860 & 1.363(7 ) & 2.87(7 ) & 0.56(1 ) & 9.3(1 ) & 1.2(1 ) & 0.56(1 ) + 2004 jul 15 & 53201 & 4654 & 28085 & 1.317(7 ) & 2.87(7 ) & 0.55(1 ) & 9.4(2 ) & 1.2(1 ) & 0.55(1 ) + + 2006 feb 26 & 53792 & 6733 & 22085 & 1.072(8 ) & 2.79(6 ) & 0.51(1 ) & 7.1(1 ) & 2.0(2 ) & 0.63(1 ) + 2006 apr 20 & 53845 & 6734 & 20553 & 1.157(9 ) & 2.75(5 ) & 0.50(1 ) & 7.1(2 ) & 2.0(2 ) & 0.61(1 ) + 2006 jun 14 & 53900 & 6735 & 22085 & 0.985(9 ) & 2.95(9 ) & 0.55(1 ) & 6.3(1 ) & 1.4(2 ) & 0.64(1 ) + 2006 jul 30 & 53946 & 7347 & 22176 & 0.821(8 ) & 2.72(8 ) & 0.51(1 ) & 6.8(2 ) & 1.2(2 ) & 0.70(1 ) + 2006 sep 23 & 54001 & 6736 & 22022 & 1.006(7 ) & 3.14(11 ) & 0.56(1 ) & 5.1(1 ) & 1.0(2 ) & 0.64(1 ) + + 2007 apr 6 & 54196 & 7647 & 20079 & 4.241(15 ) & 2.30(4 ) & 0.56(1 ) & 36.9(5 ) & 1.6(1 ) & 0.299(5 ) + 2007 apr 16 & 54206 & 7648 & 18387 & 3.959(15 ) & 2.27(4 ) & 0.55(1 ) & 36.0(5 ) & 1.7(1 ) & 0.286(6 ) + 2007 apr 28 & 54218 & 7649 & 19082 & 4.332(15 ) & 2.12(4 ) & 0.55(1 ) & 36.0(5 ) & 1.8(1 ) & 0.287(5 ) + + 2007 apr 3 & 54193 & 30912001 & 4905 & 0.58(1 ) & 2.7(2 ) & 0.58(6 ) & 35(2 ) & 2.0(6 ) & 0.22(3 ) + 2007 apr 13 & 54203 & 30912003 & 4829 & 0.55(1 ) & 2.8(4 ) & 0.62(4 ) & 32(2 ) & 0.7(3 ) & 0.17(3 ) + 2007 may 17 & 54237 & 30912009 & 3626 & 0.51(1 ) & 2.3(4 ) & 0.7(1 ) & 35(3 ) & 2(1 ) & 0.22(3 ) we extracted spectra of 1e 1048.1@xmath05937 from all archival , monitoring , and too observations collected from _
cxo _ , _ xmm _ and _ swift_. a description of the recent data analysis is in
[ sec : chandra ] , and the archival analysis was described in @xcite . using the fitting package ` xspec ` v12.3.1 we modeled the x - ray spectra with a photoelectrically absorbed blackbody plus power law .
we used the ` phabs ` photoelectric absorption ` xspec ` model , which assumes the solar abundance table of @xcite and uses the ` bcmc ` photoionization cross - section table from @xcite with the new he cross - section from @xcite .
we fit this model to all the observations simultaneously , allowing the column density @xmath16 to vary but with the sole constraint that it be the same for all observations .
we restricted our fit to the 0.75 kev band .
the total global fit had @xmath17 for @xmath18 degrees of freedom ( dof ) .
individually , the observations were equally well modeled ( except for a relatively high @xmath19 value in one observation , see [ sec : spec feature ] ) .
our best - fit column density obtained from the global fit is @xmath20 @xmath7 , in agreement with @xcite .
the results of our spectral fitting are listed in table [ tab : obs ] and plotted in figure [ fig : xrayspec ] .
all observations had spectra that were well modeled by a photoelectrically absorbed blackbody plus power law .
however , the _ cxo _ observation on 2007 april 6 ( observation i d 7647 ) , the first after the 2007 march event , had a relatively high @xmath21 ( @xmath22 dof ) .
there is possible evidence for an absorption line at @xmath12.7 kev ( see fig .
[ fig : spectrum ] ) . adding a gaussian line
improved the fit ( @xmath23 for @xmath24 dof ) with @xmath25 .
with the line width fixed at 0.1 kev , we measure a line energy of @xmath26 kev . allowing the line width @xmath27 to be a free parameter
, we obtain @xmath28 kev , a line energy of @xmath29 kev and @xmath30 for 351 dof .
we find no evidence of a phase dependence for this possible line .
a detailed phase - resolved spectroscopic analysis will be presented in a forthcoming paper . to test the significance of adding such a line
, we performed the following simulation .
we generated 10000 simulated spectra by adding poisson noise to a blackbody plus power - law model spectrum having the same parameters as our best - fit model .
we then determined the maximum change in @xmath19 after adding a gaussian line . to avoid local minima and ensure that we found the true minimum @xmath19 for each simulation iteration , instead of fitting for the peak energy
, we stepped through different line energies between 0.6 and 7.0 kev with a step size of 0.025 kev .
the width of each line was held fixed at 0.1 kev and its normalization was allowed to vary . in all the simulation iterations ,
none had a change in @xmath19 greater than 48.3 .
the probability that this feature is due to random chance is @xmath310.13% , accounting for the number of trials .
we note however that with the addition of the line , the overall fit is still unacceptable ; we speculate that other features may be present as well , though at marginal significance .
for example , there is a small but intriguing feature at @xmath11.3 kev , which is approximately half the energy of the line discussed above .
the evidence for a line must be considered tentative because cc mode , which uses the te mode response matrices , is not spectrally calibrated , and there exist calibration lines between 1.7 and 3 kev in the te response . however
, this is true for the other _ cxo _ observations , including those in which we find no such features . using our _ cxo _ and _ swift _
observations we were also able to study the source s pulse morphology , flux and pulsed fraction . for the _ cxo _
observations , we barycentered our corrected level 2 event lists and then extracted background - subtracted light curves in the different bands using the same source and background regions as for the spectral analysis .
we folded these lightcurves at the optimal frequency as determined by a periodogram .
the frequencies we obtained all agreed with the ephemerides determined with our contemporaneous _ rxte
figure [ fig : profiles ] ( _ left _ ) displays the _ cxo _ pulse profiles in the 1@xmath03 and 3@xmath010 kev bands .
notice that after the flare ( last 3 panels in figure [ fig : profiles ] _ left _ ) , the pulse profile changed from single to at least triple peaked . to study the pulse morphology evolution quantitatively
we decomposed the pulse profiles into their fourier components . in figure
[ fig : profiles ] ( _ right _ ) we plot the fourier components in terms of their ratio to the fundamental ( the @xmath32 and fundamental are excluded from the plot ) .
notice that before the flare the profile consisted only of the fundamental and an @xmath33 fourier component .
after the flare , the profile exhibited significant fourier components of order @xmath34 .
interestingly , despite the drastic changes in the other fourier components , the @xmath35 component remains consistent with zero throughout the flare .
similarly , for the _ swift _
observations we used the barycentered , cleaned , and pointed pc - mode data events . the xrt wt data consisted of too few pulses to be useful . _
swift _ pc - mode data have only 2.5-s resolution , providing fewer than 3 pulse phase bins for this 6.45-s pulsar .
we attempted to improve on the coarse sampling by utilizing advanced `` bin splitting '' techniques ( g. israel , private communication ) , however we find only marginal evidence for a pulse profile change from these data ( but see * ? ? ?
* ) . using our folded pulse profiles we measured root - mean - square ( rms ) pulsed fractions ,
the rms pulsed fraction is given by @xmath36 , and the peak - to - peak pulsed fraction is given by @xmath37 .
we opted to use the rms with fourier filtering ( archibald et al .
, in preparation ) because it is a more robust estimator of the pulsed fraction than peak - to - peak .
] using the method described in @xcite .
our 2@xmath010 kev rms pulsed fractions are listed in table [ tab : obs ] and plotted in figures [ fig : timing ] and [ fig : xrayspec ] .
the _ xmm _
pulsed fractions , taken from @xcite , were extracted in the same fashion .
because of the low time resolution of the _ swift _ data , the rms pulsed fraction would be artificially reduced if the pulse contained as much harmonic content as the _ cxo _ data during the flare . thus , for each _ swift _ observation we simulated , given its observed pulsed fraction and mean count rate , what its true pulsed fraction would be assuming a pulse profile with as much harmonic content as the nearest _ cxo_observation .
on average we found a reduction of no more than @xmath173% .
note in figure [ fig : xrayspec ] the observed tight correlation between total flux and pulsed fraction ( see also * ? ? ?
* ) which we discuss in detail in
[ sec : pfvsf disc ] .
we observed 1e 1048.1@xmath05937 using _ hst _ simultaneously with _
cxo _ on four occasions in 2006 : february 26 , april 20 , july 30 , and september 23 ( program i d 10761 ) .
observations were made with the near infrared camera and multi - object spectrometer ( nicmos ) instrument , a @xmath38 square pixel hgcdte detector . combined with camera 3
, it provided a focus ratio of @xmath39/17 , a field of view of @xmath40 , and a plate scale of @xmath41 pixel@xmath8 .
the detector was read out in multiaccum mode , and a spiral dither pattern with 5@xmath42 spacings was applied .
we observed using the filter f110w ( similar to the ground based @xmath43-band filter ) on two occasions and f160w ( similar to @xmath44-band ) on four occasions .
the average fwhm of the point spread function ( psf ) was @xmath45 for all nicmos observations .
observing parameters are summarized in table [ tab : irobs ] .
the data underwent `` on - the - fly - reprocessing '' involving standard pipeline routines before being retrieved from the _ hst _ archive .
the calibration routine ` calnica ` performed basic data reduction steps , and ` calnicb ` produced a single final image for each observation , combining the dithered mosaic input images .
daophot @xcite for iraf v2.12.2 was used to perform psf photometry .
absolute calibration of our measurements was done using the nicmos photometric keywords for which the uncertainties are believed to be less than 5% ; we found the aperture correction to correspond with the calibration keywords by simulating psfs with the tiny tim software package . in order to determine limiting instrumental magnitudes in each of the observations ,
we performed psf photometry on a set of a hundred trial images to which a single artificial star was added .
this artificial star was placed in a @xmath46 box in a blank region near the nominal position of the ir counterpart . by varying the brightness of the star
, we define the @xmath47 detection limit at the instrumental magnitude for which the psf photometry recovers the artificial star with an uncertainy of 0.3mag ( see * ? ? ?
* ) . at the position of 1e 1048.1@xmath05937 ,
as measured by @xcite in @xmath14-band with the _ magellan _ telescope , we find one point source in the f160w image from 2006 february 26 only . to confirm the positional coincidence of this source with the _ magellan _ object , we astrometrically tied @xmath48 nearby field stars in our _ hst _ image to those from the _ magellan _ image .
a @xmath49 offset was found for the 1e 1048.1@xmath05937 candidate .
source s relative positional uncertainty , given its @xmath50 fwhm radius , is roughly fwhm/@xmath51 .
therefore , we consider this _
hst _ object the counterpart of 1e 1048.1@xmath05937 .
thus , we measure our lone detection of 1e 1048.1@xmath05937 to have @xmath52 @xmath53 mag ; uncertainties indicate the nominal errors determined by daophot and incorporate psf fitting uncertainties . limiting magnitudes for the three latter _
f160w observations in which no counterpart was seen are presented in table [ tab : irobs ] . to deepen the sensitivity ,
geometric transforms were computed to register and combine the three non - detection datasets . in the combined data , we tentatively detect an extremely faint counterpart of brightness @xmath52 @xmath54 mag , where the 3@xmath27 upper limit is @xmath55 mag .
we find no f110w counterpart down to a 3@xmath27 upper limit of @xmath56 @xmath57 mag and @xmath58 mag , on february 26 and april 20 , respectively .
results are plotted in figure [ fig : nearir ] . to examine how this compares with past results
, we also calibrated our measurements using the @xmath44 magnitudes from @xcite .
a relative offset is found between _ hst _
s f160w filter and ground based @xmath44-band , as a function of @xmath59 color .
we know the color of 1e 1048.1@xmath05937only during a brightened state ( @xmath60 in april 2002 ; * ? ? ?
* ) , and suspect that it is not constant ; therefore , this is not a perfect comparison , but provides a rough estimate . based on the magnitudes of seven nearby field objects , we measure @xmath61 mag on february 26 , which is very near what we find for @xmath52 .
lcccccc + 2005 apr 11 & 52324 & 900 & @xmath14 & @xmath220.3 & & @xmath317.0 + 2005 apr 29 & 53471 & 900 & @xmath14 & @xmath220.6 & & @xmath315.3 + 2005 may 27 & 53489 & 900 & @xmath14 & @xmath221.2 & 21.0(3 ) & @xmath313.1 ; [email protected] + 2005 jun 6 & 53527 & 900 & @xmath14 & @xmath221.4 & & @xmath312.6 + + 2006 feb 26 & 53792 & 2637 & f110w & @xmath225.0 & & @xmath310.37 + 2006 feb 26 & 53792 & 1196 & f160w & @xmath223.7 & 22.70(14 ) & @xmath310.49 ; [email protected] + 2006 apr 20 & 53845 & 2637 & f110w & @xmath224.8 & & @xmath310.45 + 2006 apr 20 & 53845 & 1196 & f160w & @xmath223.1 & & @xmath310.89 + 2006 jul 30 & 53946 & 1037 & f160w & @xmath223.3 & & @xmath310.74 + 2006 sep 23 & 54001 & 1037 & f160w & @xmath222.9 & & @xmath311.0
we have analyzed archival observations of 1e 1048.1@xmath05937 which were obtained with naos - conica ( naco ) , the near - ir adaptive optics ( ao ) instrument at _ vlt _ of the european southern observatory .
the source was observed for 900s in @xmath63-band using the s27 camera ( @xmath64 field - of - view and 27 maspix@xmath8 pixel scale ) on four occasions ( april 11 , april 29 , may 27 and june 6 ) in 2005 .
star a @xcite , located @xmath65 from the axp , was used as the wavefront - sensing ( wfs ) star . due to the relative faintness of the wfs star ( @xmath66 mag ) ,
the ao correction was only partial .
the observations , either dithered sets of 5 @xmath67s integrations ( on the first night ) or of 10 @xmath68s integrations , were corrected for dark current , flat - fielded ( using averages derived from the science frames ) and corrected for variations in the sky ( using median averages of the flat - fielded science frames ) .
finally , the reduced science images were aligned using integer pixel offsets and median - combined to create an average image for each night . using daophot , we tried psf photometry on each of the four averaged images , using different models for the psf . because the ao correction is only partial , we found that the psf is reasonably well modeled by a moffat function of exponent 2.5 in combination with a look - up table .
the width of the psf that was determined for each observation was @xmath69 ( april 11 ) , @xmath70 ( april 29 ) , @xmath71 ( may 27 ) and @xmath72 ( june 6 ) . only in the observation of may 27
did we detect the near - ir counterpart to 1e 1048.1@xmath05937 .
we calibrated the instrumental magnitudes and magnitude limits using stars x2 to x8 and a and b against the calibrated @xmath63 magnitudes of wang & chakrabarty ( 2002 ) .
the rms uncertainty in the calibration is about 0.2mag .
the near - ir counterpart to 1e1048.1@xmath05937 was detected at @xmath73 on 2005 may 27 ; the @xmath47 upper limits for april 11 , april 29 , may 27 and june 6 are @xmath74 , @xmath75 , @xmath76 and @xmath77 , respectively .
limiting magnitudes were determined with the same method described in [ sec : hst ] .
results are shown in figure [ fig : nearir ] .
note that post-2007 event optical and near - ir observations of 1e 1048.1@xmath05937were obtained with _ vlt _ and _ magellan _ @xcite .
preliminary results of the @xmath14-band observations are shown in figure [ fig : nearir ] , and details on some of the observations can be found in @xcite .
explaining the origin of the different types of axp variability across the electromagnetic spectrum is a major challenge for the magnetar model .
many fundamental questions remain open . in the case of 1e 1048.1@xmath05937 , how are timing instabilities related to radiative variability ?
axp spin - up glitches sometimes are accompanied by major radiative events , as in the 2002 outburst of axp 1e 2259@xmath78586 @xcite , and sometimes not , as in the large glitch of 1e 1841@xmath0045 @xcite .
moreover , are the _ gradual _ large - scale changes seen in the 2001 - 2002 flares of 1e 1048.1@xmath05937 of a similar origin to axp outbursts like that of 1e 2259@xmath78586 or xte j1810@xmath0197 @xcite , and how can they be accounted for by magnetar models ? also , what is the nature of the optical / ir emission observed in five out of eight confirmed axps ?
suggested origins include coherent or curvature emission in the plasma magnetosphere @xcite , but so far this has not be confirmed .
are the low- and high - energy emission mechanisms intimately connected @xcite ?
correlated post - burst flux decay in the x - ray and near - ir regimes has been seen in at least one other axp @xcite . given the many observational properties that have been characterized for 1e 1048.1@xmath05937 in this work and for other axps in similar studies , which property will emerge as being the most constraining of physics is hard to know .
the most promising behaviors are those which show clear correlations with others , or those that are common to many or all axps and sgrs . in this section , we discuss the behavior of 1e 1048.1@xmath05937 , and focus on the phenomena that are potentially the most useful for testing the magnetar or other competing models .
we first consider the quiescent phase we have observed in 2004 - 2006 , and its implications for axp models , then subsequently discuss the source s return to activity in 2007 march .
the x - ray flaring observed pre-2004 contrasts strongly with the stable pulsed fluxes we observed in 2004 - 2006 ( fig .
[ fig : timing]_c _ ) .
clearly the source s stable state is also its faintest .
the x - ray spectrum during this quiescent phase is also fairly constant , with @xmath79 evolving post - flare , on a time scale of several years .
interestingly , @xmath79 does not return to its pre - flare value as measured in 2000 december ( see fig .
[ fig : xrayspec]_a _ ) .
similarly , the photon index ( fig . [ fig : xrayspec]_b _ ) decreased in 2004 - 2006 , away from its softer pre - flare value . meanwhile the pulsed fraction ( figs .
[ fig : timing]_e _ and [ fig : xrayspec]_e _ ) was slowly rising in 2004 - 2006 , as the source grew fainter , as if , unlike @xmath79 and @xmath80 , it was slowly recovering to the pre - flare value .
thus , overall , the quiescent period is characterized by slow evolution , on timescales of years , in which the source flux and pulsed fraction slowly relaxed back to their pre - flare values , while the source spectrum varied significantly ( fig .
[ fig : xrayspec]_c _ ) and did not relax back to its pre - flare state .
this is suggestive of a fixed energy loss rate in quiescence , though perhaps with a different magnetospheric current configuration , which impacts the surface emission via return currents @xcite .
we note , on the other hand , that the 2006 and pre-2001 pulse profile were very similar if not identical @xcite , and that the @xcite simple prediction for the decay time of the sort of magnetospheric twist required for 1e 1048.1@xmath05937 is an order of magnitude too large for standard parameters . the evolution could also be purely due to the thermal component @xcite .
simultaneous with the slow evolution of the x - ray flux , pulsed fraction , and spectrum , the pulsar s rotational behavior clearly stabilized , with the source spinning down relatively smoothly at a value close to the long - term average in 2004 - 2006 ( fig .
[ fig : timing]_a _ ) . whereas previously this axp distinguished itself from others by defying attempts at phase - coherent timing and exhibiting large torque variations ( fig .
[ fig : timing]_b _ ) , in 2004 - 2006 , 1e 1048.1@xmath05937 resembled , from a timing point of view , other axps which are relatively stable rotators , at least when not glitching ( e.g. * ? ? ?
* ; * ? ? ?
this behavior demonstrates a clear relationship between timing and radiative properties , and that the `` noise '' seen in the frequency derivative during 2001 - 2004 is likely physically different from the `` timing noise '' observed ubiquitously in radio pulsars and in otherwise radiatively stable axps . in the context of the magnetar model ,
magnetospheric activity can account for both torque and x - ray flux variability , although the former is most sensitive to currents anchored closest to the magnetic poles , so that only a broad correlation between flux and torque is expected @xcite and indeed is seen when considering many axps and sgrs @xcite . that the torque and luminosity in 1e 1048.1@xmath05937 do not vary simultaneously or in a clearly correlated way
is thus not necessarily problematic if the magnetospheric current configuration is changing , although it does demonstrate that the magnetar model in this particular regard is not strongly predictive .
although the magnetar model is favored , it has also been suggested that accretion from a fossil debris disk could explain axp spin characteristics @xcite as well as all aspects of the broad band emission @xcite .
while it is now evident that accretion alone can not be responsible for all observed properties , most notably the energetic x - ray bursts , a `` hybrid '' model has been invoked that puts a thin debris disk around a highly magnetized pulsar @xcite . in this hybrid case , both the persistent luminosity and pulsar spin down are related to the mass transfer rate , @xmath81 . according to , who compare 1e 1048.1@xmath05937 s torque changes with changes in _ rxte _ pulsed flux , the scale of their variability does not obey an expected relationship for an accreting pulsar undergoing spin - down , thus presenting a challenge to models of fossil disk accretion .
it could be argued , however , that since only the pulsed flux @xmath82 was being monitored , the total x - ray luminosity @xmath83 is still an unknown quantity .
we address this by using the quantitative correlation we establish between pulsed fraction @xmath84 and total unabsorbed flux @xmath85 ( discussed in [ sec : pfvsf disc ] and shown in fig . [
fig : pulsefrac ] ) to simulate a well sampled set of phase - averaged flux data as a function of _ rxte _ pulsed flux @xmath82 , given that @xmath84 is simply @xmath86 .
the resulting `` new '' total flux , @xmath85 , is shown in figure [ fig : timing]_d _ , along with @xmath87 in figure [ fig : timing]_b _ for comparison . at its most variable in 2002 - 2004 , the absolute value of @xmath87 changed by a factor of @xmath210 in less than a year , while the maximal change in total unabsorbed flux is a factor of @xmath16 from peak to quiescence .
note that these variations are not simultaneous . for a pulsar experiencing a spin - down torque due to mass accretion while in a quasi - equilibrium `` tracking '' phase ( i.e. the axp phase )
, a strong correlation described by @xmath88 can be derived from ( * ? ? ?
* equation 3 ) , where the magnitudes of torque and @xmath87 are proportional , the radius of the magnetosphere is defined by the alfn radius which is also a function of @xmath81 , and @xmath89 .
thus , a factor of @xmath210 change in @xmath87 should be reflected by a factor of @xmath2200 simultaneous change in @xmath83 , and thus @xmath85 , if this model is correct .
clearly this is not observed , rendering this particular accretion scenario unlikely .
this point is further emphasized by the significant time offset between the changes in torque and flux .
overall , the fact that 1e 1048.1@xmath05937 s long history of highly irregular spin - down is mirrored by trends in its x - ray emission only in a broad , rather than strict , sense strongly suggests that active accretion is not happening in this case . our original purpose for making simultaneous _ cxo_/_hst _
observations was to compare low - level near - ir and x - ray flux and spectral changes , with each observed with the same instrument , in order to look for correlations . given that the near - ir source had faded considerably in the _ hst _
observations , rendering it only marginally detected in 2006 , this was not possible .
compared to the handful of near - ir detections made throughout 2002 - 2003 ( see fig .
[ fig : nearir ] ) it is clear that 2005 - 2006 marked a period of near - ir quiescence in 1e 1048.1@xmath05937 . at the time of the 2005 _ vlt _ observations in which the axp was faintly seen in @xmath14-band , its flux was consistent with the last detection in mid-2003 . with _ hst _ in 2006 we find that its @xmath44 and @xmath43 magnitudes have dropped lower than ever before observed .
there is marginal evidence for near - ir flux variability during quiescence on comparably short timescales to the variability in x - ray spectral parameters , although closely spaced near - ir monitoring observations are required to confirm this . that the near - ir faded roughly in concert with the x - ray flux
is notable , suggesting a correlation similar to that seen in axp 1e 2259@xmath78586 @xcite .
this is discussed further below .
however such a correlation can be argued to be expected in both the magnetar and disk models . in the magnetar model
it would imply that the near - ir emission is magnetospheric and hence varies , as do the x - rays , when the magnetospheric configuration varies @xcite . in the disk model ,
a correlation between near - ir emission and x - ray flux is naturally expected since the putative disk is heated via x - ray illumination @xcite , however why the torque should have varied strongly and non - simultaneously is a puzzle ( see above ) . in any case , the concurrence of the near - ir and x - ray quiescence and the timing stability is unlikely to be an accident .
similarly , the presence of the two flares and the subsequent strong torque changes are also not likely to be by chance .
1e 1048.1@xmath05937 reactivated in 2007 march , when a sudden pulsed flux increase was observed ( fig .
[ fig : timing]_c _ ) , accompanied by a large spin - up glitch , the details of which will appear elsewhere ( dib et al .
, in preparation ) . our follow - up x - ray observations with _ cxo_and
swift _ show that the spectral parameters , pulse shape , total flux , and pulsed fraction are drastically changed since this flare occurred ( figs .
[ fig : xrayspec ] and [ fig : profiles ] ) .
in particular , the total x - ray flux was @xmath27 times greater in the 2@xmath010 kev band compared with quiescence , while the pulsed fraction decreased from @xmath175% to @xmath120% .
compared to the previous long - term flares from 1e 1048.1@xmath05937 , the onset of this event took place more quickly by a factor of @xmath24 , the peak pulsed flux is @xmath110% greater , and the `` high '' state appears to be lasting longer .
optical and near - ir follow - up with the _ magellan telescope _ and _
vlt_@xcite reveal an increase in both @xmath90 and @xmath14-band of @xmath11.3 mag .
very recently , @xcite reported on _ xmm _ too observations , and revealed that nearly three months after the flare onset , the total x - ray flux has decreased slightly but is still @xmath15 times brighter than in quiescence .
previously , an anti - correlation between near - ir and x - ray flux in this axp was suggested , based on the correspondence of the highest near - ir detection with a low point in x - ray pulsed flux flux on the wrong date ; we have corrected this in our figure [ fig : nearir]_c_. ] @xcite . given the new measurements from both before and after the recent event , such an anti - correlation is highly questionable .
the significant rise in near - ir flux now appears correlated with the most recent x - ray flare .
interestingly , in 2002 , a high near - ir detection was _ between _ the two x - ray flares ; this suggests that the near - ir emission is not exactly correlated with the x - rays , but rather an enhancement can precede , follow , or last longer than an x - ray enhancement .
a near - ir brightening preceding an x - ray enhancement would rule out illuminated disk models for this source ; frequent mid- or near - ir observations are required to check this . as discussed above ,
axp 1e 2259@xmath78586 also demonstrated strongly correlated near - ir and x - ray flux decay following its large 2002 outburst @xcite , but in contrast , 4u 0142 + 61 was highly variable in the near - ir despite x - ray stability @xcite , and has shown no evidence of near - ir changes coinciding with x - ray bursts ( gonzalez et al .
, in preparation ) .
note that the claim of correlated flux decay in the case of xte j1810@xmath0197 @xcite is under dispute @xcite .
such inconsistent behaviour is puzzling for both the magnetar and disk models .
one conclusion to be drawn from figures [ fig : timing]_d _ and [ fig : timing]_e _ is that 1e 1048.1@xmath05937 s
pulsed fraction @xmath84 is strongly anti - correlated with the total x - ray flux @xmath85 , a trend already noticed by @xcite and @xcite .
now , with our larger sample of data covering a wider dynamic range , we can quantify this correlation : we find a power - law dependence of @xmath91 where @xmath92 ( see fig . [
fig : pulsefrac ] ) , with @xmath85 in the 2@xmath010 kev range . in fitting , we set the index as a free parameter , ie . not tied to any model ; therefore , some scatter is expected , as evident in the large @xmath93 produced by the fit ( @xmath94 dof ) . however , in finding the uncertainty on @xmath95 , we have scaled the @xmath84 data uncertainties to force @xmath96 , effectively scaling up @xmath97 as well .
a similar correlation has been proposed in another axp , 1rxs j170849.0@xmath0400910 @xcite , although in that case it is such that the pulsed flux remains nearly constant in the presence of @xmath250% total flux changes @xcite .
how could an increase in phase - averaged flux be met with a simultaneous decrease in pulsed fraction ?
a growing hot spot on the neutron - star surface , a result of either a changing magnitude or configuration of returning magnetospheric currents or internal processes , could at least in principle produce such an effect , if the initial hot spot size were large enough that its size as viewed from a distant observer were a significant fraction of the stellar surface .
we note that @xcite suggest that magnetar afterglows may be dominated by surface thermal , rather than magnetospheric , changes ; detailed simulations of axp pulsations using gravitational light bending and appropriate radiation beaming functions ( e.g. * ? ? ?
* ) in addition to magnetospheric scattering @xcite are needed to see if the observed correlation can be reproduced .
a prediction made by the twisted magnetosphere model of @xcite is that enhanced emission should also be spectrally harder , since both are associated with larger magnetospheric twist angles .
indeed , we observed a correlation between the total x - ray flux and spectral hardness : see figure [ fig : pulsefrac ] .
similar behaviour has also been reported in 1rxs j170849.0@xmath0400910 @xcite .
this can be interpreted as a confirmation of an important magnetar model prediction , namely that both the spectral hardness and the total flux should increase for an increasing magnetospheric twist angle @xcite .
hardening of the spectrum might also be an expected effect of increased luminosity if a large injection of thermal seed photons is repeatedly up - scattered due to resonant cyclotron scattering feedback processes in the magnetosphere , thereby shifting the photon energies higher .
recently @xcite have suggested that the hardness - intensity correlation in magnetars is a result of purely surface thermal changes , with no change in the magnetospheric configuration .
it would be interesting to apply their model to the 1e 1048.1@xmath05937 data as they span a much greater dynamic range in flux than did the sources studied by @xcite .
this is beyond the scope of our paper , however . as seen in the _ cxo
_ data obtained after 2007 march , the pulse profile of 1e 1048.1@xmath05937 changed abruptly after the x - ray flux enhancement and glitch .
several new harmonics are clearly visible , and are present in a largely energy - independent way ( fig .
[ fig : profiles ] ) .
this extra power in the higher harmonics clearly signals a complication of the surface and/or magnetospheric configuration .
this change contrasts with that seen after the 2002 outburst of 1e 2259@xmath78586 , which involved mainly an exchange of powers between the fundamental and first harmonic , with higher harmonics remaining relatively unchanged @xcite . also , the change contrasts with that seen during the 1998 sgr 1900 + 14 giant flare , in which the sgr pulse profile simplified greatly immediately post - flare @xcite .
however , the opposite occured followed the 2004 giant flare from sgr 1806@xmath020 @xcite , when a sinusoidal - to - complex evolution in pulse morphology took place , similar to that of 1e 1048.1@xmath05937 .
this diversity of behaviors in magnetars post - outburst is problematic for understanding the underlying physics ; indeed it indicates a wide variety of phenomenon phase space , which itself must be explained by models . the peculiar suppression of the 4th harmonic in the post - glitch 1e 1048.1@xmath05937 profile is particularly puzzling to us ;
perhaps it indicates an important and unchangeable symmetry in the combined emission and viewing geometries , or perhaps it is a chance occurence and will not occur again in future events . in the _ cxo _ observation made on 2007 april 6 , the one closest to and just after the 2007 march event , the x - ray spectrum of 1e 1048.1@xmath05937 could not be well fit by a simple continuum model , in contrast with the other datasets .
we identified a possible absorption line near 2.7 kev , though we note that the spectral model including this line is still not a good fit to the data .
this suggests that other low - level features may be present as well , including a possible feature at roughly half the above line s energy , @xmath11.3 kev .
the putative 2.7 kev line joins the growing list of puzzling and sometimes statistically marginal axp / sgr spectral features @xcite .
examples of some feature detections that are clearly statistically significant are , interestingly , repeatedly near @xmath113 kev , e.g. in 1e 1048.1@xmath05937@xcite , in xte j1810@xmath0197 @xcite , and very recently axp 4u 0142 + 61 ( * ? ? ? * and gavriil et al .
, in preparation ) .
the feature we see in the 2007 april data for 1e 1048.1@xmath05937 are unlike any of these .
it is perhaps more reminiscent of the absorption features seen in some `` isolated neutron stars '' ( e.g. * ? ? ?
* ; * ? ? ?
* ) , and , if similarly interpreted as an electron cyclotron line , implies a magnetic field of @xmath98 g , or @xmath99 g for a proton cyclotron line .
the latter value is of course more in line with the field expected near the surface of a magnetar .
however , the feature could also be an atomic transition ( e.g. * ? ? ?
* ) ; it is difficult to know given the present data .
regardless , the spectrum 10 days later , on 2007 april 16 , showed no such features , indicating that whatever their origin , it had subsided , and on a shorter timescale than that of any profile or flux relaxation .
the 2001 - 2002 flares observed in 1e 1048.1@xmath05937 are unique in that they had very well resolved rise times of many weeks . the 2007 april event , by contrast , involved an x - ray pulsed flux enhancement that rose at least 4 times faster . in this sense
it is perhaps more similar to the 2002 outburst of 1e 2259@xmath78586@xcite , although in that case a brief rise could have been missed by the sparse monitoring . during that outburst ,
1e 2259@xmath78586 exhibited an order - of - magnitude increase in pulsed and phase - averaged flux , as well as numerous other changes including ( but not limited to ) @xmath180 sgr - like bursts . similarly to 1e 1048.1@xmath05937 , both the total and pulsed flux increased during the 1e 2259@xmath78586 outburst , and the pulsed fraction decreased , going from 23% in quiescence down to 15% @xcite .
however , the pulsed fraction was not as tightly correlated to the pulsed flux : the pulsed fraction recovered within @xmath13 days @xcite , whereas the flux took several months to return to its pre - outburst value @xcite . in terms of energetics ,
the total energy in 2@xmath010 kev released in the 1e 2259@xmath78586 outburst was @xmath100 erg and @xmath101 erg during the rapid and gradual decay components , respectively ( see table 5 of * ? ? ? * ) . for 1e 1048.1@xmath05937
, we estimate that up to 2007 april 28 , the date of the last _ cxo _ observation , the total energy emitted is comparable , roughly @xmath102 erg ( 2@xmath010 kev ) , based on the pulsed fraction - total flux correlation and assuming a distance of 2.7 kpc @xcite , or @xmath103 erg assuming 9.0 kpc @xcite .
near - ir flux variability , possibly ( in the case of 1e 1048.1@xmath05937 ) and likely ( in 1e 2259@xmath78586 ) correlated to x - ray flux , was observed following both events . that the 2007 1e 1048.1@xmath05937event clearly involved a large rotational glitch ,
as did the 2002 1e 2259@xmath78586 event , is also a commonality , although glitches at the time of the 2001 - 2002 flares could easily have been missed because of the large amount of timing noise and our inability to phase connect the data at that time ( dib et al .
, in preparation ) .
sgr - like bursts from 1e 1048.1@xmath05937 have not been observed thus far in 2007 , but that does not preclude their existence given that our observing duty cycle is so low . similar to 1e 1048.1@xmath05937 , 1e 2259@xmath78586 exhibited a pulse profile change , albeit one involving only the fundamental and first harmonic , in constrast to the appearance of higher harmonic structure in 1e 1048.1@xmath05937 . finally , as in 1e 1048.1@xmath05937 , spectral hardening was observed during the 1e 2259@xmath78586 flare .
thus overall , the 2007 1e 1048.1@xmath05937 event has practically all properties consistent with the 2002 1e 2259@xmath78586 event , but differs from its earlier flares primarily by rise time .
the transient axp xte j1810@xmath0197 presents another interesting example for comparison .
this axp , which `` turned - on '' in late 2002 or early 2003 @xcite , has been steadily declining in x - ray flux since , reaching apparent quiescence in 2005 - 2006 @xcite .
unfortunately , very little is known about the onset of enhanced emission , ie .
, whether it was accompanied by a glitch / burst , or what the timescale of the rise was . while the exponential decay timescale of several years does resemble that of 1e 1048.1@xmath05937 s 2001 - 2002 post - flare behavior , how the most recent flare will fade is as yet undetermined .
furthermore , the peak and quiescent fluxes of xte j1810@xmath0197 differed by a much greater amount , nearly 2 orders of magnitude @xcite , while the recent total 2@xmath010 kev flux increase in 1e 1048.1@xmath05937 was by a factor of @xmath17 .
spectrally , xte j1810@xmath0197 was harder in outburst and softer in quiescence , similar to 1e 1048.1@xmath05937 , when modelled as 2-temperature blackbody @xcite or as a blackbody plus power law ( w. zhu , private communication )
. one major distinction , however , is the clear evidence for a positive correlation between pulsed fraction and total flux in xte j1810@xmath0197 @xcite .
while it may be tempting to link the transience of xte j1810@xmath0197 to 1e 1048.1@xmath05937-like post - flaring behaviour , the dissimilarities are important and remind us that we may be comparing mere coincidences . compared to the sgr giant flares , rare events in which an enormous amount of broad band energy ( @xmath104 erg ; * ? ? ?
* ) is output , the 1e 1048.1@xmath05937 event is orders of magnitude less energetic .
giant flares release most of their energy in a short ( @xmath311 s ) , initial @xmath105-ray spike , a property not observed in this event , although a small initial burst could have been missed .
furthermore , the 2007 event is prolonged , and while giant flares do gradually decay , the amount of energy in the tail is usually small compared to that of the spike @xcite .
on the other hand , the gradual evolution of the persistent properties seen before and after the 2004 december giant flare of sgr 1806@xmath020 do show some resemblance to 1e 1048.1@xmath05937 s behaviour . during this pre - flare period of enhanced burst activity in sgr 1806@xmath020 that began in mid-2003 ,
the pulsar torque , pulsed flux , total flux , and hardness increased on a @xmath1year timescale , peaking several months before the giant flare occurred in 2004 december , and continuing to decline well after the flare epoch @xcite .
the gradual changes , particularly the well - resolved factor of @xmath22 increase in total unabsorbed flux , perhaps resemble the 2001 - 2002 events of 1e 1048.1@xmath05937 more than the 2007 event . as with 1e 1048.1@xmath05937 , a correlation was found between spectral hardness and intensity in sgr 1806@xmath020 , and there appeared to be some sort of physical connection between frequency derivative ( or torque ) evolution and variability in the spectrum and pulsed flux @xcite .
however , in sgr 1806@xmath020 , there was no evidence for a correlation between the phase - averaged flux , which likely peaked around 2004 october , and pulsed fraction , which was stable in the flux - enhanced pre - flare period but was lower than average in early 2005 when the persistent flux had approached quiescent levels @xcite
. furthermore , sgr 1806@xmath020 s pulse profile was noticeably more sinusoidal while brightest @xcite , although the pulse profile likely changed as a direct results of the giant flare , rather than the slow evolution of persistent flux .
we have considered @xmath110 yrs of multiwavelength observations of 1e 1048.1@xmath05937 .
the source , in 2004 - 2006 , gradually relaxed apparently back to quiescence in most respects , only to `` awaken '' suddenly in the glitch / flare event of 2007 march . by observing 1e 1048.1@xmath05937 while quiescent , it has become apparent that previous instabilities in timing , x - ray and near - ir flux were likely all linked to the major long - term flaring events of 2001 - 2002 , whose nature is yet unknown but likely includes changes in the stellar magnetosphere s current configuration .
the asynchronous spin and x - ray flux variability we have observed is incompatible with expectations of the fossil disk accretion model . following the most recent 2007 event , we observed total x - ray flux variability that is strongly correlated with x - ray pulsed fraction , x - ray spectral hardness , and changes in near - ir flux .
these observations largely agree with expectations of the magnetar model . to date , the 1e 1048.1@xmath05937 2007 event is ongoing , as is continued multiwavelength monitoring .
we thank joe hill and lorella angelini for their help with the _ swift _ xrt data analysis , and zhongxiang wang for his help with the near - ir analysis .
we thank the anonymous referee for many judicious comments that have improved the quality of the paper .
fpg is supported by the nasa postdoctoral program administered by oak ridge associated universities at nasa goddard space flight center .
pmw gratefully acknowledges support for this work from nasa / sao through grant go7 - 8077a .
this research has made use of data obtained through the high energy astrophysics science archive research center online service , provided by the nasa / goddard space flight center , and is based on observations made with the nasa / esa hubble space telescope , obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these observations are associated with program # 10761 .
this work has been supported by sao grant go7 - 8077z , an nserc discovery grant , the canadian institute for advanced research , and le fonds qubcois de la recherche sur la nature et les technologies . ,
k. , boggs , s. e. , smith , d. m. , duncan , r. c. , lin , r. , zoglauer , a. , krucker , s. , hurford , g. , hudson , h. , wigger , c. , hajdas , w. , thompson , c. , mitrofanov , i. , sanin , a. , boynton , w. , fellows , c. , von kienlin , a. , lichti , g. , rau , a. , & cline , t. 2005 , nature , 434 , 1098 , d. m. , barthelmy , s. , gehrels , n. , kippen , r. m. , cayton , t. , kouveliotou , c. , eichler , d. , wijers , r. a. m. j. , woods , p. m. , granot , j. , lyubarsky , y. e. , ramirez - ruiz , e. , barbier , l. , chester , m. , cummings , j. , fenimore , e. e. , finger , m. h. , gaensler , b. m. , hullinger , d. , krimm , h. , markwardt , c. b. , nousek , j. a. , parsons , a. , patel , s. , sakamoto , t. , sato , g. , suzuki , m. , & tueller , j. 2005 , nature , 434 , 1107 | we report on new and archival x - ray and near - infrared ( near - ir ) observations of the anomalous x - ray pulsar 1e 1048.1@xmath05937 performed between 2001 - 2007 with the _ rossi x - ray timing explorer _ ( _ rxte _ ) , the _ chandra x - ray observatory _ ( _ cxo _ ) , the _ swift gamma - ray burst explorer _ , the _ hubble space telescope _
( _ hst _ ) , and the _ very large telescope_. during its @xmath12001 - 2004 active period , 1e 1048.1@xmath05937 exhibited two large , long - term x - ray pulsed - flux flares as well as short bursts , and large ( @xmath210@xmath3 ) torque changes . monitoring with _ rxte
_ revealed that the source entered a phase of timing stability in 2004 ; at the same time , a series of four simultaneous observations with _ cxo _ and _ hst _ in 2006 showed that its x - ray flux and spectrum and near - ir flux , all variable prior to 2005 , stabilized .
specifically , we find the 2006 x - ray spectrum to be consistent with a two - component blackbody plus power law , with average @xmath4 kev and power - law index @xmath5 at a mean flux level in the 2@xmath010 kev range of @xmath16.5@xmath6 erg @xmath7 s@xmath8 .
the near - ir flux , when detected by _
hst _
( @xmath9 mag ) and _ vlt _ ( @xmath10 mag ) , was considerably fainter than previously measured .
recently , in 2007 march , this newfound quiescence was interrupted by a sudden flux enhancement , x - ray spectral changes and a pulse morphology change , simultaneous with a large spin - up glitch and near - ir enhancement .
our _ rxte _
observations revealed a sudden pulsed flux increase by a factor of @xmath13 in the 2@xmath010 kev band . in observations with _
cxo _ and _ swift _ , we found that the total x - ray flux increased much more than the pulsed flux , reaching a peak value of @xmath27 times the quiescent value ( 2@xmath010 kev ) . with these recent data ,
we find a strong anti - correlation between x - ray flux and pulsed fraction .
in addition , we find a correlation between x - ray spectral hardness and flux . simultaneously with the radiative and timing changes
, we observed a significant x - ray pulse morphology change such that the profile went from nearly sinusoidal to having multiple peaks .
we compare these remarkable events with other axp outbursts and discuss implications in the context of the magnetar model and other models of axp emission . |
sampling and counting colorings of a graph are fundamental problems in computer science and discrete mathematics .
we consider the problem of sampling uniformly at random from the set @xmath9 of proper @xmath10-colorings of a graph @xmath11 .
a proper @xmath10-coloring @xmath12 is a labeling @xmath13 such that all neighboring vertices have different colors .
this sampling problem is also fundamental in statistical physics and corresponds to generating configurations from the gibbs distribution of the zero - temperature antiferromagnetic potts model @xcite . from the physics perspective , the underlying graph is typically taken to be the cubic lattice @xmath14 and sampling and counting reveal underlying thermodynamic properties of the corresponding physical system .
much focus has gone towards solving the sampling problem using rapidly mixing markov chains .
the idea is to design a markov chain whose stationary distribution is uniform over the set of proper colorings .
then , starting at an arbitrary coloring and simulating a random walk according to this chain for a sufficient number of steps , we get a sample from close to the desired distribution .
the number of steps required of this walk is referred to as the _ mixing time _ ( see , e.g. , @xcite ) . the chain is called _ rapidly mixing _
if the mixing time is polynomial in @xmath15 ( so it converges quickly to stationarity ) ; it is _ torpidly mixing _ if its mixing time is super - polynomial in @xmath16 ( so it converges slowly ) .
there has been a long history of studying mixing times of various chains in the context of colorings ( see , e.g. , @xcite ) .
a particular focus of this study has been on _
glauber dynamics_. for proper @xmath10-colorings this is any single - site update markov chain that connects two colorings only if they differ on at most a single vertex .
the _ metropolis _
chain @xmath17 on state space @xmath18 has transition probabilities @xmath19 , @xmath20 given by @xmath21 we may think of @xmath17 dynamically as follows . from a @xmath10-coloring @xmath12 , choose a vertex @xmath22 uniformly from @xmath23 and a color @xmath24 uniformly from @xmath25 . then recolor @xmath22 with color @xmath24 if this is a proper @xmath10-coloring ; otherwise stay at @xmath12 .
when @xmath17 is ergodic , its stationary distribution @xmath26 is uniform over proper @xmath10-colorings .
a series of recent papers have shown that @xmath17 is rapidly mixing provided the number of colors is sufficiently large compared to the maximum degree ( see @xcite and the references therein ) .
substantially less is known when the number of colors is small .
in fact , for @xmath10 small it is np - complete to decide whether a graph admits even one @xmath10-coloring .
in this paper we focus on the mixing rate of @xmath17 on rectangular regions of the cubic lattice @xmath27 .
observe that the lattice is bipartite , so it always admits a @xmath10-coloring for any @xmath28 .
it is also known that glauber dynamics connects the state space of @xmath10-colorings on any such lattice region @xcite . in @xmath29
much is known about the mixing rate of @xmath17 .
luby et al .
@xcite showed that glauber dynamics for sampling 3-colorings is rapidly mixing on any finite , simply - connected subregion of @xmath29 when the colors on the boundary of the region are fixed .
goldberg et al .
@xcite subsequently showed that the chain remains fast on rectangular regions without this boundary restriction .
substantially more is known when there are many colors : jerrum @xcite showed that glauber dynamics is rapidly mixing on any graph satisfying @xmath30 , where @xmath10 is the number of colors and @xmath31 is the maximum degree , thus showing glauber dynamics is fast on @xmath29 when @xmath32 .
it has since been shown that it is fast for @xmath33 @xcite .
surprisingly the efficiency remains unresolved for @xmath34 or @xmath35 . in higher dimensions
much less is known when @xmath10 is small .
physicists have performed extensive numerical experiments @xcite suggesting that glauber dynamics on 3-colorings is torpidly mixing when the dimension of the cubic lattice is large enough .
we prove this conjecture for the first time here by studying the mixing time of the chain on cubic lattices with periodic boundary conditions .
our focus in this paper is sampling 3-colorings of the even discrete torus @xmath6 .
this is the graph on vertex set @xmath36 ( with @xmath37 even ) with edge set consisting of those pairs of vertices that differ on exactly one coordinate and differ by @xmath38 ( mod @xmath37 ) on that coordinate . for a markov chain @xmath4 on the 3-colorings of @xmath6
we denote by @xmath39 the mixing time of the chain ; this will be formally defined in section [ sec - partitioning ] .
our main theorem is the following .
[ thm-glauber.slow.mixing ] there is a constant @xmath40 for which the following holds . for @xmath41 and @xmath42
even , the glauber dynamics chain @xmath43 on @xmath44 satisfies @xmath45 our techniques actually apply to a more general class of chains .
a markov chain @xmath4 on state space @xmath46 is _
@xmath47-local _ if , in each step of the chain , at most @xmath48 vertices have their color changed ; that is , if @xmath49 implies @xmath50 these types of chains were introduced in @xcite , where the terminology _
@xmath48-cautious _ was employed .
we prove the following , which easily implies theorem [ thm-glauber.slow.mixing ] .
[ thm-rho.local.slow.mixing ] fix @xmath51 satisfying @xmath52 .
there is a constant @xmath53 for which the following holds . for @xmath41 and @xmath42 even , if @xmath4 is an ergodic @xmath47-local markov chain on @xmath44 with uniform stationary distribution then @xmath54 here @xmath55 is the usual binary entropy function .
note that all @xmath56 satisfy @xmath52 .
we show slow mixing via a conductance argument by identifying a `` bad cut '' in the state space requiring exponential time to cross . intuitively , in sufficiently high dimension , the set of 3-colorings of the lattice is believed to naturally partition into 6 classes : each class is identified by a predominance of one ( of 3 ) colors on one of the two ( even or odd ) sublattices .
this characterization was recently rigorously verified on the infinite lattice , thereby establishing the existence of 6 distinct `` maximal entropy gibbs states '' @xcite .
that work builds heavily on technical machinery introduced by galvin and kahn @xcite showing that independent sets partition similarly in sufficiently high dimensions in that they lie primarily on the even or odd sublattices .
specifically , write @xmath57 and @xmath58 for the sets of even and odd vertices of @xmath27 ( defined in the obvious way ) and set @xmath59^d$ ] and @xmath60^d \setminus [ -(l-1),l-1]^d$ ] . for @xmath61 , choose @xmath62 from @xmath63 ( the set of independent sets of the box ) with @xmath64 . galvin and kahn showed that for @xmath65 ( for a large constant @xmath66 ) and fixed @xmath67 @xmath68 in other words , the influence of the boundary on the center of a large box persists as the boundary recedes .
notice that neither of the results of @xcite or @xcite establishing the presence of multiple gibbs states directly implies anything about the behavior of markov chains on finite lattice regions .
however , they do suggest that in the finite setting , typical configurations fall into the distinct classes described in stationarity and that it will be unlikely to move between these classes ; the remaining configurations are expected to have negligible weight for large lattice regions , even when they are finite .
galvin @xcite extended the results of @xcite , showing that in sufficiently high dimension , glauber dynamics on independent sets mixes slowly in rectangular regions of @xmath69 with periodic boundary conditions .
similar results were known previously about independent sets ; however , one significant new contribution of @xcite was showing that as @xmath3 increases , the critical @xmath70 above which glauber dynamics mixes slowly tends to @xmath71 .
in particular , there is some dimension @xmath72 such that for all @xmath73 , glauber dynamics will be slow on @xmath14 when @xmath74 .
this turns out to be the crucial new ingredient allowing us to rigorously verify slow mixing for sampling 3-colorings in high dimensions , as there turns out to be a close connection between the independent set model at @xmath74 and the @xmath5-coloring model .
note that unlike most statistical physics models , the 3-coloring problem does not have a parameter @xmath70 that can be tweaked to establish desired bounds ; this makes the proofs here significantly more delicate than the usual slow mixing arguments .
section [ subsec - overview ] provides a more detailed discussion of the elements of the proof and of some of the difficulties inherent to the sampling problem under discussion .
we begin by formalizing some definitions . given a markov chain @xmath75 on state space @xmath76 with uniform stationary distribution denoted by @xmath77 ,
let @xmath78 be the distribution of the chain at time @xmath79 given that it started in state @xmath80 .
the _ mixing time _ @xmath39 of @xmath4 is defined to be @xmath81 where @xmath82 via a well - known _ conductance argument _
@xcite , using a form of the argument derived in @xcite . as above , let @xmath4 be an ergodic markov chain on state space @xmath83 with transition probabilities @xmath84 and stationary distribution @xmath77 .
let @xmath85 and @xmath86 satisfy @xmath87 and @xmath88 .
then from @xcite we have @xmath89 let us return to the setup of theorem [ thm-rho.local.slow.mixing ]
. for even @xmath37 , @xmath6 is bipartite with partition classes @xmath57 ( consisting of those vertices the sum of whose coordinates is even ) and @xmath58 . to show torpid mixing , it is sufficient to identify a single bad cut .
we concentrate on the vertices in each 3-coloring that are colored with the first color , 0
. the objective of theorem [ thm-rho.local.slow.mixing ] will be to verify that most 3-colorings have an imbalance whereby the vertices colored 0 lie predominantly on @xmath57 or @xmath58 , and those that are roughly balanced on the two sublattices are highly unlikely in stationarity .
this is sufficient to show that the conductance is small .
accordingly let us define the set of `` balanced '' @xmath5-colorings by @xmath90 and likewise let @xmath91 by symmetry , @xmath92 .
notice that since @xmath4 updates at most @xmath93 vertices in each step , we have that if @xmath94 and @xmath95 then @xmath96 .
therefore , by ( [ conductance_bound ] ) , @xmath97 and so theorem [ thm-rho.local.slow.mixing ] follows from the following critical theorem .
[ thm - main ] fix @xmath51 satisfying @xmath52 .
there is a constant @xmath53 for which the following holds . for @xmath73 and @xmath2 even , @xmath98
for a generic @xmath99 there are regions of @xmath6 consisting predominantly of even vertices colored @xmath71 together with their neighbors , and regions consisting of odd vertices colored @xmath71 together with their neighbors .
these regions are separated by two - layer `` @xmath71-free '' moats or _
cutsets_. in section [ subsec - cutsets ] we describe a procedure that selects a particular collection of these cutsets .
our main technical result , lemma [ lem-volume.bounds.from.gk ] , asserts that for each specification of cutset sizes @xmath100 and vertices @xmath101 , the probability that a coloring has among its associated cutsets a collection @xmath102 with @xmath103 and with @xmath104 surrounded by @xmath105 is exponentially small in the sum of the @xmath106 s .
this lemma is presented in section [ subsec-results.from.galvinkahn ] and theorem [ thm - main ] is derived from it in section [ subsec - proof_of_thm ] .
the main thrust of @xcite is the proof of a result that is essentially ( but not quite ) the case @xmath107 of lemma [ lem-volume.bounds.from.gk ] .
one difficulty we have to overcome in moving from a gibbs measure argument to a torpid mixing argument is that of going from bounding the probability of a configuration having a single cutset to bounding the probability of it having an ensemble of cutsets .
another difficulty is that the cutsets we consider in these ensembles can be topologically more complex than the connected cutsets that are considered in @xcite . in part ,
both of these difficulties are dealt with by the machinery developed in @xcite .
we use a `` peierl s argument '' to prove lemma [ lem-volume.bounds.from.gk ] . by carefully modifying each @xmath108 inside its cutsets
, we can exploit the fact that the cutsets are @xmath71-free to map @xmath12 to a set @xmath109 of many different @xmath110 .
if the @xmath109 s were disjoint for distinct @xmath12 s , we would essentially be done , having shown that there are many more @xmath5-colorings in total than @xmath5-colorings in @xmath111 . to control the possible overlap
, we define a flow @xmath112 supported on pairs @xmath113 with @xmath114 in such a way that the flow out of each @xmath115 is @xmath38 .
any uniform bound we can obtain on the flow into elements of @xmath46 is then easily seen to be a bound on @xmath116 .
we define the flow via a notion of approximation modified from @xcite . to each cutset
@xmath117 we associate a set @xmath118 that approximates the interior of @xmath117 in a precise sense , in such a way that as we run over all possible @xmath117 , the total number of approximate sets used is small .
then for each @xmath119 and each collection of approximations @xmath120 , we consider the set of those @xmath99 with @xmath121 and with @xmath122 the approximation to @xmath105 .
we define the flow so that if this set is large , then @xmath123 is small for each @xmath12 in the set . in this way we control the flow into @xmath124 corresponding to each collection of approximations @xmath125 ; since the total number of approximations is small , we control the total flow into @xmath124 . in the language of statistical physics ,
this approximation scheme is a _ course - graining _ argument .
the details appear in section [ sec - proofs ] .
the main results of @xcite and @xcite are proved along similar lines to those described above .
one of the difficulties we encounter in moving from these arguments on independent sets to arguments on colorings is that of finding an analogous way of modifying a coloring inside a cutset in order to exploit the fact that it is @xmath71-free .
the beginning of section [ sec - proofs ] ( in particular claims [ claim-shift.works ] and [ claim-unique.reconstruction ] ) describes an appropriate modification that has all the properties we desire .
we describe a way of associating with each @xmath99 a collection of minimal edge cutsets , following the approaches of @xcite and @xcite .
first we need a little notation .
write @xmath23 for the vertex set of @xmath6 and @xmath126 for its edge set . for @xmath127 , write @xmath128 for the set of edges in @xmath126 that have one end in @xmath80 and one end outside @xmath80 ; @xmath129 for @xmath130 ; @xmath131 for the set of vertices in @xmath80 that are adjacent to something outside @xmath80 ; @xmath132 for the set of vertices outside @xmath80 that are adjacent to something in @xmath80 ; @xmath133 for @xmath134 ; @xmath135 for @xmath136 and @xmath137 for @xmath138 .
further , for @xmath139 set @xmath140 .
we abuse notation slightly , identifying sets of vertices of @xmath23 and the subgraphs they induce . for each @xmath99
set @xmath141 .
note that @xmath142 is an independent set ( a set of vertices no two of which are adjacent ) .
for each component @xmath143 of @xmath144 or @xmath145 and each component @xmath66 of @xmath146 , set @xmath147 and @xmath148 . evidently @xmath66 is connected , and @xmath149 consists of @xmath143 , which is connected , together with a number of other components of @xmath146 , each of which is connected and joined to @xmath143 , so @xmath149 is connected also .
it follows that @xmath117 is a minimal edge - cutset in @xmath6 .
say that @xmath117 is _ even _ if @xmath143 is a component of @xmath144 and _ odd _ otherwise .
define @xmath150 , the _ interior _ of @xmath151 , to be the smaller of @xmath152 ( if @xmath153 , take @xmath154 ) .
the cutsets @xmath117 associated to @xmath12 depend only on the independent set @xmath142 , and coincide exactly with the cutsets associated to an independent set in @xcite .
we may therefore apply the machinery developed in @xcite for independent set cutsets in the present setting .
in particular , from ( * ? ? ?
* lemmas 3.1 and 3.2 ) we know that for each @xmath155 there is a collection of associated cutsets @xmath156 such that either @xmath157 or we have the analogue of ( [ cutset.conditions ] ) with even replaced by odd .
set @xmath158 . from here on whenever @xmath159 is given we assume that @xmath160 is its associated independent set and that @xmath156 is a particular collection of cutsets associated with @xmath12 and satisfying ( [ cutset.conditions ] ) .
numerous properties of @xmath161 are established in ( * ? ? ?
* lemmas 3.3 and 3.4 ) .
we list some here that will be of use in the sequel . that the cutsets are indeed @xmath71-free regions is established by ( [ contour.prop.1 ] ) .
@xmath162 @xmath163 @xmath164 @xmath165 for @xmath166 and @xmath167 set @xmath168 and set @xmath169 .
profile _ of a collection @xmath170 is a vector @xmath171 with @xmath172 for all @xmath173 .
given a profile vector @xmath174 set @xmath175 our main lemma ( c.f .
* lemma 3.5 ) ) is the following .
[ lem-volume.bounds.from.gk ] there are constants @xmath176 such that the following holds . for all even @xmath42 , @xmath73 and profile vector @xmath174 , @xmath177 we will derive theorem [ thm - main ] from lemma [ lem-volume.bounds.from.gk ] in section [ subsec - proof_of_thm ] before proving the lemma in section [ sec - proofs ] . from here on we assume that the conditions of theorem [ thm - main ] and lemma [ lem-volume.bounds.from.gk ] are satisfied ( with @xmath72 sufficiently large to support our assertions ) .
we begin with an easy count that dispenses with colorings where @xmath178 is small .
set @xmath179 [ lem - bounding_small ] @xmath180 .
_ proof : _ for any @xmath181 and @xmath182 , let @xmath183 be the number of components in @xmath184 , where for @xmath185 ( or @xmath58 ) , @xmath186 we begin by noting that by @xmath57-@xmath58 symmetry @xmath187 where the sum is over all pairs @xmath181 , @xmath182 with no edges between @xmath188 and @xmath189 and satisfying @xmath190 and @xmath191 . indeed ,
once we have specified that the set of vertices colored @xmath71 is @xmath192 , we have a free choice between @xmath38 and @xmath193 for the color at @xmath194 , and we also have a free choice between the two possible colorings of each component of @xmath195 . a key observation is the following . for @xmath188 and @xmath189 contributing to the sum in ( [ inq-small.count ] ) , @xmath196 to see this ,
let @xmath66 be a component of @xmath197 .
if @xmath198 consists of a single vertex , then ( depending on the parity of @xmath22 ) we have either @xmath199 or @xmath200 and so @xmath201 .
otherwise , let @xmath202 be an edge of @xmath66 with @xmath203 ( and so @xmath204 ) .
if @xmath22 has @xmath10 edges to @xmath189 and @xmath205 has @xmath206 to @xmath188 , then ( since there are no edges from @xmath188 to @xmath189 ) we have @xmath207 or @xmath208 .
( here we are using that in @xmath6 , if @xmath209 then there is a matching between all but one of the neighbors of @xmath205 and @xmath22 . )
since @xmath22 has @xmath210 edges to @xmath211 and @xmath212 has @xmath213 edges to @xmath214 we have that @xmath215
. from this , ( [ inq-number.of.remaining.comps ] ) follows . inserting ( [ inq-number.of.remaining.comps ] ) into ( [ inq-small.count ] ) and bounding @xmath216 and @xmath217 by the maximum values of @xmath218 and @xmath219 ( valid since @xmath185 ( or @xmath58 ) satisfies @xmath220 for sufficiently large @xmath221 . in ( [ int4 ] )
we use the chernoff bound @xmath222}{m\choose i } \leq 2^{h({\beta})m}$ ] for @xmath223 ; in ( [ int5 ] ) we use @xmath52 .
using @xmath224 , the lemma follows .
we now consider @xmath225 by lemma [ lem - bounding_small ] and @xmath57-@xmath58 symmetry , theorem [ thm - main ] reduces to bounding ( say ) @xmath226 let @xmath227 be the set of @xmath228 such that there is a @xmath229 with @xmath230 ( we think of such cutsets as being topologically non - trivial ( `` nt '' ) ; see @xcite for an explanation of this ) and also let @xmath231 .
we assert that @xmath232 and @xmath233 this gives ( [ inq - remaining ] ) and so completes the proof of theorem [ thm - main ] .
both ( [ large2 ] ) and ( [ large3 ] ) are corollaries of lemma [ lem-volume.bounds.from.gk ] , and the steps are identical to those that are used to bound the measures of `` @xmath234 '' and `` @xmath235 '' in ( * ? ? ?
* section 3.3 ) . with the sum below running over all vectors @xmath174 of the form @xmath236 with @xmath167 and @xmath237 , and with the inequalities justified below , we have @xmath238 giving ( [ large2 ] ) .
we use lemma [ lem-volume.bounds.from.gk ] in ( [ large1 ] ) .
the factor of @xmath239 is for the choices of @xmath240 and @xmath22 .
the verification of ( [ large3 ] ) involves finding an @xmath241 $ ] and a set @xmath242 of cutsets with the properties that @xmath243 , @xmath244 for each @xmath245 and @xmath246 .
the measure of @xmath247 is then at most the product of a term that is exponentially small in @xmath8 ( from lemma [ lem-volume.bounds.from.gk ] ) , a term corresponding to the choice of a fixed vertex in each of the interiors , and a term corresponding to the choice of the collection of lengths .
the second term will be negligible because @xmath248 is small and the third will be negligible because all @xmath249 have similar lengths .
more precisely , for @xmath250 and @xmath251 we have @xmath252 ( by ( [ contour.prop.3 ] ) ) and so @xmath253 the second inequality is from ( [ cutset.conditions ] ) and the third follows since @xmath254 .
set @xmath255 .
note that @xmath248 is empty for @xmath256 ( again by ( [ contour.prop.3 ] ) ) and for @xmath257 so we may assume that @xmath258 since @xmath259 , there is an @xmath173 such that @xmath260 choose the smallest such @xmath173 set @xmath261 .
we have @xmath262 ( this follows from the fact that each @xmath249 satisfies @xmath263 ) and @xmath264 the first inequality follows from that fact that @xmath265 ; the second follows from ( [ quad.scale ] ) and the fact that each @xmath117 has @xmath266 .
we therefore have @xmath267 for some @xmath268 with @xmath206 satisfying ( [ inner.property.2 ] ) , with @xmath269 with @xmath270 for each @xmath24 and with @xmath173 satisfying ( [ inner.property.1 ] ) . with the sum below running over all profile vectors
@xmath174 satisfying ( [ inner.property.1 ] ) , ( [ inner.property.2 ] ) , ( [ inner.property.3 ] ) and ( [ inner.property.4 ] ) we have @xmath271 the right - hand side of ( [ largeeventriv ] ) is , by lemma [ lem-volume.bounds.from.gk ] , at most @xmath272 the factor of @xmath273 is an upper bound on the number of choices for @xmath173 ; the factor of @xmath274 is for the choice of the @xmath275 s ; and the factor @xmath276 is for the choice of the @xmath206 ( distinct ) @xmath277 s . by ( [ inner.property.1 ] ) and the second inequality in ( [ inner.property.2 ] ) we have ( for @xmath3 sufficiently large ) @xmath278 so that in fact the right - hand side of ( [ largeeventriv ] ) is at most @xmath279 taking @xmath206 as small as possible we see that this is at most @xmath280 and taking @xmath173 as large as possible we see that this is at most @xmath281 . putting these observation together we obtain ( [ large3 ] ) .
much of what follows is modified from @xcite and @xcite .
our strategy is as follows .
let @xmath282 be given .
set @xmath283 .
we will show @xmath284 from which the lemma follows by a telescoping product . to obtain ( [ induction ] ) we define a one - to - many map @xmath285 from @xmath286 to @xmath287 .
we then define a flow @xmath288 supported on pairs @xmath113 with @xmath121 satisfying @xmath289 and @xmath290 this easily gives ( [ induction ] ) . for each @xmath291 , define @xmath292 , the _ shift in direction @xmath293 _ , by @xmath294 , where @xmath295 is the @xmath293th standard basis vector if @xmath296 and @xmath297 if @xmath298 . for @xmath127 write @xmath299 for @xmath300 . for @xmath301
set @xmath302 .
let @xmath267 be given .
arbitrarily pick @xmath303 and set @xmath304 .
write @xmath305 for the map from @xmath306 to @xmath306 that sends @xmath71 to @xmath71 and transposes @xmath38 and @xmath193 . for each @xmath291 and
@xmath307 define the function @xmath308 by @xmath309 and set @xmath310 . _
proof : _ an easy case analysis verifies @xmath312 .
we begin with the observation that the graph @xmath313 is bipartite with bipartition @xmath314 .
this follows from ( [ contour.prop.0 ] ) . by , @xmath315 and so for each component @xmath316 of @xmath313
, @xmath12 is constant on @xmath317 and on @xmath318 and in neither case does it take on the value @xmath71 . if @xmath327 and @xmath328 then @xmath329 and @xmath330 ( we will justify this in a moment ) . since @xmath331 we have @xmath332 and we can not ever have @xmath333 , so @xmath326 in this case . to see that @xmath334 , we consider subcases . if @xmath335 then @xmath336 . if @xmath337 then @xmath323 .
finally , if @xmath338 then @xmath339 ; and @xmath340 is either @xmath71 or @xmath341 depending on whether @xmath342 equals @xmath71 or @xmath343 ( @xmath342 can not equal @xmath341 ) . if @xmath347 and @xmath348 then @xmath330 ( as in the second case above ) and @xmath349 .
since @xmath332 , we have @xmath350 . noting that it is not possible to have both @xmath351 , we finally treat the case where @xmath347 and @xmath352 . in this case
suppose ( for a contradiction ) that @xmath336 .
this can only happen if @xmath354 .
if @xmath355 , we have a contradiction immediately . otherwise , we have @xmath356 and so ( since @xmath357 ) @xmath358 , also a contradiction .
this verifies @xmath312 . because @xmath359 is disjoint from the interiors of the remaining cutsets in @xmath156 and the operation that creates the elements of @xmath360 only modifies @xmath12 inside @xmath149 it follows that @xmath311 . _
proof : _ following @xcite , we may reconstruct @xmath12 as follows .
@xmath362 we define the one - to - many map @xmath285 from @xmath286 to @xmath287 by setting @xmath363 for a particular direction @xmath293 . to define @xmath364 and @xmath293 , we employ the notion of approximation also used in @xcite and based on ideas introduced by sapozhenko in @xcite . for @xmath301 ,
we say @xmath365 is an _ approximation _ of @xmath117 if @xmath366 @xmath367 and @xmath368 where @xmath369 .
note that from ( [ contour.prop.0 ] ) and ( [ contour.prop.2 ] ) , @xmath370 is an approximation of @xmath117 . before stating our main approximation lemma , which is a slight modification of ( * ? ? ?
* lemma 2.18 ) , it will be convenient to further refine our partition of cutsets .
to this end set @xmath371 note that by ( [ contour.prop.2 ] ) we have @xmath372 so @xmath373
. [ lem-k.comp.approx ] for each @xmath374 , @xmath375 and @xmath22 there is a family @xmath376 satisfying @xmath377 and a map @xmath378 such that for each @xmath379 , @xmath380 is an approximation for @xmath117 .
we are now in a position to define @xmath364 and @xmath293 .
our plan for each fixed @xmath381 is to fix @xmath382 and @xmath383 and to consider the contribution to the sum in ( [ eq-flow.in ] ) from those @xmath384 with @xmath385 ( where for each @xmath12 , @xmath117 is a particular @xmath386 ) .
we will try to define @xmath364 in such a way that each of these individual contributions to ( [ eq-flow.in ] ) is small ; to succeed in this endeavor we must first choose @xmath293 with care . to this end ,
given @xmath379 , set @xmath387 where @xmath388 in the map guaranteed by lemma [ lem-k.comp.approx ] . to motivate the introduction of @xmath389 and @xmath390 ,
note that for @xmath391 we have ( by ( [ contour.prop.0 ] ) and ( [ contour.prop.2 ] ) ) @xmath392 @xmath393 @xmath394 and @xmath395 it follows that for each @xmath391 , @xmath396 contains all vertices whose location in the partition @xmath397 is as yet unknown .
we choose @xmath398 to be the smallest @xmath293 for which both of @xmath399 and @xmath400 hold .
this is the direction that minimizes the uncertainty to be resolved when we attempt to reconstruct @xmath12 from the partial information provided by @xmath401 , @xmath293 and @xmath188 .
( that such an @xmath293 exists is established in ( * ? ? ? * ( 49 ) and ( 50 ) ) by an easy averaging argument ) .
note that @xmath293 depends on @xmath117 but not @xmath160 .
now for each @xmath267 let @xmath229 be a particular cutset with @xmath402 .
let @xmath109 be as defined before , with @xmath293 as specified above .
define @xmath403 and @xmath404 and for each @xmath121 set @xmath405 note that for @xmath384 , @xmath123 depends on @xmath149 but not on @xmath12 itself . since @xmath406 partitions @xmath149 we easily have ( [ eq-flow.out ] ) . to obtain ( [ induction ] ) and so ( [ inq-bound.on.inner ] ) we must establish ( [ eq-flow.in ] ) .
fix @xmath374 , @xmath375 such that @xmath407 .
fix @xmath408 and @xmath291 . for @xmath12 with @xmath409 write @xmath410 if it holds that @xmath385 and @xmath411 .
we claim that with @xmath412 and @xmath374 fixed , for @xmath381 @xmath413 we could extract this directly from @xcite , but for the convenience of the reader we describe a proof below .
write @xmath414 for the set of all @xmath415 such that @xmath416 , @xmath385 , @xmath411 and @xmath121 and set @xmath417 .
say that a triple @xmath418 is _ good _ for @xmath12 if it satisfies the following conditions .
@xmath419 @xmath420 and @xmath421 we begin by establishing that @xmath422 always has a good triple . in view of lemma
[ lem - core - exists ] there is a triple @xmath425 that is good for @xmath12 and which has @xmath426 as small as possible .
choose one such , say @xmath427 .
set @xmath428 and @xmath429 .
lemma [ lem - unlabeled_main_bound ] below establishes an upper bound on @xmath123 in terms of @xmath430 , @xmath431 , @xmath432 and @xmath433 , and lemma [ lem - reconstruction ] shows that for each choice of @xmath434 , @xmath435 there is at most one @xmath12 contributing to the sum in the lemma .
these two lemmas combine to give ( [ inq-black.box ] ) .
_ proof : _ we follow ( * ? ? ?
* from just before ( 55 ) to just after ( 60 ) ) , making superficial changes of notation .
the inequality in lemma [ lem - unlabeled_main_bound ] is the @xmath5-coloring analogue of the main inequality of @xcite .
the key observation that makes this inequality useful is the following .
_ proof : _ in ( * ? ? ?
* ( 56 ) and following ) it is shown that @xmath434 and @xmath435 determine @xmath439 via @xmath440 and so @xmath149 ( via @xmath441 ) .
but then by claim [ claim-unique.reconstruction ] @xmath434 and @xmath435 determine @xmath12 .
lemmas [ lem - unlabeled_main_bound ] and [ lem - reconstruction ] together now easily give ( [ inq-black.box ] ) : @xmath442 we have now almost reached ( [ eq-flow.in ] ) . with the steps justified below we have that for each @xmath443 @xmath444 completing the proof of ( [ eq-flow.in ] ) .
in the first inequality , @xmath445 is over all choices of @xmath374 , @xmath375 , @xmath293 and @xmath188 . in ( [ overcount ] )
, we note that there are @xmath446 choices for @xmath188 , @xmath447 choices for @xmath293 and @xmath448 choices for each of @xmath374 , @xmath375 ( this is because @xmath449 , by ( [ contour.prop.3 ] ) ) , and we apply ( [ inq-black.box ] ) to bound the summand . in ( [ overcount2 ] ) we use lemma [ lem-k.comp.approx ] .
finally in ( [ overcount3 ] ) we use @xmath450 ( again by ( [ contour.prop.3 ] ) ) to bound @xmath451 .
r. bubley , m. dyer and c. greenhill , beating the @xmath452 bound for approximately counting colourings : a computer - assisted proof of rapid mixing , _ proc . of the 9th acm - siam soda 98 _ , 355363 .
m. dyer , a. frieze and m. jerrum , on counting independent sets in sparse graphs , _ siam j. comp .
_ * 31 * ( 2002 ) , 15271541 .
d. galvin , sampling independent sets from the discrete torus , submitted ( preprint available on the web at www.math.upenn.edu/@xmath453dgalvin/pdf/indmixzd.pdf ) .
d. galvin and j. kahn , on phase transition in the hard - core model on @xmath27 , _ comb .
* 13 * ( 2004 ) , 137164 .
wang , r.h .
swendsen , and r. koteck , three - state antiferromagnetic potts models : a monte carlo study , _ phys .
b _ * 42 * ( 1990 ) 24652474 .
d. welsh , complexity : knots , colourings , and counting , _ london math .
note series _ * 186 * , 1993 . | we study local markov chains for sampling 3-colorings of the discrete torus @xmath0 .
we show that there is a constant @xmath1 such that for all even @xmath2 and @xmath3 sufficiently large , certain local markov chains require exponential time to converge to equilibrium . more precisely , if @xmath4 is a markov chain on the set of proper @xmath5-colorings of @xmath6 that updates the color of at most @xmath7 vertices at each step and whose stationary distribution is uniform , then the convergence to stationarity of @xmath4 is exponential in @xmath8 .
our proof is based on a conductance argument that builds on sensitive new combinatorial enumeration techniques . |
because of the intensive use of carbon dioxide in industry and research @xcite , it has become necessary to determine its thermodynamic , physical and chemical properties on an extended range of temperatures
. significant effort has been deployed to build up a database through observations and theoretical calculations @xcite . from the former point of view , we mention the case of the accurate measurements due to giauque & egan @xcite and from the latter point of view , the derivation based on the classical version of the theory of lattice dynamics , which predicts the heat capacity of carbon dioxide in the range of temperatures 15@xmath4@xmath1@xmath450 k @xcite , is in a very good agreement with that obtained through observations @xcite . however , such a good agreement is still out of reach for some other properties of carbon dioxide due to difficulties from both experimental and theoretical points of view .
for instance , the empirical determination of the latent heat of sublimation at low temperatures remains a major obstacle because of the difficulty in eliminating the superheating of the gas @xcite .
similarly , by way of example , the lagrangian classical treatment of the two - dimensional rigid rotor is intractable and the theoretical determination of the heat capacity , mentioned above , had been made possible at only sufficiently low temperatures ( @xmath1@xmath450 k ) when the harmonic approximation is valid @xcite . with that said ,
much work has to be done in order to determine further properties of carbon dioxide particularly at low temperatures , such properties are still missing in the best compendia .
we will exploit the data available in @xcite , which we refer to as g&e , and show that it is possible to evaluate the heat of sublimation @xmath5 and vapor pressure @xmath6 at temperatures 5@xmath2@xmath1@xmath2195 k. a key prerequisite is the determination of the heat of sublimation at @xmath1=0 k ( @xmath7=@xmath8 ) .
stull calculated an average value of @xmath5 by the method of least squares using the vapor pressure data measured by different workers @xcite and obtained a value of 26.3 kj@xmath3mol^-1^ ( = 6286 cal@xmath3mol^-1^ ) for 139@xmath2@xmath1@xmath2195 k @xcite .
however , the literature citations listed in @xcite show that stull did not extract data from g&e , which is even more accurate and includes data concerning the heat capacity of the solid carbon dioxide and other data that could be used to obtain @xmath5 at different temperatures . by contrast
, g&e have evaluated @xmath5 at 194.67 k using partly their measured data and available data for @xmath5 at lower temperatures @xcite .
they evaluated the integral of the heat capacity of the solid ( change in the enthalpy ) graphically from a smooth curve through their measured data and obtained a value for @xmath5 that is merely 10 cal@xmath3mol^-1^ higher than their measured value @xmath9=6030@xmath105 cal@xmath3mol^-1^ ( 25230@xmath1021 j@xmath3mol^-1^ ) .
they also evaluated the entropies of the gas and solid at 194.67 k and reached an _
excellent _ agreement between experimental data and statistics ( the experimental & spectroscopic values of the entropy of the gas @xmath11 they obtained were 47.59 & 47.55 cal@xmath3k^-1^@xmath3mol^-1^ , respectively , constituting a proof of the third law @xcite ) .
however , this cumbersome procedure had prevented them from carrying out a systematic evaluation of the latent heat and entropy at temperatures covering the range of their measured data .
furthermore , this procedure ( the graphical evaluation ) adds a human error , which is an unknown factor . in this paper
we will carry out a systematic evaluation of the fore - mentioned physical quantities on a more extended range of temperatures than that of g&e using 1 ) a computer algebra system ( cas ) , which eliminates the human error and allows an excellent adjustment of the parameters in order to achieve a better accuracy , as well as 2 ) an established formula for the vapor pressure .
it will be shown below that our reevaluated value of @xmath12 is 6030.4 cal@xmath3mol^-1^ ( 25231 j@xmath3mol^-1^ ) .
the data for the relevant quantities will be tabulated at temperatures incremented by 5 k and plotted . moreover ,
the generating codes will be provided , which allow the evaluation of any quantity at any given temperature within minutes of time . in this work , we will be relying on measured data by different workers and on some empirical formulas derived by graphical interpolation .
since some of these data are provided without accuracy and some other lack accuracy due to personal error , it will be difficult to assign accuracy to our results , as is the case in most compendia .
some values of @xmath6 ( in torr ) will be given with one significant digit while other values with 2 or 3 significant digits .
the values of @xmath5 ( of the order of 26000 j@xmath3mol^-1^ ) will be given with five digits without decimals , assuming an error not higher than 0.35% .
the accuracy of the results for @xmath6 and @xmath5 can be read by comparing with the available measured data .
[ [ heat - of - sublimation - at - pmbt0 . ] ] * heat of sublimation at @xmath13 . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + throughout this paper , we use the units and symbols recommended by the _ international union of pure and applied chemistry _ ( _ iupac _ ) @xcite .
the energy is given in j and in cal = 4.184 j , the pressure in torr , and the temperature in k. since the original data were given in calories , we perform our evaluations in this unit , taking @xmath14=1.98724 cal@xmath15k^-1^@xmath3mol^-1^ , then convert the results to joules .
the g&e heat capacity measurements , shown in the codes ( appendix ) , extend from 15.52 to 189.78 k. on such a large interval there is no best equation that will represent the data @xcite .
g&e worked on a smooth curve through the data but did not describe it . in order to represent the data ,
the alternative is to subdivide the interval into sufficiently small intervals and represent the data by a polynomial on each sub - interval in such a way that the polynomial pieces blend smoothly making a spline @xcite .
matlab provides spline curve via the command ` spline(x , y ) ` ( see appendix section ) .
it returns the piecewise polynomial form of the cubic spline interpolant with the not - a - knot end conditions , having two continuous derivatives and breaks at all interior data sites except for the leftmost and the rightmost one .
the values of the spline at the breaks ` spline(x , y , x(i ) ) ` coincide with the data values ` y(i ) ` .
cubic splines are more attractive for interpolation purposes than higher - order polynomials @xcite
. we will deal with molar physical quantities labeled by the subscripts @xmath16 & @xmath17 to differentiate between the solid and gaseous phases .
we denote by @xmath5 the latent heat of sublimation and by @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 ( @xmath24=@xmath25 ) , the internal energy , free energy , chemical potential , volume , enthalpy , entropy , respectively .
we take the zero of rotational energy to be that of the @xmath26=0 state and the zero of vibrational energy to be that of the ground state , meaning that a molecule at rest in the gas has an energy of zero at vanishing temperature ( @xmath27=0 ) .
let @xmath8 be the heat of sublimation at @xmath1=0 which is , according to our energy convention , the binding energy of the particles of the solid ( @xmath28=@xmath29=@xmath30=@xmath31=@xmath32@xmath40 ) .
the excellent agreement between the experimental & spectroscopic values of @xmath11 at 194.67 k is due to g&e accurate measurements and to the success of debye s theory at low temperatures .
g&e used debye s formula to evaluate @xmath33 for 0@xmath2@xmath1@xmath215 k. however , they did not explain their choice for debye s temperature @xmath34 . in this work ,
the energy and entropy of the solid for temperatures below 15.52 k are extrapolated by substitution of the debye heat capacity formula .
moreover , we will rely on suzuki & schnepp s assertion that the molar heat capacities of the solid carbon dioxide ( @xmath35 & @xmath36 ) are equal within an error of 10@xmath37 per cent for such small temperatures @xcite .
finally , we fix @xmath38 by equating the heat capacity due to debye with that measured by g&e at 15.52 k ( 0.606 cal@xmath3k^-1^@xmath3mol^-1^ ) .
solving the equation using a cas we find @xmath38=139.59 k. the matlab codes provided in the appendix are split into three parts . in part ( i ) , ` cd ` represents the debye heat capacity .
the vectors ` t ` & ` cp ` show the temperature data sites used by g&e ( [email protected] k ) and the corresponding measured heat capacities ( [email protected] cal@xmath3k^-1^@xmath3mol^-1^ ) , respectively .
these g&e data sites are extended by the temperature vector ` u ` and the corresponding debye heat capacity vector ` v ` , respectively .
the last two lines evaluate , at the temperature vector ` tn ` , the spline through the extended data sites ( ` t ` , ` cp ` ) , the integrals @xmath40=@xmath41=@xmath42 ( vector ` i ` ) and @xmath43=@xmath44 ( vector ` j ` ) , with @xmath45 ` tn ` . the heat of sublimation @xmath8 is determined upon solving the equation @xmath46=@xmath47 at any given temperature for which the measured @xmath5 is known .
the lead we had followed seeking for higher accuracy led us to select the value of @xmath5=@xmath48 cal@xmath3mol^-1^ at @xmath49 k ( * ? ? ?
* eucken & donath ) & @xcite .
we find @xmath8=@xmath50 cal@xmath3mol^-1^ and the calculation is shown below . with @xmath46=@xmath51 & @xmath47=@xmath52 ,
the equation @xmath46=@xmath47 reduces to @xmath8=@xmath53 . upon solving the clapeyron equation for @xmath54
we obtain @xmath54=@xmath55+p\,v_{s}$ ] , and finally @xmath56 we will make use of the g&e empirical equation to evaluate @xmath6 & @xmath57 at 170 k @xmath58\quad ( 154\leq t\leq 196\,{\rm k}),\ ] ] ( @[email protected]@xmath61@xmath62 , @xmath63=8.69903@xmath61@xmath64 , @xmath65=0.001588@xmath61@xmath66 , @[email protected]@xmath61@xmath68@xmath61@xmath66 ) , and obtain @xmath69=74.59 torr . since @xmath70=25.55 cm^3^@xmath3mol^-1^ @xcite , the last term @xmath71=0.06 cal@xmath3mol^-1^ is neglected .
the term including @xmath5 equals 6190/([email protected])=337.08 cal@xmath3mol^-1^ , and @xmath72 are the 85000^th^ components of the vectors ` i ` & ` j ` : @xmath73=`i(85000)-170*j(85000)`[email protected] cal@xmath3mol^-1^. now , we make our first hypothesis concerning the vapor .
we assume the validity of the first order virial expansion neglecting thus the next terms , and this has always been the case for carbon dioxide @xcite at such low temperatures .
we have then @xmath74 thereby we can show that the term @xmath75 in ( [ eps ] ) is the free energy of an ideal is under the above assumption @xmath76 , implying @xmath75=@xmath77 . ] gas evaluated at the point @xmath78=@xmath79 . for the molecule of co@xmath80 we have @xmath75=@xmath81 , which is the sum of the translational , rotational and four vibrational contributions @xmath82=@xmath83 @xcite . with our choice of the origin of the energy ,
these contributions write @xmath84+\theta_{\rm r}/3\}\,;$ \\
$ a_{{\rm v}i}=r\,t\ln[1-{\rm exp}(-\theta_{{\rm v}i}/t)]\,;$ & \quad $ ( t\geq 5 { \rm k})\;\&\;(i=1 - 3)$ , \\ \end{tabular}\ ] ] with @xmath85=7.575455@xmath86 in si units ( = @xmath87 ) and @xmath88=0.561 , @xmath89=954 , @xmath90=1890 , @xmath91=3360 k. we have then @[email protected] cal@xmath3mol^-1^ leading with the previously evaluated terms to @xmath8=6273.4 cal@xmath3mol^-1^. + [ [ vapor - pressure . ] ] * vapor pressure .
* + + + + + + + + + + + + + + + + + from now on we will assume @xmath8=6274 cal@xmath3mol^-1^ ( 26250 j@xmath3mol^-1^ ) . upon substituting ( [ v ] ) & ( [ fg ] ) into @xmath8=@xmath93 ( @xmath46=@xmath47 ) and rearranging the terms we obtain @xmath94 /rt\}\,,\ ] ] where @xmath95=@xmath96 , @xmath97=@xmath98\,({\rm i}$]=@xmath99 , and is coupled with the nuclear partition function and the above expression of @xmath100 no longer holds for @xmath1 of the order of @xmath101
however , as @xmath1 increases the separation of the two partition functions becomes possible @xcite .
the above formula for @xmath100 has been derived using the euler - maclaurin expansion and can be used safely for @xmath1 of the order of 5 k and higher values . ]
@xmath102=@xmath103/(2\,\theta_{\rm r})\,$ ] .
assuming that @xmath104 follows berthelot s equation @xcite @xmath105\,p({\rm torr})\ ] ] ( where @[email protected]@xmath107 k@xmath108 and , in order to express @xmath109 in cal@xmath3mol^-1^ , we take @[email protected]/([email protected]@xmath61760 ) k / torr ) , we have solved numerically both equation ( [ ptw ] ) and its linearized form and the results coincide up to an insignificant error . upon substituting exp@xmath111$]=@xmath112 into ( [ ptw ] ) , the linearized equation yields @xmath113p_{\,\rm ideal}/t\ } } \quad ( t\geq 5\,{\rm k})\,,\ ] ] where @xmath114 ( in torr ) is the corresponding pressure for an ideal gas @xmath115/rt\}\,.\ ] ] table [ tab1 ] & fig .
1 compare values of the vapor pressure derived in this work ( tw ) with those of g&e ( eqs .
( [ ptw ] ) & ( [ gep ] ) ) .
we have evaluated ( [ gep ] ) at temperatures below the left - end point 154 k , as shown in table [ tab1 ] , and the formula remains applicable , however , for temperatures above 110 k ; below this temperature , equation ( [ gep ] ) diverges from ( [ ptw ] ) . the third column ( a ) of table
[ tab1 ] shows values of the vapor pressure evaluated using antoine s equation @xcite .
the constants @xmath116=6.81228 , @xmath117=1301.679 & @[email protected] of antoine s equation have been evaluated by the national institute of standards and technology ( nist ) @xcite from g&e data .
the equation writes @xmath119\}\quad ( 154.26\leq t\leq 195.89\,{\rm k}),\ ] ] where @xmath120=@xmath121 , @xmath122=@xmath123 .
+ [ tab1 ] 1.0llll 65 & 3.4e12 & na & _ 3.3e12(c ) _ + 70 & 1.2e10 & na & _ 1.3e10(c ) _ + 75 & 2.8e9 & na & _ 3.0e9(c ) _ + 80 & 4.2e8 & na & _ 4.7e8(c ) _ + 85 & 4.7e7 & na & _ 5.2e7(c ) _ + 90 & 3.9e6 & na & _ 4.3e6(c ) _ + 95 & 2.6e5 & na & _ 2.8e5(c ) _ + 100 & 1.4e4 & na & _ 1.5e4(c ) _ + 105 & 6.8e4 & na & _ 7.3e4(c ) _ + 110 & 0.003 & _
0.003(c ) _ & _ 0.003(c ) _ + 115 & 0.01 & _
0.01(c ) _ & _ 0.01(c ) _ + 120 & 0.03 & _
0.03(c ) _ & _ 0.03(c ) _ + 125 & 0.09 & _
0.09(c ) _ & _ 0.09(c ) _ + 130 & 0.2 & _ 0.2(c ) _ & _ 0.2(c ) _ + 135 & 0.6 & _
0.6(c ) _ & _ 0.6(c ) _ + 140 & 1.4 & _
1.4(c ) _ & _ 1.4(c ) _ + 145 & 3.1 & _
3.1(c ) _ & _ 3.1(c ) _ + 150 & 6.4 & _
6.4(c ) _ & _ 6.3(c ) _ + 155 & 12.5 & 12.6(c ) & 12.5(c ) + 160 & 23.6 & 23.6(c ) & 23.5(c ) + 165 & 42.8(c ) & 42.7(c ) & 42.4(c ) + 170 & 74.6(c ) & 74.6(c ) & 74.1(c ) + 175 & 126(c ) & 126(c ) & 125(c ) + 180 & 206(c ) & 207(c ) & 205(c ) + 185 & 329(c ) & 330(c ) &
328(c ) + 190 & 511(c ) & 513(c ) & 511(c ) + 195 & 776(c ) & 781(c ) & 777(c ) + + + from table [ tab1 ] we establish the following results . equations ( [ gep ] ) & ( [ ap ] )
are still valid beyond their assumed ranges of validity ; the ranges are now extended right down below their left - end points to include temperatures above 110 and 65 k , respectively .
moreover , the vapor behaves as a polyatomic ideal gas for temperatures below 155 k. an instance of calculation is provided in the codes provided in part(ii ) of the appendix , which show the evaluation of the ideal - gas pressure equation ( [ pi ] ) & the real - gas pressure equation ( [ ptw ] ) at 160 , 180 & 194.67 k. the evaluated pressures are represented by the 3-vectors ` pi ` and ` ptw ` , respectively .
+ [ [ heat - of - sublimation . ] ] * heat of sublimation . * + + + + + + + + + + + + + + + + + + + + + + combining different thermodynamic entities we establish the equation @xmath124\,p(t)\,,\ ] ] where the last two terms add a correction for gas imperfection , @xmath125 is the vapor pressure and @xmath126 is the ideal - gas enthalpy given by @xmath126=@xmath127@xmath60@xmath128 $ ] @xmath60@xmath129 $ ] ( eq .
( [ fg ] ) ) . looking for extreme values
we can first ignore the correction for gas imperfection then justify it later .
we have solved graphically the equation @xmath130=0 ( @xmath131=@xmath132 ) and obtained the values 57.829 k for @xmath1 & 6503.58 cal@xmath3mol^-1^ for @xmath5 as shown in fig .
2 . we will assume @xmath133=6503.6 cal@xmath3mol^-1^ ( 27211 j@xmath3mol^-1^ ) .
tables [ tab2 ] & [ tab1 ] , however , show that at 57.829 k the vapor behaves as an ideal gas , and this justifies the omission of the correction terms in @xmath134=0 . substituting ( [ b ] ) into ( [ l ] ) , this latter splits into two equations whether we evaluate the vapor pressure using ( [ gep ] ) or ( [ ptw ] ) @xmath135\ , p_{\,\rm g\&e}\quad ( 154\leq t\leq 196\,{\rm k}),\label{gel } \\
l_{\,\rm tw } & = & \epsilon_{0}-\delta h_{s}+ h_{g}+r\,\ell_{1}[1-(3\ell_{2}/t^{2})]\ , p_{\,\rm tw}\quad ( t\geq 5 { \rm k})\,.\label{ltw}\end{aligned}\ ] ] equations ( [ gel ] ) & ( [ ltw ] ) are plotted in fig .
2 . in the codes provided in part(iii ) of the appendix
, we evaluate the r.h.s of ( [ ltw ] ) at 160 , 180 & 194.67 k ( 3-vector ` ltw ` ) .
the value of the latent heat obtained at 194.67 k is 6030.4 cal@xmath3mol^-1^ ( 25231 j@xmath3mol^-1^ ) or 6030.6 cal@xmath3mol^-1^ ( 25232 j@xmath3mol^-1^ ) whether we calculate the r.h.s of ( [ ltw ] ) or ( [ gel ] ) .
+ [ tab2 ] 1.0lll 0 & 26250 & na + 5 & 26394 & na + 10 & 26538 & na + 15 & 26676 & na + 20 & 26804 & na + 25 & 26914 & na + 30 & 27005 & na + 35 & 27077 & na + 40 & 27133 & na + 45 & 27172 & na + 50 & 27197 & na + 55 & 27209 & na + 60 & 27210 & na + 65 & 27201 & na + 70 & 27183 & na + 75 & 27158 & na + 80 & 27128 & na + 85 & 27091 & na + 90 & 27048 & na + 95 & 27002 & na + 100 & 26951 & na + 105 & 26896 & na + 110 & 26836 & _ 26836 _ + 115 & 26773 & _ 26773 _ + 120 & 26707 & _ 26707 _ + 125 & 26637 & _ 26637 _ + 130 & 26565 & _ 26565 _ + 135 & 26488 & _ 26488 _ + 140 & 26408 & _ 26408 _ + 145 & 26325 & _ 26325 _ + 150 & 26239(c ) & _ 26239(c ) _ + 155 & 26149(c ) & 26149(c ) + 160 & 26055(c ) & 26055(c ) + 165 & 25958(c ) & 25958(c ) + 170 & 25855(c ) & 25855(c ) + 175 & 25745(c ) & 25745(c ) + 180 & 25629(c ) & 25629(c ) + 185 & 25504(c ) & 25504(c ) + 190 & 25368(c ) & 25368(c ) + 195 & 25221(c ) & 25220(c ) + + + in concluding , it was of interest to further compare our results for the pressure with those used by stull @xcite that , as already stated , are less accurate than g&e values . at temperatures 138.8 , 148.7 , 153.6 , 158.7 k , we read from @xcite the values 1 , 5 , 10 , 20 torr for the pressure , while our evaluated values ( eq . ( [ ptw ] ) ) are 1.16 , 5.30 , 10.42 , 20.12 torr , respectively .
finally , values of the entropy of the solid at the tabulated temperatures @xmath1=5@xmath136 k ( 1@xmath2@xmath136@xmath239 , positive integer ) form a sub - vector of ` j ` and are obtainable upon executing the codes ` q=2500:2500:97500 ; j(q ) ` . for instance , @xmath137=`j(80000)`=14.07 , @xmath138=`j(90000)`=15.50 and @xmath139=`j(97335)`=16.52 cal@xmath3k^-1^@xmath3mol^-1^ ( 58.87 , 64.85 & 69.12 j@xmath3k^-1^@xmath3mol^-1^ , respectively ) .
concerning the numerical approach , given the accurate data for the heat capacity at constant pressure of carbon dioxide and some available data for the heat of sublimation , we employed the method of splines to generate and evaluate a smooth curve representing the heat capacity data . dealing with a large number of data sites
, we preferred to use cubic splines , which are more attractive for interpolation purposes than higher - order polynomials @xcite .
once the curve set , we proceeded to the evaluation of the change of the enthalpy and entropy of the solid .
the evaluation of the relevant physical quantities concerning the vapor was rather straightforward using almost fresh formulas from the thermodynamic literature @xcite .
we used matlab to execute the task and the calculated entities were used in subsequent vapor pressure and heat of sublimation evaluations .
now , concerning the theoretical approach , we mainly derived a formula for the vapor pressure including a correction for gas imperfection and effects for internal structure , as well as a formula for the heat of sublimation with same purposes .
the author acknowledges comments and suggestions by an anonymous referee , which helped to improve the manuscript . + * appendix * + this section is devoted to provide the main matlab codes , as a part of the numerical method , leading to the results shown in this paper .
+ _ * part(i ) * _ + part(i ) shows the data sites used by g&e ( [email protected] k ) & ( [email protected] cal@xmath3k^-1^@xmath140mol^-1^ ) .
we evaluate the spline through the extended data sites ( ` t ` , ` cp ` ) , the integrals @xmath40=@xmath41=@xmath42 ( vector ` i ` ) and @xmath43=@xmath44 ( vector ` j ` ) , with @xmath45 ` tn ` .
.... syms x z real ; f=(12/(x^3))*int((z^3)/(exp(z)-1),z,0,x ) ; g=(3*x)/(exp(x)-1 ) ; a = f - g ; cd=3 * 1.98724*a ; u=0.01:0.01:15.25 ; xn=139.59./u ; v = real(double(subs(cd , x , xn ) ) ) ; t=[0 u 15.52 17.30 19.05 21.15 23.25 25.64 27.72 29.92 32.79 35.99 39.43 43.19 47.62 52.11 56.17 60.86 61.26 66.24 71.22 76.47 81.94 87.45 92.71 97.93 103.26 108.56 113.91 119.24 124.58 130.18 135.74 141.14 146.48 151.67 156.72 162.00 167.62 173.36 179.12 184.58 189.78 ] ; cp=[0 v 0.606 0.825 1.081 1.419 1.791 2.266 2.676 3.069 3.555 4.063 4.603 5.195 5.794 6.326 6.765 7.269 7.302 7.707 8.047 8.370 8.703 8.984 9.189 9.421 9.671 9.893 10.07 10.27 10.44 10.69 10.88 11.08 11.27 11.45 11.64 11.84 12.07 12.32 12.57 12.82 13.05 ] ; tn=0.001:0.002:196.001 ; spcp = spline(t , cp , tn ) ; i=0.002*cumsum(spcp ) ; j=0.002*cumsum(spcp./tn ) ; .... _ * part(ii ) * _ + we evaluate the ideal - gas and real - gas pressures ( eqs .
( [ pi ] ) & ( [ ptw ] ) ) at 160 , 180 & 194.67 k. the evaluated pressures are represented by the 3-vectors ` pi ` and ` ptw ` , respectively . ....
eps=6274 ; t=[159.999 179.999 194.669 ] ; m=[80000 90000 97335 ] ; ms = i(m)-(t.*j(m ) ) ; pc=7.575455*(10 ^ 5 ) ; l1=9 * 304.1/(128 * 72.8 * 760 ) ; l2=6*(304.1 ^ 2 ) ; s = exp(ms./(1.98724*t ) ) ; ztr=(1/(2 * 0.561))*((t.^(7/2 ) ) . *
( ones(size(t))+((0.561/3)./t ) ) ) ; zv=(1./((ones(size(t))-exp(-954./t)).^2 ) ) . *
( 1./(ones(size(t))-exp(-1890./t ) ) ) . *
( 1./(ones(size(t))-exp(-3360./t ) ) ) ; pi=((760/101325)*pc).*((ztr.*zv ) . *
( s.*exp(-eps./(1.98742*t ) ) ) ) ; v=(l1*((ones(size(t))-(l2./(t.^2))).*(pi./t)))+ ones(size(t ) ) ; ptw = pi./v ; t = 160 180 194.67 pi = 23.604 204.845 739.817 ptw = 23.632 206.308 754.942 .... _ * part(iii ) * _ + we evaluate the the heat of sublimation ( eq .
( [ ltw ] ) ) 160 , 180 & 194.67 k. the output is the 3-vector ` ltw ` .
.... it = ones(size(t ) ) ; h1=954./(exp(954./t)-it ) ; h2=1890./(exp(1890./t)-it ) ; h3=3360./(exp(3360./t)-it ) ; hv=1.98724*((2*h1)+h2+h3 ) ; hg=((3.5 * 1.98724).*t)+hv-(((1.98724 * 0.561)/3)*it ) ; gi=(1.98724*l1).*(it-((3*l2)./(t.^2))).*ptw ; ltw = eps - i(m)+hg+gi ; t = 160 180 194.67 ltw(cal / mol ) = 6227.4 6125.5 6030.4 ltw(j / mol ) = 26055 25629 25231 .... 99
air liquide ` http://www.airliquide.com/en/business/products/gases/gasdata/ ` ; + j.b .
calvert ` http://www.du.edu/~jcalvert/phys/carbon.htm ` .
meyers , van dusen ( 1933 ) nat .
j. research * 10 * : 381 ; henning , stock ( 1921 ) zeits .
f. physik * 4 * : 226 ; siemens ( 1913 ) ann .
d. physik * 42 * : 871 ; eucken , donath ( 1926 ) zeits .
f. physik .
chemie * 124 * : 181 ; heuse , otto ( 1931 ) ann .
d. physik * 9 * : 486 ; ( 1932 ) * 14 * : 181 .
giauque wf , egan cj ( 1937 ) j. chem .
phys . * 5 * : 45 .
stull dr ( 1947 ) organic compounds , ind .
eng . chem . *
39 * : 517 .
suzuki m , schnepp o ( 1971 ) j. chem .
phys . * 55 * : 5349 .
schnepp o , jacobi n ( 1975 ) lattice dynamics of molecular solids . in dynamical properties of solids ,
north - holland publishing company , amsterdam .
sataty ya , ron a ( 1974 ) j. chem .
phys . * 61 * : 5471 .
grigoriev is , meilikhov ez ( editors ) ( 1997 ) handbook of physical quantities .
crc press boca raton florida .
gray de ( coordinating editor ) ( 1972 ( 1982 reissue ) ) american institute of physics handbook 3rd edition .
mcgraw - hill book company new york .
national institute of standards and technology ` http://webbook.nist.gov/chemistry/ ` .
engineering tool box ` http://www.engineeringtoolbox.com/ ` .
schaeffer cd , jr ( 1989 ) data for general inorganic , organic , and physical chemistry ` http://wulfenite.fandm.edu/data%20/data.html ` ; + the wired chemist ` http://wulfenite.fandm.edu/data%20/ ` .
moore jh , spencer nd ( editors ) ( 2001 ) encyclopedia of chemical physics and physical chemistry .
institute of physics publishing bristol & philadelphia .
stull dr , westrum ef , sinke gc ( 1969 ) the chemical thermodynamics of organic compounds .
john wiley & sons , new york .
mills i , cvita t , homann k , kuchitsu k ( 1993 ) quantities , units and symbols in physical chemistry 2^nd^ edition .
blackwell scientific publications oxford ; + iupac ` http://www.iupac.org/reports/1993/homann/index.html ` . de boor c ( 2001 ) a practical guide to splines .
springer - verlag , new york ; schilling rj , harris sl ( 2000 ) applied numerical methods for engineers . brooks / cole , pacific grove , ca ; meir a , sharma a ( 1973 ) spline functions and approximation theory ( proceedings of a symposium , university of alberta , 1972 ) .
basel , stuttgart , birkhuser ; gu c ( 2002 ) smoothing spline anova models .
springer - verlag , new york .
mcquarrie da ( 1976 ) statistical physics .
harper & row publishers new york .
herzberg g ( 1989 ) molecular spectra & molecular structure .
krieger pub .
co. malabar , fla . | we investigate the empirical data for the vapor pressure ( 154@xmath0@xmath1@xmath2196 k ) and heat capacity ( 15.52@xmath0@[email protected] k ) of the solid carbon dioxide . the approach is both theoretical and numerical , using a computer algebra system ( cas ) . from the latter point of view
, we have adopted a cubic piecewise polynomial representation for the heat capacity and reached an excellent agreement between the available empirical data and the evaluated one .
furthermore , we have obtained values for the vapor pressure and heat of sublimation at temperatures below 195 right down to 0 k. the key prerequisites are the : 1 ) determination of the heat of sublimation of 26250 j@xmath3mol^-1^ at vanishing temperature and 2 ) elaboration of a ` linearized ' vapor pressure equation that includes all the relevant properties of the gaseous and solid phases .
it is shown that : 1 ) the empirical vapor pressure equation derived by giauque & egan remains valid below the assumed lower limit of 154 k ( similar argument holds for antoine s equation ) , 2 ) the heat of sublimation reaches its maximum value of 27211 j@xmath3mol^-1^ at 58.829 k and 3 ) the vapor behaves as a ( polyatomic ) ideal gas for temperatures below 150 k. |
at sufficiently high intensities , matter no longer reacts linearly to light , and may re - emit at integer multiples ( _ harmonics _ ) of the frequency of the incoming source @xcite . according to perturbation theory , the intensity of the harmonics decreases exponentially with their order .
however , the spectrum of atoms and molecules exposed to very intense , typically infrared , laser pulses was found to present unexpectedly high harmonics @xcite , and its shape was observed to have a plateau extending non - perturbatively over many orders of magnitude a process known as _ high harmonic generation _ ( hhg ) @xcite .
the light emitted in this manner is coherent and may reach the extreme ultraviolet and soft x - ray frequency regime opening the path towards the coherent manipulation and control of matter at its natural time scale .
these properties can be of paramount importance for many technological and scientific purposes in ultrafast science
most notably , the generation of attosecond pulse trains or single isolated attosecond pulses , or the external seeding of free - electron lasers @xcite .
these advances allow to follow the electron dynamics @xcite .
unsurprisingly , a big effort has been devoted to first understanding the underlying physics , and then to controlling and fine - tuning the efficiency and spectral characteristics of the harmonic radiation .
the latter can be done by modifying the non - linear medium , or by post - processing the signal with filters , gratings , etc .
however , one advantageous alternative is to modify the characteristics of the parent pulse , which obviously will modify the spectral outcome . the most obvious manner of doing this is by systematically varying the defining parameters of this parent pulse @xcite .
however , the current availability of advanced pulse shaping tools @xcite , together with the development of closed - loop quantum control techniques @xcite , provides a superior optimization alternative @xcite . in this manner
, the successful selective enhancement of harmonic orders could be achieved when using a hollow fiber container for the generating medium @xcite .
gas jet ( _ free focusing _ ) geometries were also employed @xcite , but although some degree of control could be achieved ( for example , the extension of the cut - off frequency ) , the very selective order enhancement or depletion obtained with the hollow fibers was not observed .
this fact seems to imply that this type of selective enhancement can not be explained from the single - atom response only ; instead , the propagation effects present in the capillary set - up apparently play a fundamental role . in this context .
quasiphase matching ( qpm ) approach is commonly used to achieve independent phase control between multiple high - harmonics @xcite .
however , a full interpretation of the optimisation mechanisms can only be achieved with theoretical input , for which purpose one may utilise quantum simulations in combination with the theoretical branch of quantum optimal control @xcite ( qoct ) .
recently , schaefer and kosloff @xcite have addressed this task , showing the possibility of enhancing the emission at desired frequencies for simple few level systems and one - dimensional one.electron system . here
we address , by first principles simulations based on time - dependent density functional theory , the role of many electron interactions in the high harmonic generation , and provide compelling evidence that a single - atom hhg emission can be enhanced by few orders of magnitude in a controlled manner , with standard laser shaping techniques available in many experimental labs .
the _ three - step _ model successfully describes the key features of hhg @xcite , at least qualitatively .
it combines a quantum description of the ionisation and recombination of the electrons , with a classical description of the intermediate electronic propagation .
@xcite developed an approximate , mostly analytical , quantum description based on the _ strong field approximation _
( sfa ) : it neglects the contribution of excited bound states , the depletion of the ground state , and considers the continuum electrons to be free of the influence of the parent ion .
a more precise approach consists of propagating schrdinger s equation @xcite , an expensive method that quickly becomes prohibitive as we increase the number of electrons .
for one - electron problems the approach is perfectly feasible , and this fact has encouraged the use of the _ single active electron _ approximation ( sae ) , which assumes that only one electron is significantly disturbed by the field , and its evolution may be computed on the combination of the laser field and the potential originating by the parent ion .
this single electron picture is commonly used to describe recollision processes and hhg in atoms and relies on the fact that under hhg conditions there is one electron being ionised .
however this does nt imply the other electrons do not play a role .
there is indeed no formal justification for the use of the sae and in fact , many - body effects have been shown recently to play an important role in hhg providing an explanation of why heavier atoms emit stronger hhg than lighter ones @xcite and the giant enhancement of he hhg at 100 ev @xcite . however , the sae has been successful in explaining a few features of the hhg spectra such as the spectral cutoff , the phase structure of the spectrum and the prediction of the generation of attosecond pulses . in spite of all those experimental and theoretical efforts ,
it is clear that the topic of selective hhg generation deserves further microscopical analysis , and in this work , we explore the optimisation possibilities of one and two electron systems ( the hydrogen and the helium atoms ) , isolating the single atom response , so that we can learn how much selectivity in the hhg spectrum can be obtained from isolated atoms that can nicely complement qpm schemes in enhancing further the hhg selective emission . for this purpose ,
we employ a global optimisation scheme that acts on the envelope of the generating pulse , maintaining the fundamental frequency and minimising undesired ionisation ( and for molecules also dissociation ) processes . for the case of helium , we report results obtained both with the single active electron approximation , and with time - dependent density - functional theory ( tddft ) @xcite , in order to assess the influence of the electron - electron interaction in the optimisation process . as
many - electron effects may be relevant , tddft appears as the ideal framework to capture them in the hhg spectra ( see for example ref .
@xcite ) as it combines a very good compromise between accuracy and computational efficiency .
the present optimisation scheme has been implemented in the first - principles code octopus @xcite , that allows the treatment of more complex molecular and extended systems .
however for the purposes of the present work , it is better to stay at the simplest level of one and two electron systems . larger electronic systems would offer a wider range of possibilities for hhg enhancement .
within the dipole approximation and in the length gauge , the experimentally measured harmonic spectrum can be theoretically approximated by the following formula : @xmath0 i.e. the power spectrum of the second derivative of the expectaction value of the dipole moment @xmath1 ( although see ref .
@xcite for a discussion on the pertinence of using , alternatively , the first derivative or the dipole moment itself ) .
this object is given by : @xmath2 where @xmath3 is the ( static ) ionic potential , @xmath4 is the number of electrons , @xmath5 is the laser pulse electric field , and @xmath6 is the polarization vector ( see the methods appendix below for some extra details about the theory ) .
note that this expression can be read as both the acceleration of the electronic system , and as the corresponding back - reaction of the nucleus ( or nuclear center of mass , if we are dealing with a molecule ) .
this is not surprising since the electromagnetic emission must be related with a charge acceleration .
the expression corresponds , except for the mass factor , with the classical _ force _ acting on the nucleus , considered as a point particle .
we will therefore rewrite eq .
( [ eq : hhgfunction ] ) as : @xmath7 where @xmath8 is the fourier transform of : @xmath9 from a tddft perspective , the use of this force functional is convenient since it can be explicitly written as a density functional : @xmath10 usually , the electric field @xmath5 is factorised into a sinusoidal function determining the fundamental frequency @xmath11 , and an _ envelope _
function @xmath12 that determines the overall laser - pulse shape : @xmath13 this factorisation and the concept of a _ fundamental frequency _ is meaningful for long and quasi - monochromatic pulses , but as the technology has reached the optical period limit , it has started to lose its relevance .
nevertheless , the existence of a fundamental frequency is implicit when speaking of harmonics , which are defined as radiation at integer multiples of precisely that frequency .
these will only be well defined if the envelope function is _ smooth _ compared to the sinusoidal term , i.e. its frequencies are much lower than @xmath11 .
therefore , in this work , we investigate the possibility of manipulating the envelope function @xmath12 , leaving the sinusoidal factor @xmath14 unchanged , in order to influence the shape of the hhg spectrum .
this manipulation can not be unconstrained , as the envelope must be composed of frequencies much lower than @xmath11 .
moreover , we have searched for solutions that preserve the fluence or total integrated energy of the pulse : @xmath15 this type of requirement of a specific structure for the solution field ( in terms of frequencies , fluence , etc . )
can be respected following essentially two routes : by imposing penalties on undesired features of the pulses in the definition of the optimising function , or by constraining from the beginning the search space .
this latter option can be achieved by establishing a parametrisation of the control field ( in this case , the envelope ) that enforces the required condition , and is the route that we have chosen for this work .
the search for the optimum is in this manner performed in the space of parameters that determine the control field ; the remaining necessary ingredient is the definition of a _ merit _ function that encodes the physical requirements .
moreover , the assumption of low frequencies for @xmath12 implies that the spectrum of @xmath16 is concentrated around @xmath11 .
therefore , the @xmath17 term in eqs .
( [ eq : d2dmudt2 ] ) , ( [ eq : ft ] ) , and ( [ eq : d2dmu ] ) does not contribute to the hhg spectrum in the region we are interested in and in the following we will safely ignore it . we have shown how the hhg spectrum may be explicitly computed solely in terms of the system electronic density @xmath18 . for systems with more than one electron , this fact is convenient since it allows to use time - dependent density - functional theory @xcite ( tddft ) ( see methods ) .
one can substitute the propagation of the real interacting system by the propagation of a system of fictitious non - interacting electrons whose density is however identical to that of the real one , despite the fact that its wave function is a single slater determinant .
hereafter , we will restrict the discussion to one and two - electron systems , the extension to systems with larger number of electrons is straightforward in the tddft framework .
the one - electron case obviously does not need a tddft treatment , although it may be treated as such by considering one single occupied orbital .
for such one - orbital problem , the exchange and correlation potential must cancel the hartree term : @xmath19(\vec{r},t ) = - v_{\rm h}[n](\vec{r},t)\,,\ ] ] so that the resulting equation reduces to the initial schrdinger equation . for two - electron systems ,
we use the exact - exchange approximation ( exx ) to the xc term , which for this two - electron case amounts to setting : @xmath19(\vec{r},t ) = - \frac{1}{2}v_{\rm h}[n](\vec{r},t)\,,\ ] ] note that in this form tddft is identical to time - dependent hartree - fock that provides a food description of the non - linear properties of two - electron systems except for the description of charge - transfer excitations ( see for example ref .
@xcite ) . the electric field amplitude will be determined by the specification of a set of @xmath20 parameters @xmath21(t)$ ] .
the evolution of the tdks system is in consequence also governed by the choice of parameters @xmath22 , i.e. the orbitals and density are functionals of the parameters : @xmath23 $ ] , @xmath24.$ ] we may then use the tools of qoct to find the set @xmath22 that maximizes a given target function @xmath25 , defined in terms of a functional of the density of the system , i.e. : @xmath26 = \tilde{f}[n[u]]\,.\ ] ] this functional @xmath27 is designed to favour the desired behaviour of the system ( in this case , a certain form of the hhg spectrum , to be detailed below ) . note that it is defined in terms of the system density , and not in terms of the full many - body wave function .
this definition ensures that the substitution of the real by the kohn - sham system in the optimization entails no further approximation .
the theory must however be developed in terms of a functional of the kohn - sham orbitals , which can be easily defined as : @xmath28 = \tilde{f}[\mu\varphi^*\varphi]\,,\ ] ] where @xmath29 is the occupation of the orbital , i.e. one or two for one- or two - electron calculations , respectively .
we must now choose a form for @xmath30 in such a way that its maximization leads to the desired hhg optimization , namely the selective increase of one harmonic peak that should leave the neighboring ones as low as possible .
there is substantial liberty to design @xmath30 , and it is not evident what functional form should lead to better results . @xmath28
= \sum_{k } \alpha_k h[\varphi](k\omega_0 ) \,,\ ] ] where @xmath31 takes a positive value for the harmonic to be enhanced , and negative values for the ones that we wish to reduce .
however , this choice proved to be problematic , since the modulation of the source signal with the envelope function leads to displacements , sometimes substantial , of the harmonic peaks with respect to the precise integer multiples @xmath32 .
a general definition that solves this problem ( and that includes the previous one as a particular case ) , is : @xmath33 = \int\!\!{\rm d}\omega \alpha(\omega)h[\varphi](\omega ) = \int\!\!{\rm d}\omega \alpha(\omega ) \vert \vec{f}[\varphi](\omega)\vert^2\,,\ ] ] where we have made explicit the fact that both @xmath34 and @xmath35 , defined in eqs .
( [ eq : hhgfunction ] ) and ( [ eq : ft ] ) are functionals of the time - dependent evolution for the system .
the function @xmath36 permits to establish some finite window around each harmonic peak @xmath32 , that will be positive for the harmonic orders that we want to enhance , and negative for the ones that we want to reduce .
finally , a third option is to seek for the maximum of the spectrum in these frequency windows around the harmonic orders , i.e. : @xmath37 = \sum_{k } \alpha_k \max_{\omega \in [ k\omega_0 - \beta , k\omega_0 + \beta ] } \lbrace \log_{10 } h[\varphi](\omega)\rbrace \,,\ ] ] where the real number @xmath38 determines the size of the window .
once the function @xmath25 has been defined ( through the definition of the target functional @xmath30 ) , it remains to use some optimization algorithm to find the optimal @xmath22 set .
there are numerous options , and we may divide them on two groups , depending on whether or not they require the computation of the gradient of @xmath25 in addition of the computation of the function itself .
the methods that employ the gradient are of course more efficient , as long as this gradient can itself be computed efficiently .
the simplest scheme is steepest descents , but one can also use conjugate gradients or , in our case , the broyden - fletcher - goldfarb - shanno ( gfbs ) quasi - newton method . for the function @xmath25
, the gradient is given by @xcite : @xmath39 = 2 \int_0^t\!\!{\rm d}t\ ; \nabla \varepsilon[u](t ) { \rm i m } \langle \chi[u](t)\vert \hat{\vec{r}}\cdot\hat{\pi } \vert \varphi[u](t)\rangle\,.\ ] ] this expression uses an auxiliary orbital @xmath40 $ ] defined by the following equations of motion : @xmath41(\vec{r},t ) & = & -\frac{1}{2}\nabla^2 \chi[u](\vec{r},t ) + v^*_{\rm ks}[n[u]](\vec{r},t)\chi[u](\vec{r},t ) \\\nonumber & & + \hat{k}[\varphi[u](t ) ] \chi[u](\vec{r},t ) \\ \label{eq : chit } & & -i \frac{\delta f[\varphi[u]]}{\delta \varphi^*[u](\vec{r},t)}\ , , \\ \label{eq : bc - final } \chi[u](\vec{r},t ) & = & 0\,.\end{aligned}\ ] ] the operator @xmath42[t]]$ ] is defined as : @xmath43(t ) ] \chi[u](\vec{r},t ) = -4 { \rm i}\varphi[u](\vec{r},t ) % \times % \\ % & & { \rm i m } \int{\rm d}^3r'\;\chi^*[u](\vec{r}',t ) f_{\rm hxc}[n[u](t)](\vec{r},\vec{r } ' ) \varphi[u](\vec{r}',t)\,,\ ] ] where @xmath44 is the so - called _ kernel _ of the kohn - sham hamiltonian , which , for our two - electron case treated within the exx approximation , is given by : @xmath45(\vec{r},\vec{r } ' ) = \frac{1}{2}\frac{1}{\vert \vec{r}-\vec{r}'\vert}$ ] , and is null for the one - electron case ( zeroing the full @xmath46 operator ) .
the functional derivative of @xmath30 , needed in eq .
( [ eq : chit ] ) , for the hhg target defined in eq .
( [ eq : hhgtarget1 ] ) , is : @xmath47(t)\cdot\nabla v(\vec{r})\;\varphi(\vec{r},t)\,,\ ] ] where @xmath48(t ) = 2\mu \int\!\!{\rm d}\omega\;\alpha(\omega ) { \rm re}\left [ \vec{f}[\varphi](\omega)e^{-i\omega t}\right]$ ] .
however , we can not compute this functional derivative for the target defined in eq .
( [ eq : hhgtarget2 ] ) due to the presence of the `` max '' function , at least in a simple and efficient manner . in consequence ,
when using this target definition we could not make use of any of the optimization algorithms that make use of the gradient , and turned to the gradient - free newuoa algorithm @xcite , which is a very efficient scheme for optimization problems with a moderate number of degrees of freedom , such as the ones treated here .
in fact , for the optimizations attempted in this work , we observed numerically that the target of eq .
( [ eq : hhgtarget2 ] ) provided much better results than the target of eq .
( [ eq : hhgtarget1 ] ) , and therefore we will only show below gradient - free optimizations ; in a forthcoming publication , where the target is the hhg cut - off extension , we will present gradient - based optimizations based on a target of the type given in eq .
( [ eq : hhgtarget1 ] ) . therefore , it remains to specify the set of parameters @xmath22 that determine the envelope of the electric fields .
the requirements are : ( i ) the envelope should have a given cut - off frequency ; ( ii ) the field should smoothly approach zero at the end points of the propagation time interval ; ( iii ) the total integral of the field should be zero , and ( iv ) the _ fluence _ or total integrated intensity of the pulse should have a constant pre - defined value .
this last condition is merely a choice , and not a physical constraint that experimentalists face .
the first step to parametrize the applied time - dependent electroc field @xmath5 in order to enforce all these constraints is to expand the envelope in a fourier series : @xmath49 where @xmath50 this series fixes the maximum possible ( _ cut - off _ ) frequency to @xmath51 .
note that it explicitly omits the zero - frequency term , which is a desired restriction , in order to fulfill : @xmath52 .
the manifold spanned by the @xmath53 coefficients is not yet , however , our parameter space , since we still want to enforce the conditions @xmath54 , and fix the fluence : @xmath55 . as discussed in ref .
@xcite , these conditions reduce the degrees of freedom from @xmath56 to @xmath57 : the final parameters @xmath58 are finally the hyperspherical angles that characterize a sphere of constant fluence , determining the fourier coeffiencients : @xmath59 $ ] . in all the oct calculations to be shown below we have fixed the wavelength of the fundamental frequency @xmath11 to 800 nm , a very common value used in laboratories equipped with a ti : sapphire source . the total pulse duration is fixed to 200 cycles , @xmath60 , which corresponds to 533 fs approximately .
the envelope function @xmath61 is then restricted to have frequencies no larger than @xmath62 .
the fluence [ eq . ( [ eq : fluence ] ) ]
is then fixed to a value ( around 5.0 a.u . )
that ensures a sufficiently non - linear response of both the hydrogen and helium atoms , while not causing a substantial ionization .
fixing the fluence does not imply fixing the peak intensity ; however the simultaneous existence of a maximum frequency puts a limit on it ; in practice , the peak intensities observed in the optimal pulses are in the range of 5 10@xmath63 - 10@xmath64 w/@xmath65 .
the optimization are started from randomly generated sets of parameters @xmath22 .
since the procedure finds local maxima , we have performed several searches for each case , choosing afterwards the best among them . in order to have some `` reference '' to compare the optimal run to
, we define a reference pulse as : @xmath66 i.e. a cosinoidal envelope that peaks at @xmath67 with a value of @xmath68 , chosen to fulfill the fluence condition .
the calculated hhg spectrum emitted by the hydrogen and helium atoms , irradiated by the reference pulse , is depicted in fig . [
fig : h - he ] .
note that there is a range of harmonics with comparable intensities forming a plateau ( 9 to 19 in h and 15 to 21 in he ) .
because of this , we have selected that range ( shaded in the plot ) to perform the selective optimisations .
the range is displayed , this time with a linear @xmath69 axis scale , in the inset .
for the case of he we show the exx and sae results . as in he
two electrons populate the only orbital in a spin - compensated configuration , the sae approximation , in this case , consists in neglecting the interaction between the electrons during the action of the field , freezing the potential to its ground state shape . in this adiabatic - dft context
, it amounts to ignoring the time - evolution of the hartree , exchange and correlation potentials , and is useful to gauge the relevance that correlations may have on the hhg optimisation .
hhg spectrum of the hydrogen ( top ) and he ( bottom ) atoms , with the reference pulse of eqn .
( [ reference - pulse ] . for the case of the he hhg
spectra we show two results : one solving the tddft equations using the exx functional ( green ) and the other solving the single - active - electron ( sae ) ( red ) equation , commonly used by the strong - field community . to make more clear the comparison between exx and sae we shifted the sae spectra by 0.5@xmath11 in the x - axis .
the shaded area contains the harmonics of interest .
this area is also displayed in the inset , with a linear @xmath69-axis scale . ]
hhg spectrum of the hydrogen ( top ) and he ( bottom ) atoms , with the reference pulse of eqn .
( [ reference - pulse ] . for the case of the he hhg
spectra we show two results : one solving the tddft equations using the exx functional ( green ) and the other solving the single - active - electron ( sae ) ( red ) equation , commonly used by the strong - field community . to make more clear the comparison between exx and sae we shifted the sae spectra by 0.5@xmath11 in the x - axis .
the shaded area contains the harmonics of interest .
this area is also displayed in the inset , with a linear @xmath69-axis scale . ]
let s discuss first the case of optimising the hhg spectra of h. we used the target given by eq .
( [ eq : hhgtarget2 ] ) to optimise the odd orders from the 9th to 19th .
to enhance the 9th harmonic , for example , we set @xmath70 , and @xmath71 ( all other @xmath31 are zero ) . in this manner ,
the sum of all coefficients is zero , avoiding any improvement of the merit function due to a mere overall reduction or increase of the spectrum .
the results are displayed in fig .
[ fig : hplot ] . from top to botton , in the left panels , the spectra produced by the optimal fields for the 19th , 17th ,
, 9th harmonic . in the right panels , the optimal fields themselves ; their envelopes in real time , as well as their power spectrum .
the resulting fields produce considerably higher harmonic outputs than the unshaped , reference field . to quantify this point we introduced an _ enhancement factor _ that is displayed in each plot , defined as : @xmath72 } \lbrace h[\varphi](\omega)\rbrace } { h_{\rm ref}(j\omega_0 ) } \,,\ ] ] where @xmath73 is the spectrum obtained with the reference field , and @xmath34 the one obtained with the optimal field ( the computation of the max function is not needed for the former , because due to the regularity of its envelope function , @xmath73 always peaks at the precise integer multiples @xmath74 ) .
this enhancement factor greatly vary from case to case ( i.e. it is 6 for @xmath75 , and 208 for @xmath76 ) .
note that the plots do not share the same @xmath69-scale ; they are scaled in each case to the value of the maximum of the plot .
optimized hhg spectra ( left panels ) , and corresponding optimal fields ( right panels ) , for the hydrogen atom case .
the optimal fields are plotted in the time domain ( only the envelope function @xmath61 is shown ) , and in the frequency domain .
the hhg spectra are shown in a linear scale , normalized in each case up the value of the maximun value .
the enhancement factor defined in eq .
( [ eq : enhancement - factor ] ) is also shown . ]
we turn now our attention to the case of the helium atom , that contains two electrons .
the interaction between these is treated here with tddft , within the exx approximation . as in the previous case we performed optimisations based on the target given by eq .
( [ eq : hhgtarget2 ] ) , now for the orders 15th to 21st , fixing the coefficients @xmath31 in an analogous manner .
the results are displayed in fig .
[ fig : heplot ] . from top to bottom , in the left panels , the spectra produced by the optimal fields for the 21st , 19th , 17th , and 15th harmonic . in the right panels , the optimal fields themselves . the enhancement factors achieved are quite large , and as in the case of hydrogen , rather different from case to case .
this rather large enhancement of the wanted harmonic is not accompanied by a full depletion of the neighbouring ones in fact , they are also increased .
this partial selectivity is also similar to the hydrogen results . to quantify the role of electron - electron interactions we show in the same fig .
[ fig : heplot ] the sae results ( red curve ) .
qualitatively , the sae results are not very different to the ones obtained with the exx functional , in terms of intensity enhancements and selectivity .
the fact that the calculated optimal fields and the spectra are different for both exx and sae illustrate not only the intrinsic non - linearity of the optimisation algorithms and the rather large number of possible local maxima , but also the fact that electron interaction does play a role in the generation and optimisation of harmonics . indeed , by looking in more detail to the results shown in fig .
[ fig : heplot ] for the 15th and 19th harmonic optimisation , we see that exx with respect to sae provides a better selectivity and harmonic enhancement , measured by the height of the desired harmonic and the quenched of the neighbouring ones .
therefore electron correlation seems to play a role in the optimisation of harmonics .
this fact , together with the common knowledge that heavier noble gases emit stronger hhg radiation that light ones ( whereas the sae that predicts similar spectra ) @xcite support our findings about the limitations of the sae approximation and the role of electron interactions .
in fact we can expect larger enhancement factors to be reached by applying the present optimisation techniques to heavier atomic / molecular systems .
optimized hhg spectra ( left panels ) , and corresponding optimal fields ( right panels ) , for the helium atom case . as in fig .
( [ fig : h - he ] ) we show in green the results within tddft using the exx functional and in red the ones using the sae approximation .
the optimal fields are plotted in the time domain ( only the envelope function @xmath61 is shown ) , and in the frequency domain .
the hhg spectra are shown in a linear scale , normalized in each case up the value of the maximun value .
the enhancement factor defined in eq .
( [ eq : enhancement - factor ] ) is also shown . ]
in conclusion , we have investigated , by theoretical means , the possibility of tuning the shape of the hhg spectrum of the hydrogen and helium atoms by shaping the slowly varying envelope of a 800 nm , 200-cycles long laser pulses . for this purpose
, we have optimised a functional designed to enhance selected harmonics .
the allowed modifications of the pulse are very constrained , since we enforce a maximum envelope frequency no larger than 1/60 of the fundamental frequency .
this means very slowly varying envelopes .
however , the picture that emerges of our analysis is that these relatively small modifications produce strong variations of the spectra , allowing for significative increases of the harmonic intensities .
these enhancements are not fully selective , since the neighbouring harmonics also increase , but to a lesser extent .
the outcome depends of the precise definition of the target functional , which is a topic to be investigated further .
there is ample freedom to choose this object , and a different option may yield better selectivity while perhaps reducing the total enhancement , or vice - versa .
the spectra have been computed with a fully quantum mechanical description , by explicitly computing the time - dependent dipole moment of the systems .
the results presented here correspond to the single - atom response we have not propagated maxwell s equations in a atomic gaseous medium .
therefore , this work demonstrates the relevance of the single atom response for hhg and how this single - atom response is significantly altered by the envelope of the laser pulse , even for the small modifications allowed in our scheme .
we have shown that few orders of magnitude hhg enhancement factor can be reached at the single - atom level .
thus , if this fact is combined with the phase matching method used for hhg generation we would be able to reach much higher global harmonic enhancement in atomic and molecular gases .
moreover , our results illustrate the role of electron - electron interactions in this optimisation and control of hhg .
this can be qualitatively rationalised in terms of the larger hilbert space spanned by the interacting system as compare to the simplest single - active electron scheme ( or any other non - interacting electron approach ) .
the propagation of the tddft equations are those of the single - particle orbitals forming the slater determinant , a set of equations usually called `` time - dependent kohn - sham '' ( tdks ) equations : @xmath77(\vec{r},t)\varphi_i(\vec{r},t)\ , , \\
\label{eq : initial - value } \varphi_i(\vec{r},0 ) & = & \varphi_i^{\rm gs}(\vec{r})\ , .
% \label{eq : tddensity } % n(\vec{r},t ) & = & 2\sum_{i=1}^{n/2}\vert\varphi_i(\vec{r},t)\vert^2\,,\end{aligned}\ ] ] the initial values specifed by eqs .
( [ eq : initial - value ] ) are given by the ground - state kohn - sham orbitals , computed with static dft .
the time - dependent density of the system may be retrieved from the ks orbitals with the simple formula : @xmath78 where @xmath79 is the occupation of each orbital , which is equal to two if we consider a spin - compensated system of @xmath4 electrons , doubly occupying a set of @xmath80 spatial orbitals @xmath81 ( @xmath82 ) .
the potential that appears in those equations , @xmath83 ( the `` kohn - sham potential '' ) is a functional of this density , and is defined as : @xmath84(\vec{r},t ) = v(\vec{r } ) + \varepsilon(t)\vec{\pi}\cdot\vec{r } + v_{\rm h}[n](\vec{r},t ) + v_{\rm xc}[n](\vec{r},t)\,,\ ] ] where the hartree potential @xmath85 is given by : @xmath86(\vec{r},t ) = \int\!\!{\rm d}^3r'\;\frac{n(\vec{r}',t)}{\vert \vec{r}'-\vec{r}\vert}\,,\ ] ] and @xmath87 is the static external potential .
the time - dependent external potential for these one - electron equations is given by @xmath88 , in terms of objects already defined .
we have studied the two simplest atoms , hydrogen and helium . for the helium atom ,
we have used tddft with the exact - exchange functional . in order to assess the possible relevance of the electron - electron interaction , we have repeated the helium atom calculations employing the single active electron ( sae ) approximation , which in this case amounts to freezing the hartree , exchange and correlation functional to its ground - state value during the propagations . in this manner
, we are effectively ignoring the electron - electron interaction during the propagation , and may gauge the relevance that it may have on the possibility of changing hhg spectra via smooth variations of the envelope function .
for the purpose of studying the hhg of atoms in linearly polarized pulses , one - dimensional ( 1d ) models have been routinely employed in the past , and we have adhered to this practice , since it provides a good qualitative picture , while substantially reduces the computational cost . the nucleus - electron interaction has the soft - coulomb form : @xmath89 for an electron placed at @xmath90 and a nucleus of charge @xmath91 placed at @xmath92 . the constant @xmath93 may be tuned to reproduce some atomic property ( e.g. ionization potential ) , although in this case we have simply fixed it to one for both hydrogen and helium .
everything has been implemented in the octopus code @xcite .
the wavefunctions , potential , densities , etc . are represented in this code by the values they take at points of a real space grid .
the laplacian operator , needed to compute the kinetic part of the hamiltonian , is computed using a 9-point finite difference formula .
the propagations are performed by dividing the full time interval into short time steps @xmath94 $ ] , and approximating the short - time evolution operator @xmath95 with the exponential mid - point rule : @xmath96 the action of the exponential on a state vector is computed by making use of the lanczos polynomial expansion ( see ref .
@xcite for a discussion of the propagation schemes used in octopus ) .
the full details about the combination of tddft and qoct were explained in refs .
@xcite and @xcite .
this work was supported by the european commission within the fp7 cronos project ( i d 280879 ) .
ar acknowledges financial support from the european research council advanced grant dynamo ( erc-2010-adg-267374 ) , spanish grant ( fis2010 - 21282-c02 - 01 ) , and grupos consolidados upv / ehu del gobierno vasco ( it578 - 13 ) . | high harmonic generation ( hhg ) provides a flexible framework for the development of coherent light sources in the extreme - ultraviolet and soft x - ray regimes .
however it suffers from low conversion efficiencies as the control of the hhg spectral and temporal characteristics requires manipulating electron trajectories on attosecond time scale .
the phase matching mechanism has been employed to selectively enhance specific quantum paths leading to hhg .
a few important fundamental questions remain open , among those how much of the enhancement can be achieved by the single - emitter and what is the role of correlations ( or the electronic structure ) in the selectivity and control of hhg generation . here
we address those questions by examining computationally the possibility of optimizing the hhg spectrum of isolated hydrogen and helium atoms by shaping the slowly varying envelope of a 800 nm , 200-cycles long laser pulse .
the spectra are computed with a fully quantum mechanical description , by explicitly computing the time - dependent dipole moment of the systems using a first - principles time - dependent density - functional approach ( exact for the case of h ) .
the sought optimization corresponds to the selective enhancement of single harmonics , which we find to be significant .
this selectivity is entirely due to the single atom response , and not due to any propagation or phase - matching effect .
in fact , this single - emitter enhancement adds to the phase - matching techniques to achieving even larger hhg enhancement factors .
moreover , we see that the electronic correlation plays a role in the determining the degree of optimization that can be obtained . |
observations indicate that a large fraction of stars reside in triple systems ( * ? ? ?
* and references therein ) .
in a dynamically stable non - coplanar triple the gravitational interaction between the inner and outer binary drives periodic / quasi - periodic variations of the mutual inclination between the two orbits and of the eccentricity of the inner binary @xcite .
such eccentricity oscillations , first discovered by @xcite , can lead to a close approach between the inner binary components which can produce a variety of exotic astrophysical phenomena , including compact - object and stellar binary mergers ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , x - ray binaries , type ia supernovae and gamma - ray bursts @xcite .
lidov - kozai ( lk ) oscillations are usually described as a secular phenomenon , occurring on timescales much longer than the orbital period of the outer binary orbit ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the standard approach is to make use of the so - called orbit - averaging technique in which one computes the mutual gravitational interaction ( torques ) between two mass weighted ellipses instead of point masses on orbits .
the basic assumption is that the evolution of the orbital parameters occur on a timescale much longer than the dynamical timescales of the system ( e.g. , * ? ? ?
the orbit averaged equations of motion can describe the evolution of a wide range of systems , but become inaccurate once the triple hierarchy is moderate .
@xcite showed that the orbit - average treatment , even at the octupole level of approximation , becomes inaccurate in hierarchical systems where the tertiary comes closer to the inner binary than a certain distance . at the high eccentricity phase of a lk cycle the timescale for the angular momentum perturbation due to the tertiary can become comparable to or shorter than the dynamical timescales of the system rendering the secular approach no longer valid @xcite . in this case , the binary can be driven to a much closer separation than the secular theory would otherwise predict .
this can have important consequences for the evolution of the inner binary .
for example , @xcite , @xcite and @xcite demonstrated that in this non - secular regime , a compact object binary can be pushed inside the @xmath3hz gravitational wave ( gw ) frequency band of advanced ligo ( aligo ) detectors with a finite eccentricity .
similarly , in the case of stellar binaries , the two stars could be driven to a direct collision . @xcite and @xcite suggested that direct stellar collisions driven by the lk mechanism might have an important role in the formation of blue straggler stars ( bsss ) in stellar clusters .
globular clusters ( gcs ) are a natural environment for forming triples and higher order multiples . on one hand ,
the dense stellar environment of gcs is hostile to triples .
in fact , given the high densities of gcs , triples may be involved in close encounters which tend to destroy them . on the other hand , the high stellar densities that characterize the core of gcs favor dynamical binary - binary interactions that naturally lead to the formation of stable triples @xcite , including black hole ( bh ) triples and mixed , compact object - star , triples that could not otherwise form .
the inner binaries in the triples may be driven to merge before their next interaction .
interestingly , such mergers have been suggested as a channel for the formation of compact - object mergers @xcite , and more recently as a channel for the formation of bsss through mass transfer and merger @xcite . in this paper
we study the evolution of dynamically assembled triples inside gcs with particular focus on triples containing an inner main - sequence ( ms ) stellar binary and triples with an inner bh binary .
we do not consider here the evolution of mixed star - compact object systems , or binaries containing evolved ( e.g. , red giant ) stars ; we reserve a study of these latter systems to a future investigation . in this paper
we use a monte carlo code to simulate the long - term evolution of gcs @xcite , and identify all triple systems formed during the cluster evolution .
we then evolved these triples using a high - accuracy direct three - body integrator which includes tidal and general relativity ( gr ) effects for both the inner and outer orbit .
scattering with other stars is also accounted for , but only in the sense that the triples are evolved for a time equal to the timescale over which they will be disrupted through gravitational encounters with other stars .
the results of the direct integrations are used to determine the rate of bh and stellar binary mergers induced by the lk mechanism in gcs .
although previous studies have shown that triples can form quite frequently in gcs , their detailed long - term dynamical evolution in the cluster environment has not been yet investigated .
our study sheds light on exactly this point , i.e. it allows us to assess for the first time the role of the host cluster evolution and its properties on the formation and dynamics of the triples .
this allows us to place reliable constraints on the contribution of mergers from triples to bss populations and on the event rate and properties of inspiraling bh binaries as a source of gw radiation for aligo detectors ( e.g. * ? ? ?
* ; * ? ? ?
we note that the dynamics of triples with either an inner stellar or compact object component has been investigated by many authors and in a variety of contexts ( including stellar triples , planets around stars , supermassive bh mergers ) , using the standard orbit averaged formalism and therefore fully neglecting non - secular effects ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the new ingredients of our study compared to previous work dealing with similar problems are : ( i ) the use of realistic initial conditions for our triple populations , that were directly obtained using monte carlo models of the long - term dynamical evolution of gcs ; ( ii ) the use of accurate three - body integrations that allows an essentially exact treatment of the triple dynamics , avoiding the approximations that are made when using the secular equations of motion .
the paper is organized as follows .
we begin with a discussion of dynamically assembled triples in gcs and their properties in section [ bd - oaa ] , and show that for most triples the orbit average approximation is expected to break down . in section [ meth+ics ]
we discuss our gc models , triple initial conditions and adopted numerical methods .
we present our results on bh - triples and associated aligo event rates in section [ res - bhs ] . in section
[ st - tr ] we describe the results concerning stellar triples and their implications for bss formation in gcs .
we further discuss the implications of our results in section [ disc ] .
section [ summ ] sums up .
we consider a binary of components of masses @xmath4 and @xmath5 orbiting a third body of mass @xmath6 . we indicate the eccentricities of the inner and outer orbits , respectively , as @xmath7 and @xmath8 , and semi - major axes @xmath9 and @xmath10 ; we define @xmath11 as the argument of periapsis of the inner binary relative to the line of the descending node , @xmath12 as the argument of periapsis of the outer orbit , and @xmath13 as the orbital inclination of the inner orbit relative to the outer orbit . in what follows
we only consider triple systems that satisfy the stability criterion of @xcite , @xmath14^{2/5}\\ & & \times(1 - 0.3i/\pi)~ , \nonumber\end{aligned}\ ] ] with @xmath15 .
if the system satisfies the stability criterion ( [ stab ] ) it is also reasonable to describe the dynamics of the entire system as the interaction between an inner binary of point masses @xmath4 and @xmath5 and an external binary of masses @xmath16 and @xmath17 .
we define the angular momenta @xmath18 and @xmath19 of the inner and outer binary : @xmath20^{1/2}~,\ ] ] and @xmath21^{1/2}~,\ ] ] with @xmath22 the gravitational constant .
the dimensionless angular momenta are @xmath23 and @xmath24 with @xmath25 the angular momentum of a circular orbit with the same semi - major axis , @xmath26 .
if the changes in the orbital properties of the inner binary caused by its gravitational interaction with the third body occur on a timescale longer than both the inner binary orbital period , @xmath27 , and the tertiary orbital period , @xmath28 , it is convenient to average the equations of motion over the rapidly varying mean anomalies of the inner and outer orbits ( e.g. , * ? ? ?
the resulting double averaged hamiltonian is @xmath29 with @xmath30 $ ] and @xcite @xmath31 the quadrupole - level secular perturbation equations can be easily derived from the conserved hamiltonian ( [ ham ] ) .
the resulting evolution equation of the inner binary orbital angular momentum due to the torque from the tertiary is ( e.g. , * ? ? ?
* ) : @xmath32 the characteristic timescale for the eccentricity oscillations is @xcite : @xmath33 using equation ( [ eq : eccen ] ) and taking the relevant limit @xmath34 , we find that the maximal change in the inner binary angular momentum due to the interaction with the third object and over the inner binary orbital period is @xmath35 and the change in the binary angular momentum over one orbit of the tertiary is @xmath36 thus , the change experienced over the timescale associated with the period of the outer orbit is naturally @xmath37 larger than the change experienced over the inner binary orbital period .
equation ( [ ham ] ) describes the gravitational interaction of two weighted ellipses rather than point masses on orbits and is accurate as long as @xmath38 , i.e. , changes of the inner binary angular momentum occur on a timescale longer than the tertiary orbital period .
if instead the change in the inner binary angular momentum over the inner and outer orbital periods is of order @xmath39 , then the system no longer meets the conditions required for the orbit averaged approximation , and more accurate direct three - body integrations are needed . following the discussion above , in order to address the reliability of the orbit averaged approach in describing the evolution of a triple system it is useful to compare the change in angular momentum of the binary to the critical angular momentum @xmath40 , where @xmath41 is the scale at which other dynamical processes will dominate the evolution of the inner binary . since in this paper
we focus on triples in which the inner binary can be driven to a strong close interaction or even a merger , we identify these processes with energy loss due to gw radiation and tidal dissipation in the case of bh and stellar binaries , respectively . , dashed lines give the inner binary semi - major axis . in the upper panel the initial conditions for the bh triple system were : @xmath42 , @xmath43 , @xmath44 @xmath45 , @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath50 . in the lower panel the initial conditions for the stellar triple were : @xmath51 , @xmath52 , @xmath53 @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 . in the upper panel
the asterisks denote the minimum separation attained by the two bhs every orbital period . at @xmath61
this system satisfies the condition @xmath62 , so that the jump in periapsis over one inner orbit is of order the periapsis distance itself . in the lower panel at @xmath63
the triple satisfies the condition @xmath64 , so that the averaging procedure over the outer orbit is no longer justified . in both examples
the inner binary reaches higher eccentricities in the three - body simulation than in the secular integration .
this allows the two bhs in the upper panel to merge through gw energy loss , and the two stars in the lower panel to approach each other close enough that mass transfer would commence .
the dot - dashed lines indicate the critical value of @xmath65 below which gw radiation will dominate ( upper panel ) , and below which the two stars will collide ( lower panel ) .
these two examples clearly illustrate the limitation of the standard secular approach even when including octupole order terms to the equations of motion.,title="fig:",width=206 ] + , dashed lines give the inner binary semi - major axis . in the upper panel the initial conditions for the bh triple system were : @xmath42 , @xmath43 , @xmath44 @xmath45 , @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath50 . in the lower panel the initial conditions for the stellar triple were : @xmath51 , @xmath52 , @xmath53 @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 . in the upper panel
the asterisks denote the minimum separation attained by the two bhs every orbital period . at @xmath61
this system satisfies the condition @xmath62 , so that the jump in periapsis over one inner orbit is of order the periapsis distance itself . in the lower panel at @xmath63
the triple satisfies the condition @xmath64 , so that the averaging procedure over the outer orbit is no longer justified . in both examples the inner binary reaches higher eccentricities in the three - body simulation than in the secular integration .
this allows the two bhs in the upper panel to merge through gw energy loss , and the two stars in the lower panel to approach each other close enough that mass transfer would commence .
the dot - dashed lines indicate the critical value of @xmath65 below which gw radiation will dominate ( upper panel ) , and below which the two stars will collide ( lower panel ) .
these two examples clearly illustrate the limitation of the standard secular approach even when including octupole order terms to the equations of motion.,title="fig:",width=206 ] the condition that the inner binary undergoes rapid eccentricity oscillations over the _
orbital period can be written as @xmath66 ; this requirement translates into a condition on the outer perturber separation , @xmath67 we call this regime `` moderately '' non - secular . in this regime systems
can experience eccentricity oscillations over the outer binary period and can be driven to higher eccentricities than secular theory would otherwise predict , which in the case of compact object binaries could result in substantially reduced merger times @xcite .
if @xmath68 , the inner binary will undergo eccentricity oscillations on the timescale of the _ inner _ orbital period ; this can be achieved if the tertiary comes closer to the inner binary than the separation @xmath69 we call the regime defined by this equation `` strongly '' non - secular ( see also @xcite and equation 7 in @xcite ) . in this situation
the outer perturber is close enough to significantly change the angular momentum of the eccentric binary at the last apoapsis passage leading to a jump in angular momentum of order @xmath39 .
the angular momentum of an eccentric inner binary can jump from several times @xmath70 ( far enough to avoid gr or tidal effects ) to an arbitrarily small value .
equation ( [ bd2 ] ) defines the region of parameter space within which a compact object binary might be expected to enter the aligo frequency band with a finite eccentricity @xcite or , in the case of inner stellar components , the binary members might experience a close interaction leading to a stellar collision @xcite .
here we quantify the role of non - secular effects for triple systems forming in stellar clusters by analyzing the properties of all dynamically stable triples that were produced in the gc monte carlo models of @xcite ( see section [ mc - models ] ) .
the monte carlo models give the masses , stellar radii , semi - major axes , eccentricities and stellar type of all triples formed during the cluster dynamical evolution ; these properties were used to make figure [ period - pdf ] which shows the period distributions as well as the distribution of the ratio @xmath71 of all bh triples ( left panels ) and stellar triples ( right panels ) that formed in these models . in figure [ period - pdf ]
we have distinguished triples which satisfied the criterion ( [ bd1 ] ) ( green histograms ) and those meeting the more stringent condition equation ( [ bd2 ] ) ( blue histograms ) from the remaining systems which satisfied neither conditions . in the case of stellar triples we evaluated equation ( [ bd1 ] ) and ( [ bd2 ] ) setting @xmath72 with @xmath73 and
@xmath74 the physical radii of the inner binary components @xcite , while for bh triples we adopted a conservative dissipation scale of @xmath75 @xcite .
figure [ period - pdf ] shows that for the majority of bh triples formed non - secular dynamical effects are expected to become important before gr terms can significantly affect the evolution of the inner binary .
we conclude that the standard lk secular theory will fail in describing the evolution of these systems due to the breakdown of the double averaging approximation .
only a few stellar triples in our sample satisfy the condition ( [ bd2 ] ) , mainly because of their large value of @xmath41 .
however a large fraction of stellar triples still met the condition equation ( [ bd1 ] ) , meaning that the averaging procedure over the outer orbit is not justified for many of these systems . in what follows we study the dynamical evolution of triples that form through dynamical interactions in the dense stellar environment of gcs .
we use a high accuracy three - body integrator to study how the inner binary in the triples might evolve to become a possible source of gw radiation for aligo detectors , or , in the case of stellar binaries , how the evolution might lead to a mass transfer event or even a stellar merger .
we begin in the next section by describing the initial conditions and numerical methods adopted in our study .
the direct three - body integrations presented below were performed using the archain integrator @xcite .
archain employs an algorithmically regularized chain structure and the time - transformed leapfrog scheme to accurately integrate the motion of arbitrarily tight binaries with arbitrarily large mass - ratio .
the code also includes post - newtonian ( pn ) non - dissipative 1pn , 2pn and dissipative 2.5pn corrections to all pair - forces .
we refer the reader to @xcite for a more detailed description of archain .
we also used the octupole level secular equations of motion given in @xcite to evolve the systems , which allowed us to directly compare the predictions of the direct integrations to the orbit averaged results . in the case of triples containing an inner stellar binary ( stellar triples ) , to the pn terms we also added terms that account for dissipative tides as well as apsidal precession induced by tidal bulges . in order to do so we modified archain including terms to the equations of motion which represent tidal friction and quadrupolar distortion .
velocity - dependent forces were included via the generalized mid - point method described in @xcite ; this method allows to time - symmetrize the algorithmic regularization leapfrog even when the forces depend on velocities , permitting the efficient use of extrapolation methods .
the tidal perturbation force has the form @xmath76 where @xmath77 ( here set to a fixed value of @xmath78 sec ) is the time lag , and @xmath79 ( set to @xmath80 ) is the apsidal motion constant .
equation ( [ tides - f ] ) differs from the classical treatment of near - equilibrium tides to the extent that we have suppressed terms that are due to the difference in rotation rate between the binary and each component .
the orbital elements evolution equations that correspond to equation ( [ tides - f ] ) are given by equations ( 9 ) and ( 10 ) of ; these standard equations were implemented in the secular integrations to account for dissipative as well as non - dissipative tides . to illustrate the importance of non - secular effects for the dynamics of the systems under consideration , we show two examples in figure [ examples ] .
we consider a bh triple that is in the strongly non - secular regime defined by equation ( [ bd2 ] ) , as well as a stellar triple that is in the moderately non - secular regime defined by equation ( [ bd1 ] ) .
the evolution of the two systems was computed using both the secular equations of motion and archain .
the examples of figure [ examples ] demonstrate how the double - orbit - averaging procedure can lead to misleading results when applied to the dynamics of systems that evolve to attain very high orbital eccentricities . in the upper panel
the three - body integration predicts that the inner binary will merge within a few lk oscillations whereas the secular calculation predicts that the system does not merge in the considered time interval .
more importantly , in the direct three - body integration the binary was found to have a substantial eccentricity of @xmath81 at the moment its peak gravitational wave frequency ( equation [ [ gw_f ] ] below ) reached @xmath3hz . in the lower panel
the stellar binary also attains higher eccentricities when integrated with archain , which might induce a mass transfer event or reduce the circularization time due to tidal friction when compared to the secular integrations .
the initial conditions for the triple systems used in this analysis were taken from the monte carlo simulations of @xcite .
this study presents 42 monte carlo dynamical simulations of massive star clusters containing large populations of stellar bhs .
the monte carlo technique simulates two - body relaxation in the fokker - planck approximation through representative pair - wise scattering interactions . our implementation in the cluster monte carlo ( cmc ) code ( * ? ? ? * and references therein ) includes direct integration of strong three- and four - body ( binary single and binary binary ) encounters , direct physical collisions , single and interacting binary stellar evolution , and tidal interactions with the galaxy .
primordial triples are not included in these models .
however , during strong binary binary interactions , it is possible to form stable hierarchical triple systems @xcite .
limitations in cmc currently require that these triples be broken artificially at the end of the timestep .
nonetheless , whenever a stable triple is formed , its properties are logged , including the masses , radii , and star types for the stellar components , and the semi - major axes and eccentricities for the inner and outer orbits .
since we lack the information regarding the mutual orientation of the two orbits , we sample @xmath82 and @xmath83 randomly from a uniform distribution with @xmath84 . the initial orbital phases in the three - body integrations were also randomly distributed .
we extracted the properties for every stable triple system formed through dynamical interactions over 12 gyr across all of the 42 models presented in morscher et al .
2015 , and from five additional models provided by the same authors ( morscher , private communication ) .
the main properties of the 47 cluster models are reported in @xcite ; these span a large range of masses ( from @xmath85 to @xmath86 ) , of virial radii ( from @xmath87pc to @xmath88pc ) , and metallicities ( z=0.0005 , 0.0001 and 0.005 ) .
we note that , because these triple systems are destroyed in the monte carlo simulation , it is possible for the components of these triples to subsequently form new triple systems , when in reality they could in principle survive for a significant period of time . in the high - density stellar environment of gcs , triples may be perturbed through encounters with other passing stars on timescales that can be shorter than the relevant lk timescale .
such encounters will alter the orbital properties of the triple significantly , or even disrupt it . to account for this we set the final integration time in our simulations equal to the encounter timescale for collisions with other stars @xcite , @xmath89^{-1}\end{aligned}\ ] ] where @xmath90 is the central velocity dispersion of the cluster in units of @xmath91 , @xmath92 is the mass of the triple , @xmath93 is the mass of an average star in the cluster , and @xmath94 is the local number density of stars in units of @xmath95 .
the values of @xmath96 , @xmath93 and @xmath97 in equation ( [ coll ] ) were obtained directly from the monte carlo models and correspond to to their values in the cluster core at the moment the triple system was formed .
we then integrate each triple until a time @xmath98 was reached , or until the inner binary had merged . in total ,
the monte carlo models contain 8864 triples . of these
5238 are bh triples ; 3626 have at least one non - bh component in the inner binary , and of these , in the inner binary , 2940 have two ms stars , 260 have a bh plus another ms star , 218 have a white dwarf plus another ms star , 4 have a neutron star ( ns ) plus a ms star , 163 have a ms star plus an evolved star ( red giant or later type ) , 2 have a bh plus an evolved star .
the remaining systems include 32 triples with an inner white dwarf binary , 5 triples with an inner bh plus white dwarf binary , and 2 triples with an inner bh plus ns binary .
no triple with an inner ns binary is seen in any of these monte carlo models . of the 5238 bh triples ,
3852 satisfy the condition @xmath99 , whereas 2519 of the 2940 stellar triples with an inner ms binary satisfied this condition . .
triangle symbols represent systems that satisfy the condition equation ( [ bd2 ] ) , open circles are systems which satisfy the condition equation ( [ bd1 ] ) but not the condition equation ( [ bd2 ] ) and plus symbols are systems which satisfy neither of these two conditions .
we evolved each of these systems ten times selecting randomly the initial inclination as detailed in the text .
blue colored points are systems in which in at least of the 10 realizations we found one merger within @xmath98 .
red points are systems in which for at least one of the ten realizations the inner binary eccentricity was larger than @xmath100 at the moment its peak gw frequency became larger than @xmath101hz .
as predicted , most eccentric mergers occur in triples for which the orbit averaging over the inner binary breaks down .
, width=249 ]
for each of the 3852 bh triples extracted from the monte carlo models we made ten realizations , each adopting a random orientation of the two orbits and random orbital phases .
of the total 38520 bh triples that we evolved with archain , 2080 resulted in a merger of the two inner bhs .
the total number of bh mergers obtained when using the orbit averaged equations of motion was 1796 instead .
we found that the merger time of such binaries after the triple was formed was typically very short and always shorter than @xmath102yr .
thus , the bh mergers discussed in what follows will occur inside the gcs even when the recoil velocity at formation can be large enough to eject the triple from the cluster core . in this section
we discuss the properties of the merging compact binaries including their eccentricity , mass , merger time distribution , and event rates .
eccentric binaries emit a gw signal with a broad spectrum of frequencies .
the peak gravitational wave frequency associated with the harmonic which leads to the maximal emission of gw radiation can be estimated as @xcite @xmath103^{1.5}}~.\ ] ] we approximately follow the time evolution of the gw frequency of the inner bh binaries through equation ( [ gw_f ] ) .
this allowed us to estimate the eccentricity of the bh binary when its gw frequency is @xmath104hz , i.e. , when it would be large enough to be into the aligo frequency band @xcite .
figure [ fig3 ] gives the ratio of the triple survival timescale , @xmath98 , to the lk timescale plotted against the value of @xmath71 .
from this plot we see that non - secular dynamical effects are expected to become important to the evolution of most bh triples in our models , possibly leading to the formation of eccentric gw sources .
also , most bh mergers are found to occur in moderately hierarchical triples with @xmath105 .
this is expected for at least two reasons : ( 1 ) the closer the outer body to the inner binary the less significant is the quenching of the lk cycles due to relativistic precession of the inner binary ( e.g. , * ? ? ?
* ) , so that typically the smaller the ratio @xmath71 the larger the maximum eccentricity attained by the inner binary and , consequently , the shorter its merger time ; ( 2 ) triples with small @xmath71 ratio , have also larger @xmath106 ratio , which naturally leads to a higher chance for a merger before the triple is disrupted by gravitational encounters with other stars . in figure [ fig3 ]
we also identify those systems for which at least one of the 10 random realizations led to a merger ( blue points ) and those for which at least one realization produced a bh binary merger with an eccentricity larger than 0.9 at @xmath107hz frequency ( red symbols ) .
as predicted on the basis of our discussion in section [ bd - oaa ] , most eccentric mergers occur at @xmath105 , near the boundary for instability .
figure [ fig4 ] shows the eccentricity distribution of all bh binary mergers in our simulations , when the associated @xmath108 _ first _ enters the @xmath109hz and @xmath110hz frequency bands .
figure [ fig5 ] and [ fig6 ] show the corresponding merger time and mass distribution of the merging binaries .
the distributions showed in figures [ fig4 ] , [ fig5 ] and [ fig6 ] were obtained by weighing the number of mergers for each cluster model by the likelihood of that model to represent a typical gc in the milky way ( mw ) .
more in detail , the weights are obtained by creating a kernel density estimate ( @xmath111 ) of the mw gcs on the fundamental plane ( which we take to be mass and ratio of the half to core radius ) , then estimating the weight for each model using the kernel density estimate at the position of that model on the fundamental plane . in this way ,
cluster models that are more likely to be drawn from the same distribution of mw gcs are more heavily weighted .
the weight for a model of total mass @xmath112 , core radius @xmath113 and half mass radius @xmath114 after @xmath115gyr of evolution is computed as : @xmath116 where we have divided by @xmath117 , the kernel density estimate of the models themselves .
this serves to normalize the distribution , so that regions of parameter space that are over - sampled by the models are given lower weights ( see * ? ? ?
* for more details ) .
the fraction of mergers with a given property ( e.g. , total mass , eccentricity ) is then simply @xmath118 , with @xmath119 the number of mergers occurring in the @xmath120 cluster model .
@xcite showed that for eccentricities less than @xmath121 at @xmath122hz , circular templates will be effective at recovering the gw signal of eccentric sources .
figure [ fig4 ] shows that approximately @xmath1 of all bh mergers in our three - body integrations had an eccentricity @xmath123 at @xmath104hz frequency .
this percentage drops to @xmath124 at @xmath110hz frequency .
the difference with the results of the secular equations of motion is evident in these plots .
the secular integrations clearly underestimate the number of eccentric mergers producing just a few percent of inspirals with @xmath125 at @xmath109hz frequency .
the direct three - body integrations also produce a significant population ( @xmath126 of the total ) of highly eccentric sources in the aligo frequency window , which are fully missed when evolving the triples with the secular equations of motion .
these sources will start to inspiral due to gw radiation energy loss within the aligo band with an extremely high eccentricity , @xmath127 , and a characteristic gw frequency .
in fact , the angular momentum of bh binaries that evolve into the strongly non secular regime can be arbitrarily small at the moment energy loss due to gw radiation starts to become dominant .
for this reason the gw frequency at which the binary enters the aligo band , which depends on the binary angular momentum via equation ( [ gw_f ] ) , could be significantly larger than @xmath128 .
( see the example of figure 3 in * ? ? ?
we note that the detectability of such sources also depends on how much energy is radiated around @xmath129 where aligo is more sensitive .
so it is important to see at what frequency these sources first enter the aligo sensitivity window .
based on our simulations we found that all eccentric bh binaries will enter the aligo band at @xmath130 and @xmath131 of them had @xmath132 when they first enter the aligo frequency band ( see also figure 7 in * ? ? ?
we conclude that most eccentric sources could be detected by aligo , provided that an efficient search strategy is used .
figure [ fig5 ] displays the average merger time distribution of bh binaries induced by the lk mechanism in our models .
the figure shows a clear peak around 1gyr of evolution with a long tail extending to the present epoch .
the declining merger rate is mostly due to the decreasing total number of bhs in our models with time .
as the rate depends on the total number of bhs in the cluster core it drops with time as bhs are continuously ejected from the cluster through strong dynamical interactions .
figure [ fig6 ] displays the chirp mass , @xmath133 , as well as the total mass distribution of the merging bh binaries .
these distributions appear to be consistent with that of dynamically formed binaries shown in figure 2 of @xcite .
figure [ fig6 ] shows a mass distribution peaked around @xmath134 , while the highest chirp mass of bh binaries formed by pure stellar evolution over all our models was @xmath135 .
this latter value for the chirp mass is the maximum for primordial binaries in our gc simulations that evolved without significant dynamical encounters ( see also * ? ? ?
* ; * ? ? ?
the fact that the typical chirp mass of the merging bh binaries in figure [ fig6 ] is substantially larger than @xmath135 is a result of the naturally stronger segregation of the heaviest bhs as well as the preferential ejection of the lightest bhs from the clusters during dynamical interactions .
as the heavier bhs segregate more efficiently to the cluster center and the lighter bhs are ejected through binary single interactions , the former will tend to dominate the cluster core , where binary - binary interactions can lead to the formation of stable bh triples and to lk induced mergers .
when @xmath136hz .
insert panel gives the corresponding cumulative distributions .
, width=220 ] one of the advantages of using realistic cluster monte carlo models is that it allows us to make direct predictions about the mass and merger time distribution of the coalescing bh binaries , which is fundamental to make reliable estimates for the event and detection rates .
as we have complete information about the distribution of sources in time and chirp mass , we can compare our ensemble of gc models to observations of mw and extragalactic gcs and estimate the merger rate for a single aligo detector . to compute the rate of detectable sources per year from triple systems ,
we follow a similar procedure to @xcite .
the rate is expressed as the following double integral over source chirp mass and redshift : @xmath137 where * @xmath138 is the rate of bh binary mergers with chirp mass @xmath139 at redshift @xmath140 due to the lk mechanism . * @xmath141 is the fraction of sources with chirp mass @xmath139 at redshift @xmath140 that are detectable by a single aligo detector .
* @xmath142 is the comoving volume at a given redshift @xcite .
we assume cosmological parameters of @xmath143 , @xmath144 , and @xmath145 , from the combined planck results @xcite .
* @xmath146 is the time dilation between a clock measuring the merger rate at the source versus a clock on earth .
the rate is expressed as the product of the mean number of triple sources per gc , the distribution of sources in @xmath147 space , and the spatial density of gcs per @xmath148 in the local universe ( 0.77@xmath149 , from @xcite , supplemental materials ) .
symbolically , this becomes @xmath150 .
[ cols="<,^,^,^,^,^,^ " , ] [ tab1 ] the components of @xmath138 are computed as follows .
the mean number of sources per gc is found by performing a weighted linear regression between the number of triple mergers a gc produces over its lifetime , and the final mass of the gc at 12 gyr .
the number of triple mergers for a given cluster is obtained by multiplying the total number of mergers obtained from the three - body integrations by a factor @xmath151 ; where the first term takes into account the fact that we have only considered inclinations @xmath84 and the second term the fact that for each triple we have performed ten integrations with random orbital orientations and phases .
the weights are assigned using @xmath152 , the numerator from equation ( [ wgts ] ) .
this linear regression is then multiplied by a universal gc luminosity function @xcite , and integrated over mass ( assuming a mass - to - light ratio of 2 ) to yield a mean number of sources .
this average is then multiplied by the distribution of sources , found by randomly drawing a number of mergers from each model according to the weights in equation ( [ wgts ] ) , then generating a kernel density estimate of these sources in chirp mass / redshift space . note that unlike @xcite , we do not separately consider the populations of low - metallicity and high - metallicity gcs . finally , we model the efficiency of an aligo detector by calculating the fraction of sources at chirp mass @xmath139 and redshift @xmath140 that aligo could detect , marginalized over all possible sky locations and binary orientations .
we use imrphenomc waveforms @xcite and the projected zero - detuning , high - power aligo noise curve . ]
note that this template family does _ not _ consider the eccentricity of sources , which could dramatically alter the gravitational waveforms @xcite .
such simplification does not significantly alter the total detection rate estimates presented here ; in fact , our simulations show that at @xmath153 frequency @xmath154 of coalescing bh binaries have an eccentricity @xmath121 , which is small enough that circular templates will be efficient at recovering their gw signal @xcite .
as @xmath155 is dependent on the random draw of mergers , we report the mean of @xmath156 over 100 separate draws .
we also provide optimistic and pessimistic rate estimates to approximate the uncertainties associated with our assumptions .
the optimistic rates are computed by assuming the @xmath157 value of the 100 draws of @xmath156 , the @xmath158 uncertainty for the linear regression between gc mass and number of mergers , a highly optimistic upper limit on the gc mass function ( @xmath159 ) , and the upper - limit on @xmath160 of @xmath161 from @xcite .
conversely , the pessimistic rate estimate assumes the corresponding @xmath162 uncertainties , a pessimistic upper limit on the gc mass function ( @xmath163 ) , and the lower - limit of @xmath160 of @xmath164 . for all lk driven events , we find that a single aligo detector could detect @xmath165 events per year ( pessimistically @xmath166 , optimistically @xmath167 ) .
we performed the same analysis described above for all eccentric sources produced in our three - body simulations , i.e. , we computed the detection rate solely for eccentric systems _ assuming perfect templates for detecting eccentric binaries_. thus , the results of this computation _ do not _ represent the number of sources that aligo could detect with current searches .
instead , they suggest an estimate of the number of sources that aligo could detect if optimal matched - filtering searches for eccentric sources were to be used . for events with eccentricities greater than 0.1 , this rate is @xmath168 ( pessimistically @xmath169 , optimistically @xmath170 ) , while for mergers with eccentricities greater than 0.9 , the rate becomes @xmath171 ( pessimistically @xmath172 , optimistically @xmath173 ) .
we note also that that the corresponding waveforms of most eccentric gw sources discussed here are characterized by a sequence of bursts of energy emitted near periapsis rather than a continuous waveform . because of this
, matched filtering techniques could be unpractical and a power stacking algorithm might be employed instead @xcite .
the disadvantage is that this will not give an optimal signal - to - noise ratio which might somewhat reduce the detection rates for eccentric inspirals we find .
table 1 summarizes the main results of our rate computation giving the predicted aligo detection rate of binary bh mergers from gcs .
this table also gives the corresponding compact binary coalescence rates per gc per myr .
table 1 gives the rate of lk induced bh mergers ( hierarchical triples ) , and the rate of mergers due to the lk mechanism that have an eccentricity larger than 0.1 at @xmath109hz frequency ( eccentric inspirals ) . from our monte carlo models we also computed the merger rate of hard escapers that will merge in the galactic field within one hubble time ( ejected binaries ) and the rate of mergers occurring _ inside _ the clusters that are not due to the lk mechanism ( in - cluster mergers ) .
these latter include mergers due to binary - single dynamical interactions , as well as those mergers due to binary stellar evolution and single - single captures ( see * ? ? ? * for details ) . in agreement with previous calculations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) we find that the total event rate is largely dominated by outside cluster mergers .
we stress again that the bh mergers due to the lk mechanism will occur inside the gc .
thus , by comparing the rate estimates for triples to the rate of in - cluster mergers due to other processes in table 1 we can conclude that the detection rate of bh mergers occurring _ inside _ the cluster is dominated by the rate of lk induced mergers .
this is somewhat surprising given that in - cluster mergers that do not occur in triples account for approximately @xmath174 of the total binary bh mergers over 12gyr .
the low aligo event rate we find is a consequence of the fact that the majority of these mergers occur early in the cluster s lifetime , and are relatively low - mass . since aligo is less sensitive to low - mass , high - redshift sources , the rate of detected sources drops .
we conclude that mergers induced by lk oscillations in hierarchical triples are likely to be the main contributor to in - cluster gravitational wave sources for aligo detectors , and represent @xmath175 of the anticipated total detection rate of bh binary mergers assembled in gcs .
but for stellar triples .
blue symbols are systems in which one component of the inner binary crossed the roche limit during the evolution . , width=249 ]
in total , our monte carlo models produced 2940 triples in which both components of the inner binary were ms stars ( i.e. , stellar triples ) .
following the same procedure adopted above , we made ten realizations for each triple , taking a random orientation between the two orbits and random orbital phases .
we then integrated the triples forward in time using archain if the condition ( [ bd1 ] ) ( with @xmath176 $ ] ) was satisfied ; the rest of the initial conditions were integrated using the octupole level secular equations of motion based on the double average hamiltonian ( e.g. , * ? ? ?
* ; * ? ? ?
each integration was terminated either at @xmath98 , or when one of the binary components had passed within its roche limit . in many of our simulations
the inner binary reaches very high eccentricities , implying a high probability that the stars will cross each other s roche limit and transfer mass . to account for the possibility of crossing the roche limit
, we check whether the condition @xmath177 ( or @xmath178 in orbit average integrations ) was met , with @xmath179 the instantaneous distance between the two inner binary components and @xcite : @xmath180 as often done in the literature , we stop the integration and assumed that the binary had merged if one of the two binary components had passed within its roche limit ( e.g. , * ? ? ?
we found that in 10327 out of the overall 29400 triples , one of the inner binary components crossed its roche limit .
after accounting for the fact that we have explored a limited range of orbital inclinations , we find an overall probability of @xmath124 that a stellar triple assembled dynamically in a gc will lead to a mass transfer event and possibly to a stellar merger .
figure [ fig7 ] gives the ratio of the triple survival timescale , to the lk timescale plotted as a function of @xmath71 .
mass - transfer systems are indicated by blue colored symbols .
this plot shows that the survival time - scale for these systems can be many orders of magnitude larger than the corresponding lk time - scale .
the ratio @xmath106 for these systems is also typically larger than that for bh triples ( compare to figure [ fig3 ] ) , reflecting the different tertiary period distribution of the two populations shown in figure [ period - pdf ] and the larger binary mass in the case of bh components . in figure [ fig7 ]
we identify those systems for which the orbit average approximation breaks down . in agreement with our previous discussion ,
most of these systems correspond to compact configurations in which the tertiary comes closer to the inner binary than a separation @xmath181 .
we also identify those systems in which in at least one of the ten random realizations one of the inner binary components crossed its roche limit . not surprisingly , close outer orbits are more likely to result in a mass transfer event between the inner binary components . in figure [ fig8 ]
we consider the distribution of orbital periods .
figures [ fig9 ] and figure [ fig10 ] display respectively the distribution of merger times and mass of the merger remnant for systems that crossed their roche limit .
these distributions were obtained using the weighing technique described in section [ res - bhs ] , i.e. , weighing the number of systems for each cluster model by the likelihood of that model to represent a mw - like gc . in figure [ fig8 ]
we show the initial and final orbital period distributions of the systems that did not cross the roche limit , as well as the initial period distributions of the systems that crossed the roche limit during the evolution . as shown in this figure the peak of the period distribution of the surviving binaries shifts from @xmath182days to @xmath183day .
this is a consequence of energy loss due to tidal dissipation that the binary experiences during the high eccentricity phases of lk cycles .
if during the evolution the timescale associated with precession due to tidal and pn terms becomes shorter than the lk timescale , the eccentricity oscillations due to the interaction with the tertiary are suppressed . in this case , after many lk cycles tidal dissipation can shrink the binary orbital semi - major axis by a large factor . causing the binary to effectively decouple from the tertiary .
figure [ fig8 ] shows that although merging systems are produced at any value of @xmath27 and @xmath28 , they are more likely to originate from systems with the tightest outer orbits and widest inner binaries .
in fact , about half of the systems with @xmath184days or with @xmath185days led to stellar mergers .
generally in order for the inner binary to attain a quasi - radial orbit such that its components can merge , its eccentricity has to be excited on a timescale much shorter than the typical extra precession timescale ( such as tides and relativistic precession ) .
it can be easily shown that the effectiveness of such extra sources of precession decreases with the ratio @xmath186 ( e.g. , * ? ? ?
* ) , which is consistent with the fact that mergers are more likely to occur for wide inner binaries and close outer orbits .
the distribution of merger times for stellar binaries is shown in figure [ fig9 ] .
given that the merger time through lk cycles is generally very short , the number of merging systems at a given time is also a good measure of the number of hierarchical triples that are formed dynamically in the cluster core at that time interval .
figure [ fig9 ] shows that @xmath187 of all stellar mergers occur in the first gyr of cluster evolution .
this is expected given that it is only during these early times that the stars dominate the cluster core where stellar binary - binary encounters can take place at a high rate and lead to the assembly of stellar triples .
after @xmath183gyr of evolution , the bhs have already decayed to the center through dynamical friction becoming the dominant population in the cluster core . as the bhs
become dominant , interactions involving stellar binaries become less frequent , naturally reducing the formation rate of triples with stellar components and favoring the formation of bh triple systems instead .
finally , figure [ fig10 ] displays the total mass distribution of all binaries that underwent roche - lobe overflow during their evolution .
this plot shows that the merging systems have masses up to @xmath188 with a peak at about @xmath189 .
the most massive mergers , including a few at @xmath190 , were found to occur only very early in the first few gyrs . as illustrated by the dashed - blue lines in figure [ fig10 ] , no merger with mass larger than @xmath191 was found during the last @xmath88gyr of cluster evolution .
these late mergers have a typical mass of @xmath192 , as expected being about twice the mass of an average ms star of an old single stellar population .
bsss are main sequence stars that appear to be hotter and more luminous than the turn - off point for their parent cluster population . as
such they appear to be younger than the rest of the stellar population ( e.g. , * ? ? ? * ) .
bsss are created when a ms star is replenished with new supply of hydrogen in its core .
hence , they can remain in the ms longer than expected of an undisturbed star .
this can happen through mass transfer in a binary system or via merger / collision with at least one ms star ( see * ? ? ?
* for a recent review ) .
@xcite suggested that triples might be a natural progenitor for bsss , as the high eccentricities attained during lk cycles can lead to tidal interaction , mass loss or even a merger of the inner binary components .
although dynamically formed triples are expected to form quite efficiently in gcs , their long term dynamical evolution in gcs has not been yet investigated .
hence , until now their role in bss formation remains largely unconstrained . in what follows we use our monte carlo models coupled with the results of the three - body integrations to determine the relative contribution of triples to bsss in gcs .
besides bsss possibly produced by the lk mechanism , we also computed the number of bsss due to binary - mediated interactions , i.e. , mass transfer and physical collisions .
we computed the number of bsss at the observation time of @xmath193 in each model in the following way .
we first estimate the ms turn - off mass ( @xmath194 ) given the metallicity and age of the model cluster in question . from the snapshot containing all stars
we then extract those that are still on their ms and has mass @xmath195 .
in theory , any ms star with mass above @xmath194 are bsss , however , observationally , it is hard to distinguish bsss from the ms if their masses are not high enough compared to @xmath194 .
this simple mass - based prescription results in very good agreement between observationally extracted bss numbers , their stellar properties , and known correlations between bss and observable properties of the cluster ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for more details ) .
figure [ fig11 ] shows the relative importance of bss production via traditional binary - mediated channels and triples .
since the triples were not followed in the monte carlo models , bsss created via lk cycles leading to mass transfer or merger between the inner binary can only be calculated for each triple produced in the models . for these cases
we can not directly consider the subsequent stellar evolution of the bsss .
for example , whether followed by a merger the bss created via a triple - mediated channel later evolves off of the ms before the observation time of @xmath196 , can not be directly determined in this setup .
indeed , in detailed models it had been found that bsss can live in the ms for several billion years after formation .
it is also expected that the median age since formation for the observed bsss can vary depending on the dominant formation channel for the bsss in a real cluster ( e.g. , * ? ? ?
the real age since formation for bsss , even in these detailed models might not be accurate , since the residual rejuvenated lifetime for a real bss depends intricately on the details of the degree of h - mixing in the core which in turn depends on the details of the interaction that produced it ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . instead , we report two quantities for the triple - mediated bsss- 1 ) the total number of mergers or roche - filling architectures created from dynamically formed triples during the full evolution of the model clusters , and 2 ) the same , but for binaries with a total mass @xmath195 and created only within the last @xmath197 , an estimated upper limit of today s observed bsss in typical galactic gcs @xcite .
we find that the relative importance of triple - mediated channels is low compared to binary - mediated channels for bss production .
the importance of triple - mediated channels is @xmath198 for the cluster models considered in this work .
there is a lot of interest in bsss found in binaries .
often these systems are thought to be produced via stable mass transfer in a binary where the donor stays bound to the accretor . on the other hand ,
these may also be produced via triple - mediated channels where the inner binary merges and the outer binary is hard and remains unbroken ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we find that the contribution from triple - mediated channels for binary bsss is also small being @xmath198 of the number of mergers due to binary - mediated processes .
in addition , we find that the binary bsss coming via binary - mediated channels are _ not _ simply coming from mass transfer in a binary .
collisionally produced bsss can frequently acquire a binary companion via binary - single scattering encounters .
since bsss are more massive compared to typical cluster stars at late ages , initially single bsss , created earlier via a merger or physical collision can frequently acquire a binary companion via exchange with another normal star in a binary through binary - single interactions .
this is consistent with earlier work that investigated the detailed history and branching ratios of various binary - mediated formation channels for bsss produced in similar models @xcite .
gyr of the cluster evolution .
these distributions have been normalized to the total number of merging systems .
, width=220 ]
the first gw signal from a compact object binary is likely to be detected in the coming years by advanced ground - based laser interferometers @xcite .
most gw searches adopt matched filtering techniques in which the detector signal is cross - correlated with banks of theoretical gravitational waveforms to enable detection of the weak signals . mainly due to computational limitation ,
such banks of theoretical waveforms are exclusively composed of circular binaries . however , there are many reasons to extend the searches to also include eccentric inspirals .
first , there have been recent developments of dedicated techniques that will potentially enable the detection of eccentric sources @xcite .
second , eccentric inspirals produce a gw signal which is distinct from that of circularly inspiraling binaries , and provides more insights on the strong - field dynamics @xcite . finally ,
as discussed below , there are several dynamical mechanisms that can lead to the formation of eccentric gw sources .
the detection of such eccentric sources could provide with important information about properties ( e.g. , triple / binary fraction , central densities ) of compact object populations in gcs and galactic nuclei that will be otherwise inaccessible .
several mechanisms have been proposed for the formation of eccentrically inspiraling compact binaries for high frequency gravitational wave detectors .
these include : ( i ) bh - bh scattering in steep density cusps around massive black holes ( mbhs ; * ? ? ?
* ; * ? ? ?
* ) , ( ii ) binary - single compact object encounters in star clusters @xcite and ( iii ) lk resonance in hierarchical triple systems forming in the dense stellar environment of galactic nuclei @xcite or stellar clusters ( e.g. , * ? ?
* this paper ) . in this section
we discuss the rate for eccentric compact binary coalescences which could be detectable by aligo . with the caution that rate estimates remain subject to significant uncertainties
, we find that bh binary mergers from hierarchical triples in gcs are likely to dominate the production of eccentric inspirals that could be detectable by advanced ground - based laser interferometers .
* single - single captures .
* in galactic nuclei with mbhs , nuclear relaxation times can be often less than a hubble time , and can result in the formation of steep density cusps of stellar - mass bhs as a consequence of mass segregation against the lower mass stars ( e.g. , * ? ? ?
@xcite proposed that in such dense population environments bh binaries can efficiently form out of gravitational wave emission during bh - bh encounters . after forming , these bh binaries rapidly inspiral and merge within a few hours .
@xcite found that @xmath199 of these coalescing binaries will have an eccentricity larger than @xmath100 inside the aligo frequency band .
the predicted rate for eccentric capture of @xmath200 bh - bh encounters is @xmath201 @xcite .
table 1 of @xcite presents the merger rate per mw - like galaxy and for different cusp models .
the plausible pessimistic , likely , and optimistic rates can be taken from models a@xmath202 , e-2 , and f-1 of their table 1 .
the most optimistic scenario , given by model f-1 , produces a merger rate of @xmath203 .
the realistic scenario , identified here with model e-2 , predicts @xmath204 .
the pessimistic scenario , given by model a@xmath202 , produces @xmath205 .
we can compare these merger rates with those corresponding to eccentric inspirals from gcs .
after multiplying the merger rates in our table 1 by a factor @xmath206 ( roughly the number of gcs in the mw ; * ? ? ?
* ) , we see that the realistic merger rate for @xcite scenario appears to be about one order of magnitude smaller than the realistic merger rate of eccentric inspirals from gcs .
@xcite carried out @xmath207-body simulations of star clusters around mbhs to study the formation of bh binaries in nuclear clusters and to put constraints on the aligo detection rates for the @xcite model .
@xcite derived an aligo expected detection rate of @xmath208 depending on the maximum horizon distance assumed and the mass ratio of mbh to the surrounding cluster .
as also note by @xcite , these rates are likely to be an overestimate of the real event rate from bh - bh captures .
they were in fact obtained by assuming that the mass fraction of bhs is @xmath209 of the total cluster mass , which is only reasonable within about one - tenth of the mbh influence radius of a mass - segregated nuclear cluster @xcite . given that most of the captures were found to occur near the half - mass radius of the nuclear cluster , where mass segregation will not significantly alter the initial number fraction of different mass species @xcite , a more reasonable choice would be to simply adopt a bh mass fraction similar to that expected on the basis of the initial mass function of the underlining population .
for example , in the galactic center there is evidence for a standard kroupa - like initial mass function @xcite which will result in @xmath210 only of the total mass in bhs @xcite . using this standard choice for the initial mass function and adopting the same quadratic scaling on the bh mass fraction of @xcite will cause the estimated detection rates to drop by approximately 3 orders of magnitude .
* binary - single encounters . *
the in - cluster detection rate reported in table 1 includes bh mergers due to single - binary interactions . assuming that @xmath210 of these mergers will have a finite eccentricity within the aligo band @xcite results in a negligible contribution of binary - single encounters to the the population of eccentric inspirals when compared to the triple channel .
@xcite studied binary - single stellar scattering occurring in dense star clusters as a source of eccentric ns - ns inspirals for aligo . during the chaotic gravitational interaction between the three bodies , a pair of compact objects
can be driven to very high eccentricities such that they can inspiral through gravitational wave radiation and merge .
@xcite argued that @xmath210 of dynamically assembled ns - ns merging binaries will have a substantial eccentricity at @xmath211 frequency .
@xcite also give a simple order of magnitude estimate of the merger rate for dynamical ns - ns star eccentric inspirals assembled in gcs of @xmath212 . to compare to the event rate we found for eccentric
lk induced mergers in gcs we recompute the @xcite rates by rescaling with the different gc number density used in section [ d - rates ] .
@xcite calculate the rate of eccentric binaries in the aligo band adopting @xmath213 .
taking @xmath214 instead , assuming that aligo will see ns - ns mergers up to a sky - averaged distance of @xmath215 @xcite , and optimistically assuming that mergers are distributed uniformly over the lifetime of the globulars ( as opposed to happening early on in the lifetime of the cluster ) , this translates into a possible aligo detection rate for ns - ns eccentric inspirals of @xmath216 .
we note that @xcite give in their table ii the potential detection rates for the model of @xcite .
these rates appear to be larger by one order of magnitude than our rate estimate above .
we believe however that the rates computed by @xcite were mistakenly enhanced due to a misinterpretation of @xcite results .
@xcite found that about `` @xmath210 of _ dynamically _ assembled non - eccentric binaries '' will have a finite eccentricity at @xmath128 .
@xcite computed the rates in their table ii as the @xmath210 of the _ total _ compact merger rate given in table v of @xcite which corresponds , however , to predictions for _ field _ mergers . in order to correct for this
we have to multiply the rates in table ii of @xcite by the expected fraction of all compact object mergers that are dynamically assembled in star clusters . taking this fraction to be @xmath174 of all ns - ns mergers @xcite , and assuming that @xmath210 of these are eccentric inspirals results in @xmath217 , @xmath218 , and @xmath219 for the low , realistic , and high detection rates for binary nss .
these estimates appear to be in good agreement with the rate of @xmath216 we previously computed .
although very simplified , the simple calculations presented above show that , even when assuming a uniform rate of mergers , the rate of eccentric ns binaries in the aligo frequency band is likely 1 - 2 orders of magnitude smaller than the realistic estimates for eccentric bh binaries forming through the lk mechanism in gcs . *
massive black hole mediated compact - object mergers in galactic nuclei .
* @xcite studied the evolution of compact object binaries orbiting a mbh . near the galactic center , where orbits are nearly keplerian , binaries form a triple system with the mbh in which this latter represents the outer perturber of the triple .
the mbh can subsequently drive the inner binary orbit into high eccentricities through lk resonance , at which point the compact object binary can efficiently coalesce through gw emission .
@xcite found that @xmath220 of such coalescing binaries will evolve through the dynamical non - secular evolution similar to that described in section [ bd - oaa ] , leading to finite eccentricity sources in the aligo sensitivity window .
@xcite estimated a bh - bh eccentric inspiral rate for a mw - like galaxy between @xmath221 and @xmath222 .
these rates are somewhat comparable to the rates for bh - bh mergers due to gw captures derived by @xcite ; however , they are still smaller than the rates for eccentric inspirals assembled in gcs for a mw - like galaxy given in table 1 , and , even under the most optimistic assumptions , they are still only @xmath223 of the realistic and @xmath224 of the high / optimistic rate in @xcite
. given the realistic bh - bh detection rate with aligo of @xmath225 @xcite , the predicted overall rate of eccentric compact binary coalescence in the aligo band induced by mbh mediated lk cycles in galactic nuclei is @xmath226 @xcite .
we conclude that , the overall contribution of eccentric gw sources from galactic nuclei is likely negligible compared to the rate of eccentric mergers arising from hierarchical triples in gcs .
the bsss observed abundantly in all clusters have long been identified as a way of identifying important stages in the host cluster s history .
however , interpretation of observed trends between the bss number and cluster properties is dependent on the relative importance of the various bss formation channels in these clusters .
especially , there has been a long history of carefully considering correlations between the number of bsss produced in a cluster and several dynamically important observed cluster properties , including the total mass , total number of binaries , and the central collision / interaction rate , and what they mean @xcite .
although , it has been pointed out that formation of bsss from triple - mediated interactions may be important , at least in open clusters @xcite , this has never been self - consistently investigated in realistic and evolving cluster models @xcite .
as part of this study we have considered evolution of triples created dynamically in the cluster models and analyzed them as potential bss progenitors .
upon comparison of the estimated numbers of the bsss resulting from triple - mediated formation channels to those formed via binary - mediated channels ( mass transfer and physical collisions due to binary - mediated interactions ) , we find that , at least for gc - like densities , contribution from triple - mediated channels is quite low . for the triple - mediated bsss
, we did not consider whether upon formation the rejuvenated ms star will remain in the region expected of bsss in a color - magnitude diagram .
the rejuvenated lifetime of bsss is model dependent and hard to estimate . for the binary - mediated bsss
we assume full - mixing of hydrogen . in case of triple - mediated bsss
, we simply compare the total number of bsss created throughout the lifetime of the gc and that formed within the last @xmath197 and with mass @xmath195 .
thus , the actual contribution from triple - mediated interaction may be even lower if many of these bsss evolve off of their rejuvenated ms before the observation time of @xmath196 .
we note that above we have not considered bss formation through binary evolution in short period binaries affected by the lk mechanism coupled with tidal friction .
the shrinkage of the inner binary semi - major axis as well as the high eccentricities induced via the lk mechanism , combined with an expansion in the size of each inner binary star , could lead to coalescing / strongly - interacting post - ms binaries . in order to address the importance of this alternative channel we determined whether the stars would undergo mass transfer by requiring their periapsis distance to be smaller than the roche limit computed through equation ( [ roche ] ) and setting the stellar radius of the most massive component equal to that evaluated at the tip of the red giant branch @xcite .
the periapsis distance was evaluated using the orbital parameters at the end of the three - body integrations .
we only considered stars with mass @xmath227 so that their ms life time was shorter than the host gc age .
also , we did not consider systems that will have undergone mass transfer on the red giant branch even without the help of lk cycles and tidal dissipation .
we found that only 694 of the total 19073 binaries that did not merge during the ms would merge on the red giant branch through a combination of lk cycles , tidal friction and the expansion of the stellar radius caused by stellar evolution .
this calculation suggests that the number of bsss formed through a combination of lk dynamics and stellar evolution in our models is negligible . finally , note that we did not consider primordial triples that could somewhat increase the number of mergers from stellar triples shown in figure [ fig11 ] .
however , we think that mergers occurring in primordial triples will have a negligible contribution to bsss compared to mergers from dynamically assembled triples .
in fact , due to the high stellar densities of gcs the outer binaries in primordial triples are likely to be involved in several dynamical encounters which will lead to their disruption in the first few gyrs of cluster evolution .
the majority of triples in the catalog of @xcite have outer period @xmath228days , which according to equation ( [ coll ] ) corresponds to a typical survival time @xmath229yr within the core of a gc and @xmath230 in the cluster outskirts ( see also fig . 2 of * ? ? ?
moreover , given the old age of gcs stellar merger products that occurred in primordial triples would have already evolved off the main sequence , so that only dynamically formed triples could then produce bsss at later stages of cluster evolution .
in this paper we studied the long - term evolution of dynamically assembled triples in gcs .
the initial conditions for the triples were obtained directly from detailed monte carlo models of the clusters .
these systems consisted mostly of triples containing an inner stellar binary and triples containing a pair of bhs .
the triple initial conditions were integrated forward in time using a high accuracy three - body integrator which incorporates relativistic corrections and tidal terms to the equations of motion .
the direct integrations allowed us , for the first time , to put constraints on the contribution of the lk mechanism to the population of coalescing bh binaries and bsss in gcs .
the main results of our study are briefly summarized below .
\(1 ) we find that for the majority of dynamically assembled triple systems in gcs , the time - scale for changes of angular momentum during the high eccentricity phase of a lk cycle becomes shorter than the orbital time - scales .
we conclude that the standard orbit average equations of motion , often employed in previous studies , can not accurately describe the evolution of such triples .
\(2 ) bh triples assembled dynamically in gcs often evolve such that the inner binary angular momentum changes on timescales of order or shorter than the inner binary orbital period ( see figure [ examples ] ) .
this can lead to the formation of eccentric gw sources in the frequency band of aligo detectors ( see figure [ fig3 ] ) .
\(3 ) we estimated a realistic aligo detection rate of bh binary mergers due to the lk mechanism of @xmath0 .
we also compared this rate to that for bh mergers due to other processes ( see table 1 ) .
we find that mergers induced by lk oscillations are likely to be the main contributor to in - cluster gw sources for aligo detectors , and represent @xmath175 of the total detection rate of bh binary mergers assembled in gcs .
\(4 ) about @xmath1 of all lk mechanism induced bh mergers in gcs will have an eccentricity larger than @xmath231 at @xmath128 frequency , such that eccentric waveform templates and/or dedicated search strategies will be needed for their efficient detection ( see figure [ fig4 ] ) .
most of these eccentric sources will have an extremely large eccentricity ( @xmath232 ) at the moment they first enter the aligo frequency band . assuming perfect waveform templates , we find a possible detection rate for eccentric mergers as large as a few events a year with aligo .
we also show that the triple channel is likely the dominant formation mechanism for the production of eccentric inspirals that will be potentially detectable by aligo .
\(5 ) in the case of triples containing a stellar binary , the inner binary angular momentum often changes on timescales shorter than the tertiary orbital period , but rarely shorter than the _ inner _ binary orbital period .
this implies that `` clean '' collisions between two stars , in which all passages prior to the collision are greater than the tidal dissipation scale , are rare ( see figure [ period - pdf ] ) .
\(6 ) we estimated the number of bsss expected to form through mergers induced by the lk mechanism as well as those formed through other processes , including binary stellar evolution , binary - single scattering and direct collisions ( see figure [ fig11 ] ) .
we find that the contribution of lk induced mergers to the population of bss populations in gcs is typically @xmath198 of the total . even in clusters with relatively low central densities ( @xmath233 ) only up to @xmath124 of bss binaries could have formed in dynamically assembled triples .
we are grateful to s. mikkola who wrote the archain algorithm and who generously assisted us in using it .
during the course of this work , we have benefited from conversations with several colleagues , including cole miller and hagai perets .
we thank the referee for useful comments that helped to improve this paper .
fa acknowledges support from a ciera postdoctoral fellowship at northwestern university .
this work was supported by nasa atp grant nnx14ap92 g and nsf grant ast1312945 .
sc also acknowledges support from nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 ; the grant identifying number is hst - ar-12829.004-a .
vk and far also acknowledge support from nsf grant phy-1066293 through the aspen center for physics .
merritt , d. 2013 , dynamics and evolution of galactic nuclei , by david merritt .
isbn : < isbn>9780691158600</isbn>. 544 pp .
| 6 x 9 | 5 halftones
. 136 line illus .
princeton : princeton university press , 2013 , | hierarchical triple - star systems are expected to form frequently via close binary - binary encounters in the dense cores of globular clusters . in a sufficiently inclined triple , gravitational interactions between the inner and outer binary can cause large - amplitude oscillations in the eccentricity of the inner orbit ( `` lidov - kozai cycles '' ) , which can lead to a collision and merger of the two inner components . in this paper
we use monte carlo models of dense star clusters to identify all triple systems formed dynamically and we compute their evolution using a highly accurate three - body integrator which incorporates relativistic and tidal effects .
we find that a large fraction of these triples evolve through a non - secular dynamical phase which can drive the inner binary to higher eccentricities than predicted by the standard secular perturbation theory ( even including octupole - order terms ) .
we place constraints on the importance of lidov - kozai - induced mergers for producing : ( i ) gravitational wave sources detectable by advanced ligo ( aligo ) , for triples with an inner pair of stellar black holes ; and ( ii ) blue straggler stars , for triples with main - sequence - star components .
we find a realistic aligo detection rate of black hole mergers due to the lidov - kozai mechanism of @xmath0 , with about @xmath1 of these having a finite eccentricity when they first chirp into the aligo frequency band . while rare , these events are likely to dominate among eccentric compact object inspirals that are potentially detectable by aligo . for blue stragglers , we find that the lidov - kozai mechanism can contribute up to @xmath2 of their total numbers in globular clusters .
# 1 # 1 |
negatively charged nitrogen - vacancy centers in diamond@xcite , phosphorous atom impurities in silicon@xcite and other schemes based on point defects embedded in semiconductors have been widely studied as alternatives to develop quantum bits@xcite .
one of the necessary conditions for quantum computing is long electron spin decoherence times to insure a minimum of fault - tolerance @xcite . in diamond
s nitrogen - vacancy centers@xcite , silicon vacancies in silicon carbide@xcite , silicon@xcite and any iii - v based quantum dots@xcite the fluctuating nuclear bath is the main source limiting spin coherence time .
the nuclear dipole - dipole interaction is believed to be the dominant mechanism behind the diffusion - induced electron - spin decoherence@xcite . to protect the dynamics of the nuclear spins of point defects from the decoherence induced by the environment ,
semiconductors mainly composed of spin - zero isotopes as silicon and carbon are preferred over iii - v semiconductors@xcite .
even though the two stable isotopes of ga , @xmath0ga and @xmath1ga , have nuclear spin @xmath2 , in dilute nitride gaasn , point interstitial defects give rise to paramagnetic centers that have very peculiar and useful properties .
one of them is the spin dependent recombination ( sdr ) @xcite . in ga(in)nas alloys ,
ga@xmath3 paramagnetic centers with only one bound electron can selectively capture another conduction band ( cb ) electron with the opposite spin orientation @xcite . due to this mechanism ,
paramagnetic centers act as a spin filter that blocks the recombination of cb electrons with the same spin and efficiently capture electrons whose spin is in the opposite direction . in the centers ,
the bound and captured electron form a singlet state that is destroyed as either one of the electrons recombines to the valence band ( vb ) .
it is important to note that while the lifetime of a cb electron with the opposite spin to the paramagnetic electrons is a few pico - seconds , the lifetime of a spin polarized cb electron when the majority of bound electrons are polarized parallel to it may extend to nanoseconds . as a consequence cb electrons spin polarization can reach over 80% under circularly polarized incident light .
additionally the photoluminescence ( pl ) intensities can be as high as 800% under circularly polarized optical excitation compared to a linearly polarized one @xcite .
the increase in cb electron population allows even for the detection of electron spin polarization by electrical means due to a giant photoconductivity effect under circulalrly polarized light@xcite . whereas in diamond and silicon the optical excitation acts directly on the point defects , in gaasn
the bound electron is dynamically spin polarized due to the recombination of spin polarized cb electrons on the paramagnetic centers .
although this mechanism is radically different , ga@xmath4 centers themselves are very similar to nitrogen vacancies and silicon phosphorous atoms .
recent experiments on gaasn subject to a weak magnetic field in faraday configuration @xcite have shown consistently an enhancement in the spin - filtering mechanism in comparison to the zero magnetic field case .
the general agreement among different models@xcite is that the hyperfine interaction ( hfi ) between the bound electron and the nucleus in the centers is the key element behind this phenomenon . at low magnetic fields ,
the spin - filtering effect is reduced due to spin state mixing induced by the hfi . for higher values of the magnetic field such that the zeeman energy exceeds the hfi , the pure bound electron spin states and ,
consequently , the spin filtering effect are recovered . even though the role of the hfi is well established , some aspects of these phenomena lacks full understanding . some observations point to the existence of an overhauser - like effective magnetic field@xcite
whose origin is yet unclear .
it manifests as a shift in the band to band pl intensity or the degree of cb electron spin polarization as functions of the magnetic field .
both features shift to the positive and negative regions of the magnetic field for right and left circularly polarized light respectively .
another aspect that needs further study is the nuclear interaction between ga@xmath4 centers and adjacent ga atoms that would lead to nuclear spin relaxation ( nsr ) . the overhauser - like magnetic field has been correctly reproduced in ref .
[ @xcite ] , however , nuclear spins in this model are assumed to relax very rapidly and the origin of the relaxation mechanism is unclear . on the other hand , the model presented in ref .
[ @xcite ] considers two phenomenological and arbitrary nsr times for traps with one and two bound electrons . despite the improvements in this work , the model considers ga centers with @xmath5 nuclear spins instead of @xmath2 in order to simplify the kinetic equations . in this paper
we examine the spin dynamics in gaasn alloys .
we propose a model based on the master equation for the density matrix that describes the main interactions between cb electrons , vb holes and paramagnetic traps .
it addresses the problems on spin relaxation mechanisms and overhauser - like magnetic field .
in fact , our results show that the overhauser - like magnetic field strongly depends on the nsr times and the mechanisms behind them
. the model developed here is based on the preexisting two charge model ( tcm ) @xcite .
in addition to the sdr processes , it contains the mechanisms that give rise so nsr and hfi .
nsr is addressed through the wangsness , bloch , and redfield relaxation theory @xcite .
two interactions are explored as possible candidates to produce the nsr : dipolar interaction between neighbouring nuclei@xcite and quadrupolar @xcite interaction with charge fluctuations in the environment . to further explore the role of hfi in ga centers ,
we have studied the time resolved electronic and nuclear spin polarizations under pulsed excitation . using these results we outline a method based on a pump - probe photoluminescence scheme to trace the coherent evolution of coupled electrons and nuclear spins as they flip - flop due to the hfi . this paper is organized as follows . in sec .
[ model ] we introduce the master equation model that describes the key processes in the spin dynamics of electrons and nuclei in gaasn .
the role of the hfi and the nsr mechanism are discussed in this section .
the mathematical forms of the dipolar and quadrupolar dissipators are introduced in sec .
[ nsrsec ] . in sec .
[ nsrcwregime ] we establish the main mechanism behind the nuclear spin relaxation in ga@xmath6 centers by comparing the theoretical results issued by the model with previously existing experimental results .
simulations in the pulsed excitation regime are presented in sec .
[ coherent ] . in sec .
[ summ ] we summarize the main results .
the system mainly consists of four elements : vb holes , cb electrons , unpaired traps ( ut ) and paired traps ( pt ) . whereas vb holes spin relaxes with a characteristic times below @xmath7@xcite , cb electronic spin relaxes on a typical time in the range @xmath8@xcite .
therefore , for the sake of simplicity , we consider vb holes to be unpolarized and electrons may occupy spin - up or spin - down states .
gallium paramagnetic traps or uts can be understood as a @xmath2 nuclear spin coupled to a @xmath5 bound electron spin via hfi . on the other hand ,
when a pt is formed if a cb electron is captured by an ut , the bound and captured electrons form a singlet state that can not interact with the nuclear spin .
thus the quantum state basis that describes this system must have : ( i ) one state for holes in the vb , ( ii ) two states for the spin up and spin down cb electrons , ( iii ) eight states that account for the nucleus - bound electron system in the uts and ( iv ) four states for the nuclear spin in the pts . the complete quantum state basis is therefore given by @xmath9 where @xmath10 is the vb hole state and , @xmath11 and @xmath12 are the spin - up and spin - down cb electron states respectively .
the following eight states @xmath13 , @xmath14 , @xmath15 , @xmath16 are the bound - electron and nuclear spin states projected along the @xmath17 axis corresponding to the ut .
finally , the pts are described by the nuclear - spin states @xmath18 , @xmath15 , @xmath19 .
the dynamics of the four parts of the system and their interactions can be described through the master equation @xmath20+\mathcal{d}\left(\hat\rho\right),\ ] ] where @xmath21 is the density matrix , @xmath22 is the hamiltonian and @xmath23 is the dissipator .
the hamiltonian contains the internal interactions among the four components of the system .
we are interested in the combined effect of an external magnetic field and the hfi in ga centers , therefore the hamiltonian must contain zeeman and hfi terms .
the interactions among the different parts of the system and the surroundings are accounted for by the dissipator .
these are mostly interactions with the electromagnetic field , occurring during recombination or excitation processes , or interactions with the nuclear spin environment .
the main processes introduced in our model are schematized in fig
the hamiltonian is given by @xmath24 the first and second terms in the hamiltonian correspond to the zeeman interaction for cb and bound electrons in ga centers respectively .
in these terms @xmath25 , @xmath26 , @xmath27 is the external magnetic field , @xmath28 is the bohr magneton and , @xmath29 and @xmath30 are the cb and bound electrons gyromagnetic factors .
the hfi term , the third one on the right - hand side of eq .
( [ spinhamiltonian ] ) , couples the bound electron and the nuclear spin in uts .
the hyperfine parameter is given by @xmath31 .
the spin operator for cb electrons that appears in the first zeeman term of the hamiltonian is @xmath32 . in uts ,
the bound electron spin and nuclear spin operators entering the hfi term are @xmath33 , and @xmath34 respectively .
as the singlet state formed in the pts interacts neither with the external magnetic field nor with their nuclear spin @xmath35 , it does not appear in the hamiltonian .
the relaxation mechanisms of cb electrons and nuclei in the ga centers are described by the dissipator @xmath36 .
also , the photoexcitation and recombination of electrons will be accounted for by @xmath37 .
our approach to formulating a suitable dissipator consists in expanding the relevant operators as linear combination of the elements of an operator vector space . in principle
this set should be formed in the basis ( 1 ) by linearly independent @xmath38 hermitian matrices .
however , the density matrix structure is considerably simplified by assuming that the four components of the system ( cb , ut , pt , vb ) are interconnected only by the dissipator . as the four parts of the system are exclusively coupled by the recombination or excitation processes , this is a reasonable assumption .
thus , the density matrix operator can be presented in the block diagonal form : @xmath39 where the four blocks @xmath40 ( @xmath41 ) , @xmath42 ( @xmath43 ) , @xmath44 ( @xmath45 ) and @xmath46 ( @xmath47 ) are the partial density matrices of vb holes , cb electrons , uts and pts respectively . given
that @xmath21 takes the form of a block diagonal matrix since no coherences can arise between the four components , it suffices to consider the smaller vector space of @xmath48 hermitian matrices that generate the four blocks .
we start by finding an internal space of hermitian matrices @xmath49 that spans the @xmath50 relevant elements of the @xmath38 matrix .
the generators in this set can be chosen in such a way that they are hermitian and orthogonal with respect to the scalar product given by the trace @xmath51=\delta_{i , j}\text{tr}\left[\hat\lambda_{i}^{2}\right].\ ] ] this choice conveniently links the inner product with the expected value of a given operator @xmath52 @xmath53= \langle \hat\rho , \hat o\rangle,\ ] ] acting on @xmath54 . in this manner any operator can be expanded as a linear combination of the elements of ( [ or : lambda ] ) as @xmath55}{\text{tr}\left[\hat\lambda_q^2\right ] } \hat \lambda_q.\ ] ] a very convenient set of operators is the one formed by the generators of the unitary groups @xmath56 ( vb holes ) , @xmath57 ( cb electrons ) , @xmath58 ( pts ) and @xmath59 ( uts ) .
the operators forming this set are not only of physical significance but they are also linearly independent and orthogonal with respect to the trace .
explicitly , the set of operators in ( [ or : lambda ] ) is given by @xmath60 where @xmath61 , @xmath62 , @xmath63 and @xmath64 generate the vb , cb , ut and pt bolcks .
the vb hole population density operator can be represented by the matrix @xmath65 the operators that generate the the cb block can be compiled in the matrix @xmath66 where the electron population density in the cb is given by @xmath67 and their spin operators are @xmath62 for @xmath68 . here
@xmath69 is the @xmath43 identity matrix , and @xmath70 for @xmath68 are the standard pauli spin matrices that fullfil the usual spin commutation relations
@xmath71=i \sum_{k=1,2,3}\epsilon_{ijk}\hat s_k.\ ] ] this definition allows us to write the matrices that generate the ut and pt blocks in the compact forms @xmath72 and @xmath73 according to this scheme the population density of uts is @xmath74 , and the one for pts is @xmath75 .
similarly , the cb electrons spin operators are : @xmath76 , @xmath77 and @xmath78 .
we have the same case for the bound electron spin operator components in uts where @xmath79 , @xmath80 and @xmath81 .
the operators of the nuclear spin of uts and pts can be expressed as linear combinations of the elements of @xmath82 as @xmath83 where @xmath84 and @xmath85.\ ] ] the dissipator can be separated in six parts @xmath86 here @xmath87 contains the vb hole and cb electron photogeneration terms .
the sdr processes that mainly consist of the selective capture of cb electrons in uts according to their relative spin orientation and the subsequent recombination to the vb are described by the @xmath88 dissipator .
cb and bound electron spin relaxation are accounted by @xmath89 and @xmath90 .
nsr in ut s and pt s is introduced through the dissipators @xmath91 and @xmath92 .
the term @xmath87 that models the generation of electrons is given by @xmath93 where @xmath94 and @xmath95 are the spin up and spin down electron generation rates .
the unitary vector @xmath96 points in the direction of the incident light . to build the @xmath97 part of the dissipator we resort to the tcm@xcite given by the following kinetic equations ( @xmath98=1 ) @xmath99 in the previous equations ,
the population densities of cb electrons and vb holes are given by @xmath100 and @xmath101 respectively .
the density of uts is @xmath102 , @xmath103 is the density of electron singlets hosted by the centers ( pts ) , and consequently @xmath104 is the total density of ga centers .
the vectors @xmath105 and @xmath106 represent the average of free and bound electron spin polarizations .
the spin dependent capture of electrons in the ga centers is ensured by the recombination rate terms @xmath107 and @xmath108 where @xmath109 is a constant .
notice that these two terms vanish when the system is fully polarized , i.e. @xmath110 , @xmath111 , @xmath112 and @xmath113 .
the recombination rate of one of the electrons trapped in the ga centers to the vb is given by the terms @xmath114 where @xmath115 is a constant .
we thus require that the dissipator s structure is such that the master equation reduces to eqs .
( [ mag : eq3])-([mag : eq2 ] ) when the hfi is lifted ( @xmath116 ) .
this may be achieved by identifying @xmath100 , @xmath101 , @xmath102 , @xmath103 , @xmath105 and @xmath106 with the quantum statistical average of the corresponding operators , namely @xmath117 $ ] , @xmath118 $ ] , @xmath119 $ ] , @xmath120 $ ] , @xmath121 $ ] and @xmath122 $ ] . as well , the quantum statistical average of any generator of @xmath82 is given by @xmath123 $ ] and therefore , the density matrix can be expanded as @xmath124 } { \mathrm{tr}\left[\hat{\lambda}^{2}_{q}\right ] } \hat{\lambda}_{q}=\sum_{q=1}^{85 } \frac{\lambda_{q}}{\mathrm{tr } \left[\hat{\lambda}^{2}_{q}\right]}\hat{\lambda}_{q}.\end{aligned}\ ] ] the sdr part of the dissipator @xmath88 can also be expanded in terms of the elements of @xmath82 as @xmath125 } { \mathrm{tr}\left[\hat{\lambda}^{2}_{q}\right]}\hat{\lambda}_{q } = \sum_{q=1}^{85 } \frac{c_q } { \mathrm{tr}\left[\hat{\lambda}^{2}_{q}\right]}\hat{\lambda}_{q}.\ ] ] to determine the coefficients @xmath126=\mathrm{tr}[\hat \lambda_q\hat{\mathcal{d}}_{\mathrm{sdr}}]$ ] we insert ( [ dissipator : sdr ] ) in the master equation ( [ masterequation ] ) and multiply by @xmath127 , @xmath61 , @xmath128 , @xmath129 , @xmath130 or @xmath131 . by taking the trace of the resulting equation
we readily find the coefficients @xmath132 & = & -c_p p n_2,\label{bal:1}\\ c\left[\hat n\right ] & = & -c_n\left(nn_1 - 4\boldsymbol{s}\cdot\boldsymbol{s}_c\right),\\ c\left[\hat n_1\right ] & = & -c_n\left(nn_1 - 4\boldsymbol{s}\cdot\boldsymbol{s}_c\right ) + c_pn_2p,\\ c\left[\hat n_2\right ] & = & c_n\left(nn_1 - 4\boldsymbol{s}\cdot\boldsymbol{s}_c\right ) -c_pn_2p,\\ c\left[\hat{\boldsymbol{s}}\right ] & = & -c_n\left(\boldsymbol{s}n_1-\boldsymbol{s}_cn\right),\\ c\left[\hat{\boldsymbol{s}}_c\right ] & = & -c_n\left(\boldsymbol{s_c}n-\boldsymbol{s}n_1\right)\label{bal:6}.\end{aligned}\ ] ] at this point we have considerable freedom since these equations only define 10 of the 85 coefficients needed to fully determine the @xmath88 dissipator . however , the choices get narrowed down by imposing the symmetry and invariance properties that the system is expected to satisfy . as the most basic requirement , the master equation must be invariant under any arbitrary rotation in accordance with the space s isotropy . the tensors @xmath64 and @xmath63 must therefore transform by the corresponding laws . the complete set of coefficients is thus given by @xmath132&=&-c_p p t_{0,0},\label{balance:1}\\ c\left[\hat n\right]&=&-c_n \left ( s_{0 } u_{0,0,0 } -4 \sum_{r=1}^3 s_{r}u_{r,0,0}\right),\label{balance:2}\\ c\left[\hat s_k\right]&=&-c_n \left ( s_{k } u_{0,0,0 } -s_{0 } u_{k,0,0 } \right),\label{balance:3}\\ c\left[\hat u_{0,j , i}\right]&=&c_p p t_{j , i } \nonumber\\ & & \,\,\,\,\,\,\,- c_n\left(s_{0 } u_{0,j , i } -4 \sum_{r=1}^3 s_{r } u_{r , j , i}\right),\label{balance:4}\\ c\left[\hat u_{k , j , i}\right]&=&-c_n \left(s_{0 } u_{k , j , i } - s_{k } u_{0,j , i } \right),\label{balance:5}\\ c\left[\hat t_{j , i } \right]&= & -c_p p t_{j , i}\nonumber\\ & & + \,\,\,\,\,\,\
, c_n \left ( s_{0 } u_{0,j , i } -4 \sum_{r=1}^3 s_{r } u_{r , j , i } \right),\label{balance:6 } \end{aligned}\ ] ] where @xmath133 and @xmath134 .
the spin precession terms @xmath135 and @xmath136 , absent in eqs .
( [ balance:1])-([balance:6 ] ) , are accounted for by the zeeman terms in the hamiltonian ( [ spinhamiltonian ] ) since they concern the coherent evolution of the system .
the sdr part of the dissipator also excludes the spin relaxation of cb and bound electrons .
these terms enter the dissipator through @xmath89 and @xmath90 as @xmath137 } \hat{\lambda}_{q}=-\frac{2}{\tau_{s}}\sum_{i=1}^{3}s_{i}\hat{s}_{i},\label{diss : s}\\ \hat{\mathcal{d}}_{\mathrm{sc } } & = & -\frac{2}{\tau_{sc}}\sum_{i=1}^{3}s_{ci}\hat{s}_{ci } .\label{diss : sc}\end{aligned}\ ] ] the cb electron spin relaxation time due to the dyakonov - perel mechanism is given by @xmath138 , while @xmath139 is the phenomenological bound electron spin relaxation time in ga centers @xcite .
the dissipators ( [ diss : s ] ) and ( [ diss : sc ] ) yield the cb and bound electron spin relaxation terms @xmath140 and @xmath141 in the tcm .
to get an insight into the role of the possible mechanisms involved in nsr we consider three different models : non - selective , dipolar and quadrupolar . as a reference we also study the effects of the absence of spin relaxation .
the dipolar and quadrupolar interactions are dealt through the wangsness , bloch , and redfield relaxation theory @xcite summarized in appendix [ redfield : theory ] .
first , we study the non - selective@xcite dissipators for uts and pts given by @xmath142 where @xmath143 and @xmath144 are the bound electron spin and nuclear spin indices .
this dissipator is highly symmetrical .
second , we consider the relaxation due to the dipolar interactions between neighbouring ga nuclei . in this case , the hamiltonian ( [ pert : hamiltonian ] ) only contains the irreducible spherical tensors of rank @xmath145 .
the terms in this hamiltonian correspond to the angular momentum operators interacting with a random local field . substituting ( [ magdipole:0 ] ) and ( [ magdipole:1 ] ) in ( [ simplified : redflied ] ) the dissipators for uts and pts are @xmath146\right ] , \label{d1:d}\\ \hat{\mathcal{d}}_{2 } & = & -\frac{1}{3\tau_{n2}}\sum_{i=1}^{3 } \left[\hat{i}_{2i},\left [ \hat{i}_{2i},\hat{\rho}\right]\right],\label{d2:d}\end{aligned}\ ] ] where @xmath147 and @xmath148 are the @xmath149-th components of the nuclear spin operators for uts and pts respectively .
the nsr times for unpaired and paired traps are considered to be different in principle and therefore are set to @xmath150 and @xmath151 .
finally we study the relaxation owing to the quadrupole interaction with random fluctuation of the local electric field gradient .
the hamiltonian takes the form of ( [ pert : hamiltonian ] ) where @xmath152 . substituting the rank @xmath152 irreducible spherical tensors in eqs .
( [ elecquadrupole:0])-([elecquadrupole:2 ] ) in terms of the nuclear angular momentum components in ( [ simplified : redflied ] ) yield the following dissipators @xmath153\right],\label{d1:q}\\ \hat{\mathcal{d}}_{2 } & = & -\frac{1}{2\tau_{n2}}\sum_{i=1}^{5 } \left[\hat{q}_{2,i},\left [ \hat{q}_{2,i},\hat{\rho}\right]\right].\label{d2:q}\end{aligned}\ ] ] here , the operators @xmath154 and @xmath155 are related to the rank @xmath152 irreducible spherical tensors and therefore can be expressed in terms of the nuclear spin operators as @xmath156 the explicit forms of the dissipators corresponding to the dipolar interaction ( [ d1:d])-([d2:d ] ) and quadrupole interaction ( [ d1:q])-([d2:q ] ) are presented in appendix [ dissipators ] .
the model developed above is used in this section to examine the interplay of the hfi and the nsr mechanisms in the continuous wave ( cw ) and pulsed excitation ( pe ) regimes .
first , the theoretical results are compared with previous experimental observations under cw excitation in order to identify the main interaction behind nsr and hfi in ga centers .
then , we analyze the dynamics of the bound electrons and nuclear spin in the pe regime .
we outline a method for detecting the bound electron and nuclear spin coherent oscillations induced by hfi by means of a pump - probe pl scheme . in order to extract information from the model , we start by building the system of kinetic equations that follow from the master equation ( [ masterequation ] ) . by multiplying both sides of ( [ masterequation ] ) by @xmath157 , inserting the density matrix in the form ( [ ope - densidad ] ) into the resulting expression and taking the trace we obtain a set of @xmath50 differential equations of the form @xmath158\hat\rho\right ] + \mathrm{tr}\left[\mathcal{d}\hat{\lambda}_q\right]\\ = f_q\left(\lambda_1,\lambda_2,\dots,\lambda_n , t\right ) , \,\,\,\ , q=1,2,\dots , n.\label{kinetic : eqs}\end{gathered}\ ] ] unlike the tcm that only considers the sdr mechanism , zeeman interaction and electron spin relaxation , these new kinetic equations also take into account the hfi and nsr .
we study the spin dynamics of electrons and nuclei by numerically solving the system of ordinary differential equations ( [ kinetic : eqs ] ) .
the relevant parameters are then extracted from the thus obtained @xmath159 functions which in turn are the quantum statistical averages .
we assume that before the optical excitation ( @xmath160 ) the uts are equally populated and that the electrons as well as the nuclei are completely unpolarized .
therefore , initially @xmath161 and @xmath162 for @xmath163 .
notice that these initial conditions also imply that at this stage there are no pts , namely @xmath164 .
under cw stimulation , the generation of spin up and spin down electrons is given by the smooth step function @xmath165,\ ] ] where @xmath166 is the excitation power , @xmath167 is the power to generated electron ratio , @xmath168 is the spin polarization degree of the optically generated cb electrons , @xmath169 is the onset time of the excitation and @xmath170 is the duration of the onset .
the system is allowed to evolve for a sufficiently long time ( @xmath171 ) to reach steady state conditions .
some of the parameters as @xmath172 , @xmath173 , @xmath174 , @xmath175 , @xmath176 , @xmath177 and @xmath178 where estimated from previous experimental results @xcite . for the nuclei at the ga centers , the hyperfine parameter was estimated to be @xmath179 , the average hyperfine parameter of the two stable isotopes of ga@xcite .
the nsr times @xmath150 and @xmath151 are determined below by comparing the theoretical calculations with the experimental results .
as we stated above , the two key features behind the hfi in ga centers are a growth in the pl degree of circular polarization @xmath180@xcite and an overhauser - like magnetic field@xcite .
both are observed under circularly polarized excitation and a faraday configuration magnetic field .
more specifically for the first feature , @xmath181 exhibits a minimum close to @xmath182 . as @xmath183 increases , @xmath180 saturates at values above @xmath184 where zeeman energies are comparable to the hfi . in this region the hfi has been completely exceeded by the zeeman interaction and therefore the bound electrons and nuclei in the ga centers are effectively decoupled .
the degree of circular polarization of the cb to vb photoluminescense as a function of the faraday configuration magnetic field can be described by an inverted lorentzian - like curve . in fig .
2 we have calculated the cb electron spin polarization @xmath185 as a function of the magnetic field in faraday configuration @xmath186 under a circularly polarized excitation . in this case
we have chosen the spin relaxation times @xmath187 and @xmath188 that give good quantitative agreement with the experiment .
the degree of the cb electron spin polarization as a function of the faraday configuration magnetic field for the non - selective mechanism is shown in fig . 2 for different pump powers @xmath166 .
as a reference , fig . 2 also presents the behaviour observed in the absence of nsr ( @xmath189 ) as thick dotted lines .
these plots show that despite the hfi , in the absence of spin relaxation , @xmath181 does not display any sign of the spin filtering enhancement .
these results are exactly the same as those obtained with the tcm that does not contain the effects of the hfi .
therefore , in order for the effects of the hfi- as the amplification of the spin filtering effect - to be visible , nsr is essential .
the second feature of this phenomenon is a shift of the minimum of @xmath180 vs. @xmath186 that points to the existence of an overhauser - like magnetic field@xcite .
the @xmath181 curves are shifted to the positive and negative magnetic field regions depending on the helicity of the circularly polarized light .
thereby , under @xmath190 and @xmath191 excitation the minimum is located at @xmath192 and @xmath193 respectively .
the experimental data shows that @xmath194 grows with the excitation power @xmath166 until it apparently saturates at approximately @xmath195 .
the non - selective dissipator yields vanishing @xmath194 as no shift is observed for @xmath180 in fig .
[ figure2 ] .
this dissipator is too symmetric to be able to produce an overhauser - like magnetic field and hence must be ruled out as the leading nsr mechanism .
in contrast , the dipolar and quadrupolar mechanisms yield non - vanishing @xmath194 as it can be seen in figs .
[ figure3 ] and [ figure4 ] .
these two plots show @xmath181 for the dipolar and quadrupolar dissipators for various excitation powers .
even though the shifts produced by both mechanisms qualitatively agree with the experimental observations , only the dipolar one is able to accurately fit the experimental measurements as we discuss below .
in addition to the overhauser - like magnetic field , another feature that strongly depends on the nsr mechanism is the depth of the inverted lorentzian - like @xmath181 curves given by @xmath196 . to discern which of the two mechanisms is the dominant one , we compare our theoretical calculations with the experimental observations of @xmath194 and @xmath197 @xcite .
the power dependence of @xmath194 and @xmath197 is determined by finding the minima @xmath198 and maxima @xmath199 of @xmath181 for different excitation powers . in fig .
[ figure5 ] we plot the isolines for @xmath200 and @xmath201 as functions of the nsr times @xmath150 and @xmath151 .
these two correspond to the experimental results observed for an excitation power of @xmath202@xcite .
the two isolines intersect at @xmath203 and @xmath204 . in accordance with these results ,
the @xmath194 and @xmath197 isolines at other excitation power coincide at similar @xmath150 and @xmath151 values .
collecting the intersecting points of all the experimental results we find that the nsr times must fall within the ranges @xmath205 and @xmath206 .
plots of @xmath194 and @xmath197 as functions of the excitation power are shown in figs .
[ figure6 ] ( a ) and [ figure6](b ) respectively .
the best quantitative agreement with the experimental data is accomplished by using the nsr times @xmath207 and @xmath188 consistent with the ranges above .
whereas the calculated @xmath197 presents a very good agreement with the experimental data , the theoretical values of @xmath194 above @xmath208 show a significant deviation with respect to the experimental observations .
the experimental results suggest that after increasing with the excitation power , @xmath194 saturates at approximately @xmath195 . however , the computed @xmath194 as a function of the excitation vanishes for high powers after reaching its maximum at @xmath195 ( see fig .
[ figure6 ] ) . the quadrupolar mechanism , however , yields systematically non - intersecting isolines regardless of the excitation power value used to calculate them . in fig .
[ figure7 ] we present the @xmath209 and @xmath201 isolines that clearly do not intersect .
this behaviour is observed for all the excitation powers reported experimentally and therefore we must also rule out the quadrupolar mechanism . having identified the nsr mechanism and the corresponding relaxation times we are in a position to consider time resolved simulations .
our aim here is to develop a method to observe the coherent oscillations of bound electrons and nuclei in ga centers .
to do so we propose a pump probe scheme that we describe below . in pe regime ,
the generation terms are given by @xmath210,\ ] ] where @xmath166 is the pulse s average power , @xmath211 is its width and @xmath212 @xmath213 is the period between repeated pulses@xcite .
the pump pulse originates at @xmath160 and the probe is delayed @xmath214 .
figure [ figure8 ] is an outline of the proposed method .
the pump pulse is left circularly polarized and therefore most of the electrons are spin polarized in the @xmath215 direction .
likewise , the probe pulses are left circularly polarized fig .
[ figure8 ] also shows a plot of the bound electron and nuclear spin polarizations as a function of time after being excited by the pump pulse .
the pump and probe pulses widths are exaggerated to make them visible in the given time scale . in this plot
it is possible to observe the electron - nucleus flip - flops as oscillations of @xmath216 ( dashed lines ) and @xmath217 ( thick lines ) that are phase shifted by @xmath218 . as indicated in this diagram , the second pulse at time delays @xmath219 and
@xmath220 probes two extreme situations . in the first one cb electron and bound electrons
are mostly spin polarized in opposite directions .
the center is therefore more likely to capture a cb electron whose spin is oriented in the opposite direction to the majority therefore rising the population of electrons in the cb .
in contrast , in the second situation centers are more likely to capture electrons whose direction is parallel to the majority diminishing the electron population in the cb .
a good estimate of the electron and hole population in either situation is the time resolved sdr ratio given by @xmath221 where the pl intensity under circularly polarized light @xmath222 is proportional to the cb and vb density populations @xmath223 and @xmath224 .
similarly @xmath225 where @xmath226 and @xmath227 are the density populations of cb electrons and holes under linearly polarized light . if cb electrons are captured spin dependently by the ga centers then @xmath228 and @xmath229 and therefore @xmath230 . in accordance with the above considerations , @xmath231 where @xmath232
thus , it is possible to trace the oscillations of bound electrons and nuclei by successively measuring the time resolved @xmath233 for different probe pulse delays . by determining the maxima of the time resolved @xmath233 for different probe pulse delays
we have obtained the plot displayed in fig .
[ figure9 ] setting @xmath234=1 . here
the @xmath233 maxima are plotted as a function of their corresponding time delays @xmath214 as closed circles .
similar results ( not shown here ) are obtained by calculating the @xmath233 from the integrated pl .
below , the spin polarization of bound electrons @xmath217 and nuclear spin polarizarion @xmath216 are shown for reference .
this plots demonstrate that it is possible to trace the coherent oscillations of the spin polarization of bound electrons interacting with the nuclei by means of the time resolved @xmath235 ratio .
we have analyzed the spin dynamics of electrons and nuclei in gaasn by developing a model based on the master equation approach .
the main mechanisms behind the spin dependent recombination are considered as well as the hyperfine interaction in ga paramagnetic traps .
we have demonstrated that the nsr in centers plays an essential role in reproducing the two most significant signatures of the hfi in ga centers .
first , the amplification of the spin filtering effect under a faraday configuration magnetic field is visible only if some nsr mechanism is present .
second , the features of the overhauser - like magnetic field not only depend on the hfi but also strongly rely on the nature of the nsr mechanism .
we have tested the dipolar interaction between neighbouring ga atoms and the quadrupolar interaction of ga centers with random charge distribution background .
we have proven that the dipolar is the only mechanism compatible with the experimental observations . indeed , a scenario where large charge distribution variations are present in the vicinity of the ga nuclei is difficult to imagine .
although most of the experimental results are correctly reproduced by our model some of the aspects regarding the behaviour of the overhauser - like magnetic field remain elusive .
one of these is the discrepancy between the saturation values in the high power regime .
this is important since it would allow to pinpoint the exact origin of the nsr mechanism . to further explore the effects of the hfi and the nsr we have tested the model in the pe regime . in particular ,
we have proposed a pump - probe scheme that allows to trace the coherent oscillations of the bound electron spin interacting with its nucleus through the hfi .
even though in principle this model is conceived for ga centers , it can be easily adapted for other type of centers where dipolar or quadrupolar interactions play an important role as the leading mechanisms of nsr .
french and russian authors acknowledge funding from lia cnrs - ioffe ras ilnacs . l.a.b .
, v.k.k . and e.l.i .
acknowledge the russian foundation for basic research ( grant no .
14 - 02 - 00959 ) .
a.k . gratefully appreciates the financial support of departamento de ciencias bsicas uam - a " grant numbers 2232214 and 2232215 .
j.c.s.s . and v.g.i.s .
would like to acknowledge the support received from the becas de posgrado uam " scholarship numbers 2151800745 and 2112800069 .
we are indebted to professor j. grabinsky for the careful reading of the manuscript .
according to the wangsness , bloch , and redfield relaxation theory @xcite the interaction of a nucleus with its sorroundings can be accounted for by the hamiltonian @xmath236 where @xmath237 is a constant , @xmath238 is a @xmath239-th component of the rank @xmath240 irreducible spherical tensor and @xmath241 is a random function that describes the interaction with the surroundings . to second - order ,
the average fluctuations of the surroundings with the nucleus are given by the following dissipator @xmath242\right]},\ ] ] where @xmath243 is the average density matrix . by substituting the hamiltonian ( [ pert : hamiltonian ] ) in ( [ redflied ] ) and using the fact that @xmath244 and @xmath245 we get the general form of the dissipator @xmath246\right]\\ \times \left(\int_{-\infty}^{t}dt'\gamma^{2 } \overline{f_{k , s}\left(t\right)f_{k , r}^{*}\left(t'\right)}\right ) .
\label{gen : redflied}\end{gathered}\ ] ] the above dissipator describes the interaction of a nucleus with the fluctuations of a random electromagnetic field . the functions @xmath247 and @xmath248 comply with @xmath249 where @xmath250 is a correlation function .
@xmath251 is the correlation time of the fluctuating field and @xmath197 is the correlation amplitude when @xmath252 . with ( [ condition : fi ] ) , the dissipator @xmath253 ( [ gen : redflied ] ) is simplified to @xmath254\right],\ ] ] where @xmath255 is the nsr time . in the case of magnetic dipole interactions ,
only the irreducible spherical tensors of rank @xmath145 participate in the hamiltonian ( [ pert : hamiltonian ] ) .
they can be expressed in terms of the nuclear spin components as @xmath256 and for electric quadrupole interaction the rank is @xmath152 and their components are @xmath257,\label{elecquadrupole:0}\\ \boldsymbol{\hat{t}}_{2,\pm1}=&&\mp\frac{1}{2\sqrt{6 } } \left[\hat{i}_{x}\hat{i}_{z}+\hat{i}_{z}\hat{i}_{x } \pm i\left ( \hat{i}_{y}\hat{i}_{z}+\hat{i}_{z}\hat{i}_{y}\right ) \right]\label{elecquadrupole:1},\\ \boldsymbol{\hat{t}}_{2,\pm 2}= & & \frac{1}{2\sqrt{6}}\left[\left(\hat{i}^{2}_{x}-\hat{i}^{2}_{y}\right ) \pm i\left(\hat{i}_{x}\hat{i}_{y}+\hat{i}_{y}\hat{i}_{x}\right ) \right]\label{elecquadrupole:2}.\end{aligned}\ ] ]
the quadrupolar dissipator for pt is given explicitly by @xmath258 @xmath259 in a short form the above equations can be presented as @xmath260 then the quadrupolar dissipator for ut is given by @xmath261 the dipolar dissipator for pts is given explicitly by @xmath262\:,\nonumber \\ \\ & & \left ( \hat{\cal d}_2\right)_{-\frac12,-\frac32 } = - \frac{2}{3\tau_{n2 } } \left ( 3 \rho_{2;-\frac12,-\frac32 } - \sqrt{3 } \rho_{2;\frac12,-\frac12}\right ) , \\ & & \left ( \hat{\cal d}_2\right)_{\frac12,\frac32 } = - \frac{2}{3\tau_{n2 } } \left ( 3 \rho_{2;\frac12,\frac32 } - \sqrt{3 } \rho_{2;-\frac12,\frac12}\right ) , \\ & & \left ( \hat{\cal d}_2\right)_{-\frac12,\frac12 } = - \frac{2}{3\tau_{n2 } } \left [ 4 \rho_{2;-\frac12,\frac12 } - \sqrt{3 } \left ( \rho_{2;\frac12,\frac32 } + \rho_{2;-\frac32,-\frac12 } \right ) \right],\nonumber \\ \\ & & \left ( \hat{\cal d}_2\right)_{-\frac32,-\frac12 } = - \frac{2}{3\tau_{n2 } } \left ( 3 \rho_{2;-\frac32,-\frac12 } - \sqrt{3 } \rho_{2;-\frac12,\frac12}\right ) , \\ & & \left ( \hat{\cal d}_2\right)_{\frac32,-\frac12 } = - \frac{1}{\tau_{n2 } } \left ( 3 \rho_{2;\frac32,-\frac12 } - \rho_{2;\frac12,-\frac32 } \right)\:,\\ & & \left ( \hat{\cal d}_2\right)_{\frac12,-\frac32 } = - \frac{1}{\tau_{n2 } } \left ( 3 \rho_{2;\frac12,-\frac32 } - \rho_{2;\frac32,-\frac12 } \right),\\ & & \left ( \hat{\cal d}_2\right)_{-\frac32,\frac12 } = - \frac{1}{\tau_{n2 } } \left ( 3 \rho_{2;-\frac32,\frac12 } - \rho_{2;-\frac12,\frac32 } \right)\:,\\ & & \left ( \hat{\cal d}_2\right)_{-\frac12,\frac32 } = - \frac{1}{\tau_{n2 } } \left ( 3 \rho_{2;-\frac12,\frac32 } - \rho_{2;-\frac32,\frac12 } \right),\\ & & \left ( \hat{\cal d}_2\right)_{\frac32,-\frac32 } = - \frac{4}{\tau_{n2 } } \rho_{2;\frac32,-\frac32}\:,\\&&\left ( \hat{\cal d}_2\right)_{-\frac32,\frac32 } = - \frac{4}{\tau_{n2 } } \rho_{2;-\frac32,\frac32 } \:.\end{aligned}\ ] ] 45ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevb.93.035402 [ * * , ( ) ] http://dx.doi.org/10.1038/nature10900 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.041201 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.094303 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.246801 [ * * , ( ) ] link:\doibase 10.1038/30156 [ * * , ( ) ] http://dx.doi.org/10.1038/nature11449 [ * * , ( ) ] link:\doibase 10.1103/physrevb.71.014401 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4867905 [ * * , ( ) , http://dx.doi.org/10.1063/1.4867905 ] link:\doibase 10.1103/physrevb.74.195301 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.115206 [ * * , ( ) ] http://dx.doi.org/10.1038/nature08812 [ * * , ( ) ] link:\doibase 10.1117/2.1201607.006539 [ ( ) , 10.1117/2.1201607.006539 ] link:\doibase 10.1103/physrevlett.105.187602 [ * * , ( ) ] link:\doibase 10.1126/science.1155400 [ * * , ( ) ] , link:\doibase 10.1103/physrevx.6.031014 [ * * , ( ) ] http://dx.doi.org/10.1038/nature12011 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.116601 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.121201 [ * * , ( ) ] link:\doibase 10.1134/1.2142877 [ * * , ( ) ] link:\doibase 10.1134/s0021364007030095 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3275703 [ * * , ( ) , http://dx.doi.org/10.1063/1.3275703 ] \doibase http://dx.doi.org/10.1063/1.3186076 [ * * , ( ) , http://dx.doi.org/10.1063/1.3186076 ] \doibase http://dx.doi.org/10.1063/1.3273393 [ * * , ( ) , http://dx.doi.org/10.1063/1.3273393 ] http://stacks.iop.org/0953-8984/21/i=17/a=174211 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3299015 [ * * , ( ) , http://dx.doi.org/10.1063/1.3299015 ] http://stacks.iop.org/0953-8984/22/i=46/a=465804 [ * * , ( ) ] link:\doibase 10.1103/physrevb.30.931 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0038-1098(74)90202-6 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4816970 [ * * , ( ) , http://dx.doi.org/10.1063/1.4816970 ] link:\doibase 10.1103/physrevb.83.165202 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.035205 [ * * , ( ) ] link:\doibase 10.1134/s0021364012210060 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.125202 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms2776 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.115205 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.205202 [ * * , ( ) ] link:\doibase 10.1103/physrev.89.728 [ * * , ( ) ] in \doibase http://dx.doi.org/10.1016/b978-1-4832-3114-3.50007-6[__ ] , , vol . ,
( , ) pp . link:\doibase 10.1103/physrev.142.179 [ * * , ( ) ] https://books.google.com.mx/books?id=miufce1c9cqc[__ ] , series in chemical physics ( , ) chap .
link:\doibase 10.1103/physrevlett.109.166605 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.89.146601 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1002/pssa.200673009 [ * * , ( ) ] | similar to nitrogen - vacancy centers in diamond and impurity atoms in silicon , interstitial gallium deep paramagnetic centers in gaasn have been proven to have useful characteristics for the development of spintronic devices . among other interesting properties , under circularly polarized light
, gallium centers act as spin filters that dynamically polarize free and bound electrons reaching record spin polarizations ( close to 100% ) .
furthermore , the recent observation of the amplification of the spin filtering effect under a faraday configuration magnetic field has suggested that the hyperfine interaction that couples bound electrons and nuclei permits the optical manipulation of its nuclear spin polarization . even though the mechanisms behind the nuclear spin polarization in gallium centers are fairly well understood ,
the origin of nuclear spin relaxation and the formation of an overhauser - like magnetic field remain elusive . in this work
we develop a model based on the master equation approach to describe the evolution of electronic and nuclear spin polarizations of gallium centers interacting with free electrons and holes .
our results are in good agreement with existing experimental observations .
in particular , we are able to reproduce the amplification of the spin filtering effect under a circularly polarized excitation in a faraday configuration magnetic field . in regard to the nuclear spin relaxation , the roles of nuclear dipolar and quadrupolar interactions are discussed .
our findings show that , besides the hyperfine interaction , the spin relaxation mechanisms are key to understand the amplification of the spin filtering effect and the appearance of the overhauser - like magnetic field . to gain a deeper insight in the interplay of the hyperfine interaction and the relaxation mechanisms ,
we have also performed calculations in the pulsed excitation regime . based on our model s results we propose an experimental protocol based on time resolved spectroscopy .
it consists of a pump - probe photoluminescence scheme that would allow the detection and the tracing of the electron - nucleus flip - flops through time resolved pl measurements . |
due to its structural simplicity , fese has played a leading role in many experimental and theoretical studies on fe - based superconductors since its discovery in 2008 @xcite .
the enduring interest in this compound is partially owed to the high @xmath0 ( ranging between @xmath3 ) achieved when monolayer fese films are grown on srtio@xmath4 substrates @xcite ( fese / sto ) , a ten - fold enhancement from the @xmath5 of bulk fese crystals at ambient pressure @xcite .
intriguingly , the high @xmath0 in the interfacial system proves to be robust for various oxide substrates , including srtio@xmath4 ( 001 ) @xcite , batio@xmath4 ( 001 ) @xcite , srtio@xmath4 ( 110 ) @xcite , anatase tio@xmath6 ( 001 ) @xcite , and rutile tio@xmath6 ( 100 ) @xcite .
these oxide substrates , terminated at tio@xmath6 surface when interfaced with fese , have lattice parameters significantly larger than that of bulk fese and thus apply strong tensile strain on fese thin films .
the anatase and rutile tio@xmath6 substrates even induce rather different strains along @xmath7 and @xmath8 axes of the monolayer fese .
the @xmath0 s , however , are consistently above @xmath9 , as measured by angle - resolved photoemission spectroscopy ( arpes ) .
this observation appears to rule out a direct correlation between the enhanced superconductivity and the tensile strain @xcite .
the electronic structure of the interfaces displaying enhanced @xmath0 s are also remarkably similar across the various substrates .
for instance , the fermi surface measured by arpes consists of only electron pockets at the corners of the two - fe brillouin zone , indicating substantial electron doping from the parent compound .
this observation poses a challenge to theories for the high @xmath0 based on the pairing mediated by spin fluctuations that are strongly enhanced by fermi surface nesting .
one potential solution to this problem is the involvement of bands below the fermi level in pairing ( so - called incipient band pairing ) @xcite .
another possibility is the involvement of a different type of pairing mediator such as nematic fluctuations @xcite or phonons particular to the interface @xcite .
evidence for the latter has been provided by the common observation of replica bands in the electronic structure of superconducting fese monolayers on srtio@xmath4 @xcite , batio@xmath4 @xcite , and rutile tio@xmath6 @xcite .
the replica bands observed by arpes are exact copies of the original bands crossing the fermi level in momentum space but with a weaker spectral weight .
they are interpreted as being generated by an electron - phonon ( @xmath1-ph ) interaction between the fese electrons and oxygen phonons in the substrate @xcite .
this view is supported by the fact that the @xmath10 energy offset between the primary and the replica band coincides with the phonon energy of oxygen modes in srtio@xmath4 @xcite , batio@xmath4 @xcite , and tio@xmath6 @xcite . due to the particular properties of the interface @xcite ,
this interaction is strongly peaked for forward scattering ( i.e. , peaked at small momentum @xmath11 transfer ) , as found by analyzing the electrostatic potential from the dipole induced by the oxygen modes @xcite and by first - principles calculations @xcite .
this unique momentum structure accounts for the fact that the replicas sharply trace the dispersion of the primary band , which requires the @xmath1-ph interactions be forward - focused .
such a coupling can also significantly enhance @xmath0 , due to the linear dependence of @xmath0 on the dimensionless coupling constant @xmath12 @xcite , as opposed to the exponential dependence obtained for the usual bcs case .
for example , assuming a narrow width @xmath13 for the forward scattering peak , some of the current authors found @xmath14 reproduces the measured spectral weight ratio between the replica band and the primary band and at the same time a @xmath15 @xcite . ref .
has obtained similar results after extending this approach to a more realistic band structure .
many aspects of the influence of the @xmath1-ph interactions with strong forward scattering on electronic properties and superconductivity are summarized in refs . .
in comparison , there are no qualitative or quantitative studies of the phononic properties for the problem at hand . here , we have carried out such a study to address two issues .
first , zhang _ et al . _
@xcite recently measured the phonon linewidth of a @xmath16 phonon mode penetrating from the srtio@xmath4 substrate into thin fese films using high resolution electron energy loss spectroscopy ( hreels ) and concluded a mode - specific @xmath1-ph coupling constant @xmath17 .
not only does this echo the discovery of replica bands by the arpes experiments in the same system , but it also calls for a theoretical consideration on the hreels measurements .
doing so would corroborate both the total coupling strength and momentum dependence of the @xmath1-ph coupling in fese / sto system with those inferred from the arpes measurements .
second , when a strong @xmath1-ph coupling is distributed over a subset of wave vectors , one expects tendencies towards charge - density - wave formation that can compete with superconductivity .
such tendencies will manifest themselves as kohn anomalies in the phonon dispersion and broad phonon linewidths .
one can , therefore , address this issue directly by examining the phononic self - energy .
here , we examine the phonon linewidth due to @xmath1-ph interactions with strong forward scattering using the same model adopted in ref . to study the electronic spectral function .
we first describe the details of the model and method in sec .
[ sec : model ] .
next , in sec .
[ sec : analytical ] we give some analytical results for the normal state phonon properties in the _ perfect _ forward scattering limit , where the interaction is treated as a delta function at @xmath18 .
our numerical results for both normal and superconducting state with finite @xmath13 are given in sec .
[ sec : numerical ] . here ,
our results show that the forward focused peak in the @xmath1-ph coupling results in very broad phonon lineshapes . however , in sec .
[ sec : screened_linewidth ] we reintroduce coulomb screening , which subsequently undresses the phonon propagator and suppresses these effects . finally , in sec .
[ sec : summary ] we summarize our results and make some concluding remarks in relation to the hreels experiment of zhang _
et al . _
our model hamiltonian describes a single band model of fese electrons coupled to an optical phonon branch via a momentum - dependent coupling , which reads @xmath19 here , @xmath20 ( @xmath21 ) creates ( annihilates ) an electron with wavevector @xmath22 and spin @xmath23 , @xmath24 ( @xmath25 ) creates ( annihilates ) a phonon with wavevector @xmath2 ; @xmath26 is the electronic band dispersion measured relative to the chemical potential @xmath27 ; @xmath28 is the phonon dispersion ( @xmath29 ) ; and @xmath30 is the momentum dependent @xmath1-ph coupling .
we take a simple electronic band dispersion @xmath31 - \mu$ ] , where @xmath7 is the in - plane lattice constant .
we set @xmath32 and @xmath33 , which produces around @xmath34 point an electronlike fermi pocket with @xmath35 , a fermi velocity @xmath36 along the @xmath37 line , and an effective electron band mass @xmath38 , which is similar to the electron pocket at the @xmath39 point in fese / sto seen in arpes experiments @xcite .
since we have a single band model , it only takes a trivial @xmath40 shift to map our @xmath34-point pocket onto the electron pocket in the real system centered at the @xmath39 point and any physical quantities depending only on the momentum transfer @xmath41 , such as phonon linewidth , do not depend on the position of the pocket .
since we are not considering the effects of an unconventional pairing mechanism here , we do not need to consider the possibility of @xmath42-wave instabilities due to scattering between the electron pockets .
as such , a single band model is sufficient for our purpose .
throughout we approximate the experimental phonon dispersion with a dispersionless einstein mode @xmath43 according to the observed energy separation between the replica band and the primary band @xcite , as well as the phonon dispersion of the interface , as measured by hreels @xcite .
we neglect the fermion momentum dependence in the coupling @xmath44 , where @xmath2 is the momentum transfer and adopt @xmath45 as derived from simple microscopic model @xcite . here
, @xmath46 is adjusted to fix the total dimensionless coupling strength of the interaction and @xmath13 sets the range of the interaction in momentum space .
the normalization factor @xmath47 is chosen such that @xmath48 for @xmath49 , where @xmath50 denotes an momentum integral over the first brillouin zone .
we will typically set the in - plane lattice constant @xmath51 below ; however , we will occasionally write it out for clarity . the values of @xmath28 and @xmath52 we use in the calculation include all the screening effect within the oxide substrate , but _
none _ from the fese film .
thus , we refer to them as the `` bare '' or `` unscreened '' quantities . in sec . [
sec : screened_linewidth ] , we show that such a treatment is justified in calculating the electron self - energy using the `` unscreened '' phonon propagator _ and _ the `` unscreened '' coupling @xmath52 , but it overestimates the phonon self - energy , especially the imaginary part ( phonon linewidth ) at @xmath53 , by overlooking the strong screening effect at @xmath53 from the fese film . the difference between the fully screened phonon frequency @xmath54 ( by both the substrate and the fese film ) and partially screened @xmath28 ( only by the substrate itself ) is small , however , so we do not distinguish them ( @xmath55 ) in sections [ sec : model ] , [ sec : analytical ] , and [ sec : numerical ] .
our calculation in sec .
[ sec : screened_linewidth ] shows that the difference is within 10% for most parameters .
the experimental measurements in ref . on phonon frequency in srtio@xmath4 with and without fese deposited also support this conclusion .
the electron and phonon self - energies due to @xmath1-ph interaction are calculated using migdal - eliashberg theory , where the vertex part @xmath56 is approximated with the zeroth order vertex function @xmath52 . here , @xmath57 ( @xmath58 ) is the fermionic ( bosonic ) matsubara frequency .
this is shown in fig .
[ fig : diagrams ] . as discussed in ref .
, in the forward scattering limit the vertex corrections are of order @xmath12 , and can thus be neglected in the weak coupling regime @xmath59 considered here .
( here , @xmath12 measures the fermi surface average of the mass enhancement due to the @xmath1-ph interaction , see ref . . )
note that the vertex correction is independent of the adiabatic parameter @xmath60 , in contrast to the standard migdal s approximation for @xmath61 .
( the vertex correction is always proportional to @xmath12 for either @xmath61 or @xmath62 @xcite , so our argument also applies for the forward - focused @xmath1-ph interaction .
) there are alternative treatments that do not make use of migdal s approximation @xcite in the nonadiabatic regime for momentum independent interaction @xmath63 .
these approaches are beyond the scope of this work , which instead focuses on a momentum dependent interaction .
furthermore , we calculate the dressed electron green s function ( electron propagator ) from the self - energy using the bare phonon green s function ( phonon propagator ) [ see fig . [
fig : diagrams](a ) ] and then insert this into the bubble diagram for the phonon self - energy [ see fig .
[ fig : diagrams](b ) ] .
this approach is the so - called `` unrenormalized migdal - eliashberg '' scheme @xcite , where the phonon self - energy is not fed back into the electron self - energy self - consistently . as we will show in section
[ sec : screened_linewidth ] , this treatment is justified when one includes the coulomb screening of the @xmath1-ph interaction in the problem . adopting nambu s 2-spinor scheme , the electron self - energy @xmath64\hat{\tau}_0 + \chi({{\mathbf{k}}},{{\rm i}}\omega_n)\hat{\tau}_3 + \phi({{\mathbf{k}}},{{\rm i}}\omega_n)\hat{\tau}_1 $ ] and the dressed electron green s function @xmath65 are matrices in nambu space with @xmath66 being the pauli matrices ; @xmath67 are fermionic matsubara frequencies with @xmath68 the inverse temperature ( @xmath69 ) ; @xmath70 and @xmath71 renormalize the single - particle mass and band dispersion , respectively ; and @xmath72 is the anomalous self - energy .
the electron self - energy is self - consistently calculated from the one - loop diagram in fig .
[ fig : diagrams](a ) as follows @xmath73 , \label{eq : elselfe}\end{aligned}\ ] ] where @xmath74 is the `` bare '' phonon propagator .
once we obtain the electron green s function self - consistently , the polarization bubble in fig .
[ fig : diagrams](b ) is given by @xmath75\label{eq : epoltr},\end{aligned}\ ] ] and @xmath76 is the phonon self - energy and @xmath77 is the phonon linewidth , which has been analytically continued to the real frequency axis . to perform the analytic continuation we use the spectral representation of the dressed green s function @xmath78 \notag\\ & \quad\quad \frac{1}{n}\sum_{{\mathbf{k}}}{\operatorname{tr}}\left [ \hat{\tau}_3\hat{a}({{\mathbf{k}}},\omega ' - \omega ) \hat{\tau}_3\hat{a}({{\mathbf{k}}}+{{\mathbf{q}}},\omega ' ) \right ] \bigg\},\end{aligned}\ ] ] where @xmath79 is the fermi - dirac distribution function and @xmath80 @xmath81 is obtained by the same iterative analytic continuation method @xcite we used in ref . .
finally , we find the dressed phonon propagator using @xmath82 and phonon spectral function @xmath83 in the numerical calculations , we solve the electron self - energy self - consistently on a @xmath84 @xmath85-grid .
the convergence for the self - energy is reached if the difference of the self - energies from two consecutive iterations is less than @xmath86 .
the small imaginary part included in the iterative analytic continuation is @xmath87 .
we begin by examining the perfect forward scattering limit , where several analytical results can be obtained . here
, we consider only the normal state in the low - temperature limit ( @xmath88 ) , because many qualitative features of the phonon linewidth are already manifested there . for a normal metal with a parabolic band @xmath89 , i.e. , electron gas in three - dimensions ( 3d ) , the analytical result of eq .
( [ eq : epoltr ] ) is the lindhard function @xcite .
the corresponding result for electron gas in two - dimension ( 2d ) is given in refs . .
without the @xmath1-ph interaction , we can apply the 2d electron gas result to our single band model , due to the small size of the fermi pocket from the band dispersion @xmath90 - \mu \approx \frac{k^2}{2m^{*}}-e_\text{f}$ ] , where @xmath91 , @xmath92 , @xmath93 , and @xmath94 .
this approximate band dispersion is exact at the band bottom and suitable for small @xmath85 . with this approximation ,
the imaginary part of the electron polarization without @xmath1-ph interaction is @xmath95 , \label{eq : epol0}\end{aligned}\ ] ] where @xmath96 , @xmath97 , @xmath98 is density of states of two spins , and the step - function @xmath99 for @xmath100 and @xmath101 for @xmath102 . with the inclusion of the @xmath1-ph interaction , the self - energy in eq .
( [ eq : elselfe ] ) is nonzero but diagonal in the normal state . in the perfect forward scattering limit @xmath103 , where @xmath104 .
the @xmath105-element of the self - energy is then given by @xcite @xmath106 where @xmath107 and @xmath108 .
using this self - energy and dyson s equation , we find that at low temperatures ( @xmath109 and @xmath110 ) , the dressed green s function acquires a two - pole form @xmath111 where @xmath112 and @xmath113 . here , `` m '' and `` r '' denote the main and replica band , respectively .
to simplify the calculation , we shift the two bands by the same energy @xmath114 ( which is small if @xmath115 ) , and the dressed green s function becomes @xmath116 where the shifted @xmath117 and @xmath118 . here , @xmath119 .
physically , eq .
( [ eq : greenmrb ] ) clearly indicates that the replica band exactly follows the dispersion of the main band , and its energy offset from the main band is @xmath120 ( @xmath121 ) for the part of the main band above ( below ) the fermi level . without @xmath1-ph interaction ( a ) and with forward scattering @xmath1-ph interaction ( b ) .
@xmath122 is used in panel ( b ) . the parabolic band approximation for fese / sto model @xmath123
is assumed , so @xmath124 is isotropic in momentum space .
@xmath125 , @xmath126 , @xmath127 , and @xmath128 . ] using eq .
( [ eq : greenmrb ] ) , the imaginary part of the electron polarization with the @xmath1-ph interaction in perfect forward scattering limit can be expressed in terms of the noninteracting electron polarization as follows @xmath129 here , @xmath130 is the sign of @xmath131 .
equation ( [ eq : epol ] ) is also a good approximation when the coupling function @xmath132 has a sharp peak ( @xmath133 ) .
then , the phonon linewidth is given by @xmath134 . in fig .
[ fig : impolt0 ] we show @xmath135 calculated from eq .
( [ eq : epol0 ] ) and from eq .
( [ eq : epol ] ) in panel ( a ) and ( b ) , respectively . fig .
[ fig : impolt0](a ) manifests the electron - hole continuum for 2d electron gas at low temperature , while fig .
[ fig : impolt0](b ) shows multiple scattering processes at low temperature corresponding to the three terms in eq .
( [ eq : epol ] ) : one within the main band for @xmath136 that represents the original electron - hole continuum , one between the main and replica band for @xmath137 , and one within the replica band for @xmath138 , in a descending order of weights ( @xmath139 , @xmath140 , and @xmath141 ) . as shown in fig .
[ fig : impolt0](b ) , at the fixed frequency @xmath142 , the magnitude of the imaginary part of the electron polarization has a sharp upturn at a finite momentum , leading to a peak that slowly decreases at larger momentum .
this qualitative feature persists in the full numerical result in the next section .
note that since the coupling constant is a delta function , the phonon linewidth @xmath143 is zero at all @xmath144 values despite the fact that the polarization @xmath145 is nonzero .
we now turn to the polarization and phonon linewidth for the case of an @xmath1-ph interaction with a small but nonzero width in momentum space .
figure [ fig : imppht ] shows the imaginary part of the electron polarization @xmath135 and the phonon linewidth @xmath146 for various temperatures . here
, we have parameterized the total @xmath1-ph coupling using the double fermi - surface averaged definition @xmath147 where @xmath148 is the density of states per spin and @xmath149 .
we have used this definition because the @xmath2-averaged @xmath150 equals the mass enhancement factor @xmath151 only in the limit of perfect forward scattering , while @xmath152 approximates the mass enhancement factor when the @xmath1-ph interaction is more uniform .
the latter case occurs for the larger values of @xmath13 used in fig .
[ fig : imppht ] .
in addition , @xmath152 as defined in eq .
( [ eq : lambda ] ) does not depend on temperature where as @xmath12 does .
empirically , we find @xmath153 ( see ref . for the proportionality constant ) , which can be used to approximately convert between the two definitions .
fig : imppht ] we have set @xmath154 , which is equivalent to @xmath122 for @xmath155 and within the suitable range of values that simultaneously fit both high @xmath0 value and the measured spectral weight of the replica bands @xcite . at low temperature and @xmath156 , the imaginary part of the polarization in fig .
[ fig : imppht](a ) has a peak appearing at @xmath157 , which is a feature of the electron - hole continuum ; with increasing temperature , the @xmath158 increases , and the rate of increase is faster for smaller values of @xmath13 .
the phonon linewidth , shown in fig .
[ fig : imppht](b ) , strongly peaks at @xmath53 for @xmath159 because the forward scattering coupling @xmath52 strongly suppresses the peak in the polarizability appearing at the finite @xmath11 . as the value of @xmath13 increases , however , the width of @xmath52 begins to overlap with the peak in the polarization , and a corresponding peak in the linewidth recovers at nonzero @xmath2 . in this case
, both the temperature and the width of the coupling function @xmath52 dictate the full @xmath2 dependence of the phonon linewidth .
thus , due to its sensitivity to these parameters , the momentum dependence of the phonon linewidth can be used to determine not only the overall strength of the @xmath1-ph interaction but also the width of the coupling function . for a fixed frequency @xmath142 and various momentum width parameters
@xmath159 , @xmath160 , @xmath161 , and @xmath162 in the @xmath1-ph coupling function @xmath163 .
the double fermi - surface averaged coupling constant ( defined in the text ) is fixed at @xmath164 .
the colors ( gray scales ) of lines represent low ( blue ) to high ( red ) temperatures .
( b ) similar to ( a ) but for phonon linewidth @xmath165 . ] to reproduce the replica bands observed in the arpes experiments , the width of the @xmath1-ph coupling must be narrow in momentum space with @xmath166 . based on this observation , and the results shown in fig .
[ fig : imppht ] , one might expect that the phonon linewidth in the vicinity of @xmath167 should be very large . in turn
, the real part of the phonon self - energy will also develop significant kohn anomaly , leading to an instability of the lattice .
it turns out that the coulomb interaction will prevent this from occurring , as the divergence in the coulomb interaction at @xmath167 effectively blocks the long - wavelength instability .
we will discuss this issue in the next section .
in this section we examine the effects of coulomb screening by the fese electrons on the @xmath1-ph vertex and the phonon linewidth .
[ fig : diagrams](c ) shows the diagramatic expansion of the _ screened _ @xmath1-ph vertex evaluated at the level of the random phase approximation .
the screened vertex is @xmath168 \notag \\ & \quad + g({{\mathbf{q}}})\left[-v_c({{\mathbf{q}}})\chi_0({{\mathbf{q}}},{{\rm i}}\omega_\nu)\right]^2+\dots \notag \\ & = \frac{g({{\mathbf{q}}})}{1+v_c({{\mathbf{q}}})\chi_0({{\mathbf{q}}},{{\rm i}}\omega_\nu)},\end{aligned}\ ] ] where @xmath169 is the charge susceptibility and @xmath170 is the fourier transform of the coulomb potential . in the contiuum limit ,
@xmath171 in three dimensions ( 3d ) and @xmath172 in two dimensions ( 2d ) .
the corresponding phonon self - energy is obtained by replacing the vertex function with the screened vertex with @xmath173^ * \left[-\chi_0({{\mathbf{q}}},{{\rm i}}\omega_\nu ) \right ] \notag \\ & = \frac{-|g({{\mathbf{q}}})|^2 \chi_0({{\mathbf{q}}},{{\rm i}}\omega_\nu ) } { 1+v_c({{\mathbf{q}}})\chi_0({{\mathbf{q}}},{{\rm i}}\omega_\nu)},\end{aligned}\ ] ] where we have assumed @xmath174 is real . here , we are interested in the case of an fese monolayer located a distance @xmath175 above the oxide substrate .
we place the fese electrons at @xmath176 and the ions in the termination layer of the substrate at @xmath177 . for this geometry
, we introduce an anisotropic coulomb potential @xcite @xmath178 where @xmath179 is the momentum transfer a plane parallel to the fese monolayer , and @xmath180 and @xmath181 are the zero - frequency dielectric constants parallel and perpendicular to the plane . by inverse fourier transform ,
the real space formula is @xmath182 where @xmath183 and @xmath184 . after performing the 2d fourier transforming for the in - plane
coordinates we arrive at @xmath185 to compute the screened @xmath1-ph interaction , we must use the interaction at @xmath176 for the coulomb potential since the particle - hole pairs are created in the fese layer .
putting this all together , the phonon linewidth is given by @xmath186 where we define the unscreened " phonon energy as @xmath187 ^ 2 } - { \operatorname{re}}\pi({{\mathbf{q}}},\omega)$ ] .
along a high symmetry path @xmath39-@xmath34-@xmath188-@xmath39 at @xmath189 k. results are shown for ( a ) @xmath190 and ( b ) @xmath191 , @xmath192 .
the line color ( gray scale ) encodes the different values of @xmath193 , @xmath194 , @xmath195 , and @xmath196 , as indicated by the color bar . ]
we evaluated eq .
( [ eq : screened_gamma ] ) for several values of @xmath13 and @xmath152 , and the results are shown in fig . [
fig : screened_gamma ] . since the exact values of the dielectric constants are not known for the fese interface systems , we show results for @xmath197 in fig .
[ fig : screened_gamma](a ) and @xmath191 , @xmath192 in fig .
[ fig : screened_gamma](b ) .
note that the latter values are close to the estimates obtained by kuli and dolgov ( ref . ) in the limit of perfect forward scattering . in both cases , we find that the phonon linewidth is dramatically suppressed once coulomb screening is included ; however , as the values of @xmath180 and @xmath181 are increased , the magnitude of the linewidth increases .
these results indicate that the long - range coulomb interaction prevents the formation of a competing charge ordering at long - wavelengths , which is consistent with the notion that extend coulomb interactions can suppress insulating behavior @xcite .
our results also show that this effect will be somewhat senstive to the dielectric properties of the interface , which may offer a means to tune these properties .
finally , the undressing of the phonon linewidth observed here also provides a rationale for adopting an unrenormalized migdal - eliashberg scheme , where the phonon self - energy is not fed back into the electron self - energy in a self - consistent manner . in this case , the calculated phonon self - energy is small , justifying the use of the bare phonon propagator in the electron self - energy diagrams .
comparing our results to the recent rheels measurements by zhang _
et al . _
@xcite , we find that once the coulomb screening is included , the computed linewidths are much smaller than those inferred experimentally . moreover , in the experimental data , the linewidth is finite at @xmath34 point and maximal around @xmath188 point .
our calculated linewidth is exactly zero at @xmath34 point because the screening from the coulomb potential diverges at @xmath167 .
however , we have not considered any impurity potential in our calculation , or other sources of broadening in the electron green s function , and , subsequently , the phonon linewidth once the charge susceptibility @xmath198 is computed .
regardless , @xmath188 point is not the maximal point for the linewidth in any of our calculation results .
this discrepancy could also be due to the limitation of our single band model .
the real system is multiband in nature and also shows strong magnetic fluctuations .
in this paper , we have calculated the phonon linewidth , i.e. , the imaginary part of the phonon self - energy in an unrenormalized migdal - eliashberg scheme in the weak to intermediate coupling regime for strong forward scattering @xmath1-ph interaction .
such an @xmath1-ph interaction dresses the electron propagator by simply creating the replica bands and shuffles the electron - hole continuum of 2d electron gas into three similar parts with descending weights beginning at @xmath136 , @xmath199 , and @xmath200 .
if we do not include coulomb screening , the phonon linewidth is a simple product of coupling function @xmath201 with a forward scattering peak around @xmath53 and the electron polarization with a very similar momentum structure of the electron - hole continuum of 2d electron gas .
depending on the peak width @xmath13 of the @xmath1-ph coupling constant @xmath52 and the peak of electron polarization at @xmath157 , we find the linewidth @xmath202 has a maximum value at @xmath53 or @xmath203 at low temperature , and the linewidth is broad at @xmath167 .
even if the latter happens , since the linewidth for small @xmath11 tends to increase with temperature , the maximum may shift back to @xmath53 at high temperature .
the momentum resolved phonon spectral function at @xmath204 can be understood in the same picture .
the broad linewidths at @xmath167 would normally indicate an instability to a charge - ordered phase at long wavelengths .
however , once the long - range coulomb interaction screens the @xmath1-ph interaction we find that the phonons are undressed . here , the anomalous broadening at @xmath167 is suppressed by the divergence in the coulomb interaction at @xmath167 while the total phonon linewidth is reduced throughout the brillioun zone . in this case ,
a small peak remains at nonzero momentum transfers ; however , the magnitude of this peak is much smaller than the linewidths measured by hreels @xcite .
our results suggest that the broadening of the srtio@xmath4 phonons ( with a maximum at @xmath188 point ) observed by zhang _ et al . _
are not due to the forward - focused @xmath1-ph coupling inferred from the arpes measurements and are likely from some other source . to resolve the forward - focused @xmath1-ph interaction ,
the hreels experiments should focus on smaller values of @xmath2 , which will be challenging given the large background signal at @xmath167 .
we thank t. berlijn , t. p. devereaux , m. l. kuli , e. w. plummer , and d. j. scalapino for useful discussions .
l. r. acknowledges funding from the dutch science foundation ( nwo ) via a rubicon fellowship and by the national science foundation under grant no .
phy11 - 25915 and grant no .
nsf - kitp-17 - 019 .
s. j. is funded by the university of tennessee s office of research and engagement s organized research unit program .
y. w. and e. d. are supported by the u.s .
department of energy , office of basic energy sciences , materials sciences and engineering division .
this research used computational resources supported by the university of tennessee and oak ridge national laboratory s joint institute for computational sciences .
37ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1073/pnas.0807325105 [ * * , ( ) ] http://stacks.iop.org/0256-307x/29/i=3/a=037402 [ * * , ( ) ] link:\doibase 10.1038/nmat4153 [ * * , ( ) ] link:\doibase 10.1038/nature13894 [ * * , ( ) ] link:\doibase 10.1038/ncomms6044 [ * * , ( ) ] link:\doibase 10.1103/physrevb.94.104510 [ * * , ( ) ] link:\doibase 10.1063/1.4950964 [ * * ( ) , 10.1063/1.4950964 ] @noop ( ) link:\doibase 10.1103/physrevlett.117.067001 [ * * , ( ) ] @noop ( ) http://stacks.iop.org/1367-2630/16/i=2/a=023029 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.224514 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.117.077003 [ * * , ( ) ] link:\doibase 10.1038/srep32078 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.117.217003 [ * * , ( ) ] http://stacks.iop.org/1367-2630/17/i=7/a=073027 [ * * , ( ) ] http://stacks.iop.org/1674-1056/24/i=11/a=117405 [ * * , ( ) ] http://stacks.iop.org/1367-2630/18/i=2/a=022001
[ * * , ( ) ] link:\doibase 10.1063/1.4876750 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.134513 [ * * , ( ) ] http://stacks.iop.org/1367-2630/19/i=1/a=013020 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.224506 [ * * , ( ) ] @noop ( ) http://stacks.iop.org/0953-2048/29/i=5/a=054009 [ * * , ( ) ] link:\doibase 10.1103/physrevb.94.081116 [ * * , ( ) ] link:\doibase 10.1038/ncomms1946 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.93.054517 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.174516 [ * * , ( ) ] link:\doibase 10.1103/physrevb.94.024505 [ * * , ( ) ] link:\doibase 10.1103/physrevb.42.2416 [ * * , ( ) ] link:\doibase 10.1103/physrevb.37.4965 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physrevlett.18.546 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.74.045124 [ * * , ( ) ] link:\doibase 10.1103/physrevb.56.r1645 [ * * , ( ) ] | the discovery of an enhanced superconducting transition temperature @xmath0 in monolayers of fese grown on several oxide substrates has opened a new route to high-@xmath0 superconductivity through interface engineering .
one proposal for the origin of the observed enhancement is an electron - phonon ( @xmath1-ph ) interaction across the interface that peaked at small momentum transfers . in this paper
, we examine the implications of such a coupling on the phononic properties of the system .
we show that a strong forward scattering leads to a sizable broadening of phonon lineshape , which may result in charge instabilities at long - wavelengths . however , we further find that the inclusion of coulombic screening significantly reduces the phonon broadening .
our results show that one might not expect anomalously broad phonon linewidths in the fese interface systems , despite the fact that the @xmath1-ph interaction has a strong peak in the forward scattering ( small @xmath2 ) direction .
= 1 |
it is now widely accepted that quasi - stellar objects ( qsos ) are powered by accretion onto a massive black hole ( mbh ) within a galaxy .
while mbhs are commonly found in galaxies , qsos require the central black hole to be fed by a steady supply of matter .
characterization of the host galaxies of qsos can provide important clues to understand both the formation mechanisms and the evolution of qsos .
observations of low - z qsos offer the obvious advantage of increased resolution and sensitivity for accurately measuring host properties , and relatively large samples are required to both adequately probe the full range of host properties and to explore possible relationships between host type and properties of the qso nucleus .
the pioneering optical surveys of low - z qsos in the mid 1980 s , using 4m - class telescopes @xcite , produced several common themes , as well as enigmatic problems that still remain today .
qsos were found typically to live in luminous ( @xmath16 ) hosts , with a mixture of host types ( spirals , barred - spirals , ellipticals ) , including a significant fraction of disturbed and/or strongly - interacting systems .
there was also some hint of a trend in host type with host luminosity , with spirals and luminous ellipticals favored at the low and high end , respectively , of the range of qso luminosities .
however , it often proved difficult to distinguish reliably between different host types , often due to poor sensitivity or angular resolution or both ; in addition , the mixture of objects being observed was abnormally weighted toward radio - loud and/or x - ray selected objects despite these objects being only a small fraction of the predominant radio - quiet population .
optical qso surveys with _
@xcite offered dramatically improved resolution , but for the most part contained similar biases toward radio - loud objects , resulting in a preponderance of ellipticals . in the end
, results from the _ hst _ surveys did not significantly modify earlier conclusions based on the ground - based surveys .
more recent surveys of qso hosts have favored the near - infrared , where the fraction of the observed light due to the underlying host galaxy is largest , and ground - based seeing can rival that from _ hst_. important large near - infrared qso surveys include those from the ground with 2/4m - class telescopes @xcite and , more recently , those exploiting adaptive optics ( ao ) @xcite , and , finally , those using _
@xcite . of most relevance to our own current work are the near - infrared observations by mcleod and co - workers who have imaged the host galaxies for a large volume - limited sample of optically - selected qsos from the pgbqs similar to the sample and selection criteria we have used .
other near - infrared surveys of qsos include studies with nearly equal mixtures of radio - quiet and radio - loud qsos ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and studies of far - infrared - excess " qsos ( e.g. * ? ? ?
* ; * ? ? ?
all of the near - infrared surveys generally agree with previous optical surveys in finding that qsos typically reside in luminous ( i.e. , @xmath16 ) hosts , and that spirals represent an increasing fraction of hosts at lower qso luminosities .
however , the near - infrared surveys still emphasize the overall diversity of host types , and they are still plagued by difficulties with the accurate determination of radial profiles due to a combination of problems including bad seeing , unstable psf , and ultimately the faintness of the hosts at large radii . finally , it has now become clear that optically - selected qsos may be rivaled in their numbers by newly discovered populations of near - infrared selected @xcite , and far - infrared selected @xcite qsos .
it is extremely intriguing that a very high fraction ( @xmath17% ) of the former show evidence of tidal interactions @xcite , and nearly all of the latter are found to be major mergers of large gas - rich galaxies ( e.g. , @xcite ) .
these new findings raise the obvious questions of possible evolutionary connections between these different classes of objects , all of which seem to be capable of producing luminous qsos .
it thus became clear to us that it might prove extremely useful to revisit the properties of optically - selected qso hosts and to search for trends in host type with other properties such as far - infrared luminosity and color , which might then prove useful in future studies designed to explore possible connections between the familiar optically - selected qso population and the various types of infrared - selected qsos .
the availability of ao systems on 810 m class telescopes provides a potentially important new capability for accurately measuring the distribution of light in the host galaxy of qsos .
in particular , our access to relatively large amounts of observing time using the new university of hawaii near - infrared ao system , hokupaa , on mauna kea , allowed us to carefully design a sensitive and relatively uniform survey , primarily at _
h_-band , of a large representative sample of low - z qsos randomly selected from the optical ( _ b_-band ) pgbqs .
the sample was made large enough to include a representative sub - sample of the minority population of radio - loud objects , as well as to allow for adequate sampling of a wide range of qso luminosities , but was otherwise unbiased .
an additional bonus for our new observations was the existence by 2003 of relatively complete and accurate spectral - energy - distributions ( seds ) from radio - to - xray wavelengths @xcite , including most importantly , new far - infrared fluxes for most of our targets , which were recently determined from reprocessed _ iras _ and _ iso _ data @xcite . in this paper
we present the first near - infrared imaging study of a large number of qsos ( 32 objects ) made with ao systems on the gemini 8.2 m and subaru 8.1 m telescopes on mauna kea .
we first describe , in [ sec : sample ] how our sample was selected . the observations and data reduction techniques
are detailed in
[ sec : observations ] .
the qso host galaxies are characterized through 2-d fits and visual analysis of the images in [ sec : analysis ] .
section [ sec : adopthostclass ] presents our adopted classification scheme , including descriptions of the individual objects and comparisons to the host types derived from previous imaging studies , and summarizes the host fit parameters .
section [ sec : discussion ] discusses how host type depends on other properties of the qso ( e.g. degree of disturbance , radio loudness , the near - infrared absolute luminosity of the host , and the total far - infrared luminosity ) , and compares our results with previous studies .
our main conclusions are summarized in [ sec : conclusions ] . as background to our current work ,
good brief reviews of previous problems and subtleties associated with interpreting the host types of low - z qsos can be found in @xcite and @xcite .
in addition , a more thorough review of the wide range of qso host studies their consensus and differences including theories as to the origin and evolution of qsos and their possible relation to other types of extragalactic objects , can be found in the proceedings of three recent workshops specifically devoted to the study of qso hosts @xcite .
the performance of ao systems is a strong function of the photon flux seen by the wavefront sensor ( wfs ) .
although the hokupaa ( the institute for astronomy ao system @xcite in use on the gemini north telescope from mid-2000 to mid-2002 ) and subaru ao @xcite systems can `` lock '' onto objects as faint as @xmath18 under good seeing , only very partial wavefront correction can be achieved for such faint objects .
to ensure reasonably good image quality ( fwhm@xmath19 ) , our sample was limited to bright qsos ( approximately @xmath20 ) .
this observational constraint would result in a strong bias towards intrinsically highly luminous qsos if the sample were not redshift - limited .
moreover , observation of higher redshift qsos is made more difficult for the purpose of this work by the decrease in spatial resolution , and especially surface brightness , with increasing distance .
our sample is hence limited to @xmath21 .
the ideal sample for our current observations proved to be the palomar - green bright quasar sample ( bqs : * ? ? ? * ) .
although there are more recent and larger qso surveys , the bqs still contains the majority of the nearest and brightest optically - selected , uv - excess qsos for study from the northern sky , and its magnitude limit was close to our desired brightness limit for good ao performance .
there are also extensive multi - wavelength observations in the literature for the bqs sample , which potentially allows us to further explore the relationship of the qso host properties to other properties of the qso .
the bqs contains 92 qsos with @xmath22 ( in the @xmath23 cosmology used in * ? ? ?
* ) , in the 10,714 deg@xmath24 surveyed region of the sky at @xmath25 and @xmath26 .
forty - seven of these qsos met our final redshift and brightness selection criteria ( @xmath27 , @xmath28 ) , of which we were eventually able to observe 58% ( 28/47 ) during our two year observing campaign on mauna kea .
these 28 qsos from the bqs were essentially selected at random from the complete sub - sample of 47 , depending only of the ra range observable during each observing run .
four additional objects were eventually added to the list of 28 , bringing the total number of observed targets to 32 .
two objects from the bqs with @xmath29 ( pg1119 + 120 and pg 1126 - 041 ) were successfully observed during one of the earliest observing runs , while we were still defining the scope of the overall ao campaign . rather than discard the data , we decided to include these objects here for two reasons .
first they allow us to further explore the low luminosity end of the qso sample , and second , we noted early on that the uncertainties in the @xmath10-band magnitudes listed in the original bqs are large enough such that these two objects might have properly been in our sample .
two additional objects were added to our observing list in the latter half of our overall campaign , when it was realized that we had only two `` bona - fide '' radio - loud qsos ( rlqs ) in our list of observed objects .
one of these rlqs , 4c31.63 ( @xmath30 ) , was at the time it was added , the highest redshift object observed , but still within our planned redshift limit for the survey , and its @xmath10-band magnitude was above the average overall threshold for the bqs .
the second rlq , pg2251 + 113 ( @xmath31 ) , was above our stated redshift cutoff , but satisfied our @xmath10-band limit and was the only rlq from the bqs that could be observed at the ra range available for observations during 2001 , september .
the addition of these two rlqs allowed us to bring the total fraction of rlqs in our observed sample ( @xmath32 ) up to the ratios more typically found ( i.e. , @xmath215% ) in large qso surveys .
the complete list of our 32 observed targets is given in table 1 ( vega magnitudes are used in this table and throughout this paper ) .
the listed @xmath10-band magnitudes have been updated where possible with newer more accurate values than originally given in @xcite .
the absolute magnitudes are also updated to an @xmath33 , cosmology : the @xmath34 qso criteria used by @xcite now corresponds to @xmath35 .
we note that the updated @xmath10-band magnitudes resulted in three additional objects , pg0838 + 770 , pg0844 + 349 and pg1229 + 204 , falling below the absolute magnitude cutoff , bringing the total number of lower luminosity objects to five .
both the hokupaa ao system , built by the university of hawaii and used on the gemini north telescope , and the subaru telescope ao system , were used for this project . for both systems ,
a 36-element curvature sensor is used with a 36-element bimorph mirror .
curvature ao systems are traditionally very efficient on faint targets , partly thanks to the use of photon - counting devices ( avalanche photo diodes ) with no readout noise .
this allows these systems to guide on @xmath36 sources and to routinely provide good correction ( better than 02 ) on @xmath37 point sources , which makes them well suited for qso host galaxy imaging .
the ao - corrected beam was fed to quirc for gemini and to ircs for subaru .
both cameras use @xmath38 detectors for near - ir imaging .
the contrast between the qso and a typical host galaxy is better ( lower ) at about 1 @xmath39 m ( in the rest frame ) than either in the visible or the longer infrared wavelengths .
this is largely due to a local minimum near 1 @xmath39 m in the rest - frame radio - to - xray spectral energy distributions ( seds ) of nearly all qsos , both rlqs and rqqs , except for the strongly beamed blazars e.g . ,
. 2 of @xcite , and also fig . 1 of @xcite .
moreover , the performance of ao systems is optimal in the near infrared : at shorter wavelengths , the number of actuators is too small to sample the wavefront properly . at longer wavelengths ,
the high thermal background precludes the observation of low - surface - brightness sources , and the diffraction limit of the telescope restricts the achievable angular resolution .
observations in the near - ir are therefore well - suited for qso host studies from the ground .
the seeing was constantly estimated during the observations by image quality monitoring and monitoring of the control signals of the ao system .
most of the observations were made using the @xmath0 filter , which provides the optimal resolution under good seeing conditions . when the seeing was bad ( corrected fwhm in @xmath0 band larger than 02 ) , the @xmath40 filter was preferred to the @xmath0 filter , and usually allowed high resolution imaging ( fwhm@xmath19 ) even under worse than average seeing , but at the expense of a stronger thermal background .
a few observations were done in the @xmath41 band on exceptionally good nights .
fainter qsos were preferentially observed under good seeing conditions , to avoid running out of bright targets when the seeing was average or worse than average .
these precautions resulted in a surprisingly homogeneous set of images ( angular resolution about 015 ) considering the wide range of seeing conditions and qso magnitudes of the sample .
table 2 gives the complete summary of our observing log for all of the data that are presented in this paper , including filter , exposure times , seeing ( fwhm ) , and telescope used .
twenty - six objects were eventually observed at @xmath0 , and 8 of these 26 were also observed at either @xmath41 or @xmath40 , or both .
four objects were observed at @xmath40 only , one at @xmath41 only , and one at both @xmath41 and @xmath40 .
twenty - eight objects were observed at gemini , and four objects were observed at the end of our campaign ( feb , 2003 ) using subaru . for each target , a list of about six psf candidates , all within a few arcminutes of the quasar , was prepared prior to observations .
each of those was briefly observed ( without closing the ao loop ) to test if it could , with the help of the wfs neutral density filter set , reproduce the wfs photon counts generated by the quasar .
this allowed us to match the photon counts of the quasar to within 10% for most of the objects in our sample .
the ao loop parameters ( membrane stroke and loop gain ) were not modified between observations of the target and the psf . for most of the targets ,
the psf total integration time was much smaller than the qso total integration time . because of the variability of the psf presented in
[ sec : psfstab ] , the residuals of the psf subtraction are limited by effects that do not average well with time , and there is little gain in observing the psf star for longer than a few minutes . however , long exposure times are required to record the faint structure of the host galaxy that are outside the central part of the psf , where psf subtraction is not essential ( where the psf wings are fainter than the host galaxy ) . when possible , the psf star was observed with the same frame integration time as the qso to use the psf frames for sky extraction . in this section ,
we characterize the psfs delivered by gemini+hokupaa and the subaru ao system and investigate their variability .
this preliminary work is essential for attempting the characterization of the qso hosts in our data .
the psf delivered by an ao system is a function of : * * 1 .
guide star v magnitude and color of the object imaged .
* the brightness of the star , as seen by the wfs , strongly affects the performance of the ao system . due to the wavelength - dependence of the seeing ( and the optics of the camera ) , psfs are also expected to be color - dependant . * * 2 .
atmospheric seeing . * under poor seeing conditions , the fwhm of the corrected psf is larger .
in addition , the brightness of the large halo surrounding the psf core increases with the turbulence . *
* 3 . instrumental variables .
* wind - powered telescope shake , change in the optical path , and the primary and secondary mirror wavefront all modify the psf . * * 4 .
ao loop parameters .
* loop gain and membrane stroke ( for curvature systems ) set the speed of the ao loop and the optical gain of the wfs respectively , and therefore affect the psf . despite efforts to match the brightness ( as seen by the wfs ) of the reference psf and the qso , small discrepancies in the two flux levels are unavoidable .
more importantly , we found that the variations in seeing always limit the accuracy of the psf subtraction when the flux of the psf and the qso are matched to better than 10% .
figure [ fig : psfcmpprof ] shows how the radial profile of the psf can change rapidly ( timescale of about a minute ) due to seeing variations .
the far wings of the psf ( 1 to 2 in fig .
[ fig : psfcmpprof ] ) show how much light is scattered by small - scale turbulence ( high spatial frequencies of the seeing ) , which is not corrected by the ao system . in this figure
, it can be seen that these wings are much higher for psf1 than for psf2 .
consequently , because the turbulence was stronger during the acquisition of psf1 , the ao correction is poorer and , inward of about 02 ( dotted vertical line ) , the surface brightness of the psf is lower .
the `` crossing point , '' for which the loss of surface brightness due to the poor ao correction exactly equals the gain of surface brightness due to the increased scattering by small - scale turbulence , is set by the number of actuators of the ao system .
the surface brightness at this crossing point is insensitive to seeing variations .
figure [ fig : psfcmpprof ] also shows how the brightness of the psf affects its radial profile : for brighter psfs , the ao correction is better and flux is taken from an `` intermediate '' annulus between 01 and 03 and put into the central diffraction core of the psf .
the far wings of the psf are left unchanged because of the limited number of actuators of the ao system . during real observation ,
a combination of psf variations due to changing seeing and different source brightness will affect the radial profile .
although the observer has no control over the seeing , it is possible to reduce psf variations by avoiding any change in the instrument during observations .
for example , the ao loop parameters should be kept constant . during this observing program
, it was discovered that the psf stability could be significantly improved by turning off the cassegrain rotator of the telescope . at the gemini and subaru telescopes , which have alt - azimuth mounts ,
the cassegrain units carrying the ao systems are usually rotated to preserve the field orientation on the detector .
the primary and secondary mirror create static wavefront aberrations .
if these static aberrations are rotating in front of the ao system , which has preferential directions due to the geometry of the actuators and also introduces some higher order aberrations , the resulting rotational shear between the incoming wavefront and the aberrations of the ao system changes the structure of the psf .
this very complex psf variation is difficult to predict since the individual sources of aberrations are poorly known .
figure [ fig : rot ] illustrates this effect through the comparison of two differenced psfs , one for which the cassegrain rotator angle was kept unchanged between the two exposures , and the other for which a 15@xmath42 rotation occurred between the two exposures . because some of the aberrations come from the primary and secondary mirror and some come from the ao system and the camera , a psf obtained after rotation of the cassegrain rotator
can not be reproduced by a rotation of a psf obtained at a different cassegrain rotator angle .
we have seen above that turning off the cassegrain rotator makes the psf more stable and therefore , improves the accuracy of the psf subtraction .
there is a second advantage to this technique : the psf and the image of the host galaxy can be decoupled through differential rotation .
when attempting to detect faint structures very close to the qso , we are usually very sensitive to small variations in the psf structure , such as `` flares '' of the psf .
these `` flares '' are more frequent when observing faint targets such as qsos , because the limited number of incoming photons does nt allow fast recovery from events such as a wind - induced telescope shake .
it is possible to encounter such a flare on the object and not on the reference psf , leading to the false detection of a structure in the host galaxy .
careful visual examination of individual frames can help detecting these artifacts but is not fully reliable .
when observing with the cassegrain rotator off , the psf has a constant orientation on the detector while the object rotates , which unambiguously decouples the object from the psf variations . for a `` flare '' to be misidentified as a structure in the host galaxy , it would need to rotate in the psf at a speed reproducing the field rotation . because we observed most objects continuously for more than an hour
, the total rotation during the observation is usually sufficient to efficiently decouple the object from the psf . because of the improved psf stability when observing without the field derotator ,
most of our observations were done with the cassegrain rotator turned off .
the geographical coordinates of the telescope , the ut time of the exposure , the coordinates of the target and the angle of the cassegrain rotator are required , for each exposure , to recover the field orientation .
this observing mode requires the exposure time to be short enough to avoid rotational smearing of the image .
the limit we set on the rotational smearing corresponds to half a full width at half maximum ( fwhm ) at the edge of the detector when the target is at the center .
this limit allows exposure times of 20 s to 30 s on most of the sky . for some objects ,
detector saturation was reached in a few seconds . because of the relatively long readout time of the camera ( 10 s ) on gemini / hokupaa , we chose to obtain a series of short unsaturated exposures and a series of longer exposures which were essential to reach a good flux sensitivity .
we were careful to obtain the longer exposures when the field rotation was slow .
ao psfs are time - variable and have relatively strong and extended wings compared with _
hubble space telescope _
( _ hst _ ) psfs , but the image of the underlying host is made sharp , which allows accurate retrieval of the radial profile in regions where the wings of the central psf are not significantly brighter than the host galaxy .
depending on the objects of our sample , we find that it is possible to accurately recover the host radial profile outward of about 02 ( for hosts which account for a large fraction of the total flux ) to 2
( for hosts which are much fainter than the central point source ) with classical psf subtraction techniques .
this corresponds to about 0.5 kpc to 2 kpc for most of the objects in the sample .
thanks to the large collecting area of 8 m class telescopes , the host radial profile can usually be recovered to up to 5 to 10 . beyond this angular separation ,
the error due to imperfect sky subtraction becomes larger than the host : due to the small field of view of the images , the sky level can not be accurately decoupled from the faint outer part of the host galaxy . the good angular resolution ( 01 to 02 , which is comparable to or better than that of _ hst _ at the same wavelength ) and depth of qso images obtained on a 8 m class telescope equipped with an ao system makes it possible to image small and faint structures in the host .
this is very important for this project , which relies on the ability to characterize the morphology of the host beyond the disk vs. elliptical galaxy classification . for each target ,
the frames were processed and coadded by a data reduction pipeline written in c. flat field and bad pixel maps were created from a set of dome flat exposures with increasing exposure time .
the sky frame was obtained from the dithered qso and nearby psf exposures .
the persistence effect of the detector was corrected for by ignoring the pixels that have been recently strongly exposed to light .
an empirical formula was used to select such pixels from the pixel value and the time of the exposure of the 10 preceding frames .
the first steps described above ( flat fielding , bad pixel identification and sky subtraction ) were applied to both the psf and the qso frames . for each qso ,
a coadded image was generated by derotating each qso frame before coaddition .
as shown in fig .
[ fig : exposuremap ] , because of the combination of dithers , rotations and bad pixels masking ( or pixels ignored due to the persistence of the detector ) , the total exposure time varies across the final image .
a reference psf frame is obtained through coaddition of individual psf frames , but without field derotation . to avoid increasing the noise of the psf - subtracted image , the far wings of the psf were replaced by synthetic psf wings generated from a fit of the wings radial profile .
the psf image matching the coadded qso image was obtained by coadding multiple copies of the reference psf frame with the proper rotation angles and relative exposure times .
images are shown in figure [ fig : qsoimages1 ] for each object of our sample .
the flux in the raw images ( left ) is dominated by the bright central nucleus , although the host galaxy is usually visible without psf subtraction , thanks to the good contrast and high angular resolution delivered by ao . the psf - subtracted images ( middle )
reveal more host galaxy details , especially in the central part of the image .
the quality of the psf subtraction is highly variable through the sample , and depends upon many factors ( see [ sec : psfstab ] ) . in some cases ,
the reference psf used for subtraction is less peaked ( and , therefore , has a comparatively brighter `` halo '' component ) than the actual psf during observation .
if the host galaxy vs central nucleus luminosity ratio is low , the residual of the psf subtraction is then negative in an annulus surrounding the positive residual at the center of the image ( see [ sec:2dana ] for a description of the 2-d analysis algorithm used to subtract the psf ) : this effect can be seen on pg0026 + 129 , pg0838 + 770 , pg1411 + 442 .
psf subtraction errors due to poor stability of the ao control also occur when the source is faint , the atmospheric turbulence is strong and/or clouds vary the source brightness during the observation .
these effects are responsible for the poor psf subtraction of pg0050 + 124 , pg1116 + 215 and pg1302@xmath5102 . during the observation of pg1617 + 175 , an avalanche photodiode of the ao wavefront sensor burned out , suddenly creating a bright extension in the psf delivered by the system , as can be seen in the psf subtracted image .
many of the images also show artefacts due to sky subtraction residuals .
these features are usually faint and beyond 5 from the qso .
they are most visible in the images of pg0804 + 761 ( curved feature 4 north - west of the qso and compact artefacts 10 due east , north , west and south of the qso ) , pg0838 + 770 ( about 5 to 7 south of the qso ) , pg0844 + 349 ( straight edge and extended feature 6 east of the qso ) , pg0923 + 201 ( straight feature running north - south about 5 east of the qso ) , pg1001 + 054 ( faint extended fringes 5 to 10 south of the qso ) , pg1119 + 120 ( faint extended fringes mostly in 10 south - west of the qso ) , pg1307 + 085 ( 4 faint features in a 10 square pattern centered on the qso ) , pg1626 + 554 ( wide 5-radius arc centered in the qso , mostly visible in the northern half of the image ) and pg2214 + 139 ( faint broad features in a 8 square pattern centered on the qso ) .
the flux within a 55 diameter circular aperture centered on the qso was measured for each object , with an accuracy of about @xmath435% at the 1 @xmath44 confidence level .
this aperture size allows easy comparison with previous measurements by @xcite and @xcite . for the objects observed in non - photometric conditions , only a lower limit for this flux
could be determined .
table [ tab : photom ] compares our photometry with other measurements from the literature and from the 2mass archive .
no correction was applied to account for differences between the 2mass @xmath45 and mauna kea @xmath40 photometric systems , as these two filters are very similar @xcite .
the values shown in table [ tab : photom ] are consistent with previous findings that show that many of the qsos observed in this sample appear to show some variability in the near - infrared .
the amplitude of variability shown in this table is in good agreement with @xcite : both sets of data identify pg1001 + 054 , pg1426 + 015 , pg1613 + 658 , pg1617 + 175 and pg2130 + 099 as the most variable qsos of our sample .
since the observations made under non - photometric conditions need to be calibrated with other photometric measurements of the same object , the variability of qsos can cause errors in the estimate of the host galaxy luminosity . in order to minimize such errors ,
the time difference between our observations and the photometric measurement should be kept small .
2mass measurements , rather than older measurements from @xcite or @xcite , were therefore used for photometric calibration of objects for which no reliable photometry could be derived with our observations .
in one case ( pg1116 + 215 ) , the lower limit obtained with our data is higher than the 2mass measurement : this lower limit was then adopted for photometric calibration .
we note that pg0050 + 124 was observed under very poor conditions : poor seeing and large absorption by clouds resulting in highly variable adaptive optics performance .
we did not attempt to perform radial profile or @xmath46 analysis for this object , which explains why no photometry was adopted in table [ tab : photom ] .
the 2-d analysis relies on a least - square fitting of the image according the the following equation : @xmath47 pixels too close to ( distance less than 1 to 2 depending on the qsos ) , or too far from ( more than 8 to 10 ) the center were excluded from the fit because psf calibration errors dominate the host s signal in the inner region of the image , and sky calibration is unreliable in the outer parts of the image ( due to the small field of view of the detector ) . rejecting the central part of the image in the fitting algorithm
could allow solutions where the total flux of the image is not preserved : the psf contribution could be increased to a point where its total flux exceeds the total flux in the raw image to fit the host galaxy beyond the inner edge of the fitting region . to avoid this effect ,
the total flux of the psf component was not allowed to exceed the total flux in the raw image .
bright companions , when present , were also excluded from the fit .
the value of @xmath44 was directly measured on the image .
this 2-d fit can have up to 9 free parameters : 1 for the intensity of the central source , 4 for the bulge component ( modeled in this work as a single de - vaucouleurs @xmath48 profile ) and 4 for the disk component ( modeled as an exponential profile ) .
a simulated annealing minimization routine was used to find the absolute minimum .
the @xmath46 minimization routine was first run with the full 9 free parameters to yield @xmath49 , the minimum value of @xmath46 for a exponential + @xmath48 de vaucouleurs profile host model .
then , @xmath50 ( @xmath48 host only ) and @xmath51 ( exp host only ) were computed , with 5 free parameters in each case .
the resulting reduced @xmath46 values for each of these three cases are listed in table 4 for all of our targets .
it is worth noting here that the noise component @xmath44 in equation [ equ : chi2 ] is often poorly characterized , due to the difficulty of estimating the temporal variations of the psf wings .
thus the reduced @xmath46 values listed in table 4 can not be reliably compared between targets , but different fits obtained from the same image can still be compared to classify the host galaxies . for each object ,
the 1-d radial profiles extracted from the 2-d image and the corresponding 2-d fit models ( point nucleus + host model convolved by the psf ) are shown in fig .
[ fig : qsofits1 ] to [ fig : qsofits11 ] .
the 1-d profile of the residual error , which is also shown in the same figures for the 3 fits , tends to become large at small angular distance from the center because of errors in the psf subtraction .
the domain within which the fit was performed is shown by 2 vertical red lines .
due to the limited field of view of the instrument ( 20 ) , estimations of sky background level and the psf are prone to errors beyond @xmath528 radius .
this explains the peculiar behavior of some profiles beyond @xmath528 ( for example , the abrupt drop at 10 in the psf profile for pg1440 + 356 ) , and justifies our choice to exclude the outer part of the profile for host galaxy characterization . in deciding which of the three @xmath46-model fits listed in table 4
best characterizes the host galaxy , the following criteria were used .
unless the 2-d fit strongly favors a 2-components bulge+disk host ( i.e. if @xmath53 and no more than 70% of the total host flux is allocated to either the bulge or disk in this bulge+disk fit ) , the smallest of the @xmath50 and @xmath51 is adopted as the result of the 2-d fit .
the adopted fits for each object and each filter used are listed in columns 24 of table [ tab : hostclass ] .
however _ , in a number of cases , the difference in the @xmath46 values for the exponential vs. the @xmath48 de vaucouleurs profile were small enough to render the choice of a particular host model somewhat meaningless .
in the cases where @xmath54 , we consider the result of the fit to to favor one model , but with very low confidence ( noted in parentheses in table [ tab : hostclass ] ) .
and in those cases where @xmath55 , or in the case of bad psf subtraction , we consider the fit to be inconclusive ( a question mark in table [ tab : hostclass ] ) .
since the overall qso host classification scheme used in this work takes into account _ both _ the results of the 2-d @xmath46-model analysis and morphological information visually extracted from the images , table [ tab : hostclass ] also lists information obtained from visual inspection of the host images , including the presence or absence of `` bars , '' ( spiral ) " arms , and the overall degree of `` dist(urbance ) , '' respectively .
additional visual information is given in the notes .
each host was finally classified using a combination of information which included the best - fit @xmath46-model type " and the visual appearance .
the final adopted type " listed in col .
8 of table [ tab : hostclass ] was either d@xmath56 " = disk present ,
e " = elliptical , or b+d " = bulge+disk , except for those objects with insufficient and/or ambiguous information from the @xmath46-model fits , or objects with strongly disturbed " disks , where a ? " was listed .
we note here that since the 2-d fitting analysis does not include the central part of the image ( due to psf variability ) , the type `` disk present '' does not exclude the presence of a bulge ( which would be hidden in the central part of the image ) , and the `` disk present''/``bulge+disk '' distinction is therefore unclear
. `` elliptical '' and `` bulge '' ( in the bulge+disk type ) refer to a @xmath48 for all 2-d fits performed in this work .
strongly disturbed hosts were not classified according to host type , since such a classification would be difficult and unreliable .
while the degree of disturbance is estimated purely from visual inspection of the images , the rules adopted for the host `` adopted type '' classification were as follows : * if a bar or spiral arms are obvious in the image , the host is classified as a `` disk present '' , or a `` bulge+disk '' ( the decision between `` disk present '' and `` bulge+disk '' is made using the 2-d fitting analysis results ) . *
if no spiral structure or bar is visible in the image , the host classification is given by the 2-d @xmath46-model fitting analysis .
if the host has been imaged in several spectral bands , and the corresponding 2-d @xmath46-model fitting analysis do not agree , the fitting result obtained for the spectral band with the best psf subtraction is adopted .
it is worth restating here that we have chosen to give priority to the features visible in the images rather than the results of the 2-d @xmath46-model fits , primarily because the 2-d @xmath46-model fits are prone to errors .
it can be seen from the 1-d profiles in fig .
[ fig : qsofits1 ] to [ fig : qsofits11 ] that for many objects , the 2-d fit does not produce a reliable results - we found that changing the fit domain often changes the result of the fit .
thus , it was reasoned that , for example , in several objects where a bar or spiral arms were clearly visible in the host images , the host should not be classified as an elliptical , even if the 2-d fit suggests so .
it also should be noted that not all objects gave ambiguous results .
some objects are quite easy to classify , as the @xmath46-model fit results are clear and agree with the images .
but overall , the images show that the @xmath46-model fits often give misleading results , and in @xmath219% of the objects ( 6/32 ) the hosts are so disturbed that a @xmath46-model fit is of little relevance . before presenting the final fit parameters corresponding to the host adopted type " listed in column 8 of table [ tab : hostclass ] , it is worth comparing our @xmath46-model fit results and observed visual features with results from previous published studies of these objects .
each object is discussed separately in the paragraphs below , including comparisons with previous analyses by different authors and their published `` host types , '' and including remarks about the various difficulties encountered in our own analysis , which often led to ambiguities in the final choice of host type . * * * + the image shows an obvious host , slightly elongated east - west , with no apparent structure .
the radial profile analysis fails to identify clearly the galaxy type , but gives host magnitudes ranging from @xmath57 ( bulge+disk fit ) to @xmath58 ( disk fit ) , and nucleus magnitudes ranging from @xmath59 to @xmath60 .
although our image does not allow classification of this host , the _ hst_/nicmos observations of @xcite suggest that it has an @xmath48-law profile .
a non - stellar apparent companion is visible 4.8 from the central nucleus ( @xmath61 , @xmath62 ) .
* * * + this object is also known as i zw 1 and is often classified as a seyfert galaxy , particularly since the host shows clear spiral arms @xcite . in our observations , the spiral arms are visible in the host galaxy across the field of the detector ( 20 ) .
although the observing conditions for this object were poor ( making it impossible to perform an efficient psf subtraction or photometry ) , the morphology is unambiguously that of a spiral host .
@xcite have found spectroscopic evidence for continuing star formation in the disk as well as morphological evidence for weak interaction with the small companion to the west .
@xcite find a mass in molecular gas of @xmath63 m@xmath64 , and recent observations with the bima interferometer @xcite resolve individual giant molecular clouds in a ring structure from about 1.2 to 3.5 kpc from the nucleus . * * * + the host galaxy of pg 0157 + 001 (= mrk1014 ) is large and bright , and it shows 2 prominent spiral - like features that extend from @xmath65 kpc to at least 28 kpc ( limited by the edge of the detector , 10 from the nucleus ) , as well as a bright bulge .
this galaxy has been the subject of numerous ground - based and _ hst _ imaging studies ( e.g. , @xcite ) .
the spiral features are clearly not conventional spiral arms ( see fig . 2 of @xcite ) ; instead , as many of the previous studies have emphasized , they are almost certainly tidal tails generated in a past major merger .
these tails provide strong evidence that both of the merging galaxies themselves had cold stellar disks .
the host galaxy is highly asymmetric , with the northern tail brighter than the southern tail .
several bright knots can be seen in the northern tail , which are also clearly seen in _ hst _ images ( e.g. , @xcite ) .
a small , bright companion is visible to the south - west at a projected distance of about 25 kpc .
the host is well fitted as a @xmath66 bulge+disk galaxy with a @xmath67 central nucleus , which is compatible with the structure seen in the image . @xcite
confirm the presence of significant populations of young stars , particularly along the leading edge of the northern tail .
* * * + this host is best described as a `` bar '' about 5 kpc in radius .
an extended ( kpc size ) faint companion is visible to the east of the qso , at a projected distance of 4 kpc from the qso .
the curved structure visible in the image about 8 to the north - west of the qso is a artifact from the camera .
no reliable 2-d fit was found for this galaxy in @xmath40 , and the fit in @xmath0 suggests an elliptical galaxy .
however , the bar indicates that this galaxy contains a prominent disk .
the only previous published imaging study of this object is that of @xcite , who found a profile consistent with a disk galaxy from their ground - based @xmath0-band imaging . @xcite
detect molecular gas mass of about @xmath68 m@xmath64 in this galaxy . * * * + the central part of the host of pg0838 + 770 is a 9 kpc radius bar . a noticeable twist of this bar is seen in our image , and a 20 kpc long arm is starting from the east end of the bar .
this arm seems to connect both ends of the bar , thus forming the northern half of a ring immediately surrounding the bar .
visible images @xcite show both the bar and two spiral arms outside the ring .
an extended ( about 2.5 kpc diameter ) apparent companion is also visible to the northwest at a projected distance of 12 kpc .
although the host is best fitted by a @xmath48 law , the bar and spiral arms show that a substantial disk is present . *
* * + two symmetric spiral arms are clearly visible , extending to about 9 kpc radius in our image .
@xcite describe the host as a `` barred spiral , with some complex outer structure .
'' our 2-d fit , which suggests a dominant bulge , is uncertain because of poor psf subtraction . * * * + although some previous studies have listed this object as interacting @xcite or possibly interacting @xcite , apparently because of the bright galaxy projected @xmath69 to the south , the host is unambiguously a very symmetric elliptical .
it has an apparent extended companion at a projected distance of 7 kpc ( 22 ) almost due north .
this host was also identified as an elliptical from _ hst_/wfpc2 observations @xcite .
the 2-d fit shows that the host galaxy ( @xmath70 , @xmath71 ) has a low ellipticity ( about 0.1 ) , and the central nucleus is about 1 magnitude brighter than the galaxy . * * * + this host was identified as a spiral from ground - based imaging by @xcite ; however , all subsequent imaging studies , including _ hst_/wfpc2 observations of @xcite and @xcite , have found the host too faint to classify reliably .
this is also true for our data .
the brightest part of the host is visible as a faint extension to the south of the qso in our image .
@xcite find marginal evidence favoring an elliptical profile , but they also claim a possible tidal arm to the southwest , which would indicate the presence of a disk . * * * + our image shows a highly elongated host ( long axis radius of 4 kpc ) which is probably a bar but could also be a small inclined disk .
the 2-d fit suggests an elliptical galaxy , but we choose to classify the host as a `` disk present '' from visual inspection of the image .
@xcite see only a small , uniform host from ground - based optical and near - ir imaging . *
* * + the quasar , also known as 4c13.41 and pks1004 + 13 , is a powerful steep - spectrum radio source .
the host galaxy is large , bright and smooth , and is shown to be an elliptical by the 2-d fit .
@xcite also classify the host as a bright elliptical from _ hst_/wfpc2 imaging . * * * + the psf subtraction is poor for this object .
nevertheless , the host appears to be asymmetric , with several companions : a compact companion at a projected distance of about 4 kpc to the west , a faint extended ( 2 kpc diameter ) companion at 5.5 kpc projected distance to the south , a compact companion ( star ? ) at 12.7 kpc to the south - southwest , and an elongated bright galaxy at 33 kpc to the west ( outside the field of view of the image shown in fig .
the 2-d fit to the host suggests that it is a disk , although there is considerable uncertainty in the determination .
@xcite list the host as `` probably an elliptical . ''
* * * + the host is bright and smooth , with a north - south elongation .
an elongated companion is visible 2.6 kpc ( 27 ) to the north , projected on the host galaxy .
optical images @xcite show 2 faint extended rings on the east and west sides of the brightest part of the host .
these features are too faint to be detected in our near - ir images .
our 2-d fitting gives a bulge+disk decomposition and shows that the host galaxy is almost a magnitude brighter than the central nucleus .
* * * + our image shows a bright and large host with spiral arms extending beyond 10 kpc .
the 2-d fit suggests the presence of a bulge of luminosity comparable to the disk component .
a bright apparent companion , at a projected distance of 6.6 kpc ( 57 ) , is clearly resolved and shows a small elongation . *
* * + this host is too faint to be classified in this work . @xcite
have concluded from their ground - based @xmath72-band imaging that the host is dominated by a disk . * * * + this host was identified as a spiral from ground - based visible observations @xcite and _ hst_/wfpc2 observations @xcite .
our observations show a prominent bar ( radius is 5 kpc ) , with faint spiral arms starting from their ends .
the result of the 2-d fit is a bulge+disk model about one magnitude brighter than the central nucleus .
we classify this host as a `` disk present '' because of the structures visible in our image . * * * + the quasar is a strong flat - spectrum radio source .
two compact , but resolved , bright companions are within 10 kpc of the qso . although decoupling the closer companion and the host in the 2-d fitting analysis is difficult , visual inspection of the image suggests that the host of pg 1302 - 102 may be a somewhat asymmetrical elliptical galaxy .
the 2-d fit favors a @xmath48 model about 1.5 magnitudes fainter than the central nucleus , which we adopt for the classification of this host . @xcite and
@xcite ( working from the same _ hst_/wfpc2 image ) both suggest an elliptical host . * * * + our image shows a smooth symmetrical host with no apparent structure , and the 2-d fit shows that it is an elliptical galaxy with no detectable elongation .
analysis of _
hst_/wfpc2 images gives a similar result @xcite .
the four small ojects at the corners of a perfect 10 by 10 square centered on the qso are sky subtraction artefacts . * * * + this host is smooth , symmetric and moderately elongated in the north - south direction . although it has been described as an early - type spiral from _
hst_/wfpc2 observations @xcite , our 2-d fit favors an elliptical host .
_ hst _ observations also give a closely @xmath48-law profile , but pick up weak but definite spiral features , which are not seen in our image probably because of the longer wavelength .
classification for this host is therefore somewhat uncertain , although we adopt the `` elliptical '' classification for consistency with the criteria adopted for the rest of our sample . * * * + our image shows a large smooth host , slightly elongated in the north - south axis .
ground - based observations in the visible show a smooth host , best fitted as an elliptical @xcite .
our 2-d analysis does not allow classification of this galaxy .
deep optical imaging @xcite shows an faint extension of the host to the west .
this feature is marginally detected in our image , and appears as a subtle light excess about 5 west of the qso . * * * + highly disturbed host , with a bright elongated companion at a projected distance of 3.6 kpc ( 21 ) .
three long tidal arms or loops are visible in our image .
this host had previously been classified as a spiral @xcite .
deep optical images @xcite show that the tidal features extend out to at least 30 kpc .
our 2-d fit does not allow classification of this host , and gives different answers in different bands .
we therefore classify this object as highly disturbed . * * * + strongly disturbed host , with a bright companion at a projected distance of 4.1 kpc ( 25 ) .
the host galaxy is asymmetrical : its south - west side is brighter and shows tidal features .
our 2-d fit favors an elliptical host , but the apparent tidal features lead us to classify this object as strongly disturbed . * * * + our image shows a smooth host galaxy .
the northern half of the host is slightly brighter than the southern half at a radius of about 2 kpc .
a companion is resolved ( about 2 kpc diameter ) at a projected distance of 10.8 kpc ( 48 ) to the west .
our 2-d fit suggests an elliptical , which we adopt for the host classification . * * * + two symmetric spiral arms are visible in this host , without any obvious evidence of disturbance .
the 2-d fit for this object is of poor quality , and its result ( high surface brightness disk galaxy of small size ) is probably driven by a mismatch in the psf or the background sky level . @xcite
classify the host as a spiral with a bar . * * * + this bright host galaxy shows a very complex structure , with prominent tidal arms and a bright resolved apparent companion projected 5.1 kpc ( 22 ) west of the qso .
this structure was first reported by @xcite .
similar structure is shown by an _
hst_/nicmos image @xcite .
deep optical images @xcite show tidal arms extending out to 35 . * * * + the psf subtraction is poor because of an apd failure in the ao system .
our image shows a smooth host and the apparent asymmetry in the image is most likely due to psf subtraction residuals .
an elliptical host is favored by the 2-d analysis , and adopted for host classification . due to the poor quality of the psf subtraction
, this classification might not be reliable . * * * + smooth host .
the 2-d analysis suggests an elliptical host , but the reference psf for this object is not very good , and the host classification might not be reliable .
the rings visible in the residual wide - field image of fig .
[ fig : qsoimages1]g might be due to sky subtraction residuals . * * * + highly disturbed host , with an interacting companion 7 kpc ( 16 ) north of the qso . a tidal arc , especially bright in the north and west ,
is associated with the companion .
several bright knots are visible within this tidal feature .
fainter extended features to the south - west and south - east of the qso are visible in our image . despite the high quality of our data ,
no reliable classification is obtained by our 2-d fit for this highly disturbed object .
@xcite obtained imaging with an earlier version of the hokupaa ao system at the canada - france - hawaii telescope , detecting the companion and its arc .
they also found spectroscopic evidence for a young ( 85 myr ) stellar population in the companion .
@xcite , from their _
hst_/nicmos imaging of this object , suggest that the companion is a ring galaxy , produced by a small - impact - parameter collision . * * * + this spiral host shows two symmetric , high - contrast spiral arms extending out to at least 65 from the qso .
a non - stellar apparent companion is visible 9 to the east - southeast of the qso .
these features are also seen in ground - based and _ hst_/wfpc2 optical imaging @xcite , which also show evidence for a central bulge .
strangely , @xcite classify this object as an elliptical , although with marginal confidence . despite the poor psf subtraction in our image of this object , the 2-d
fit correctly identifies this host contains a prominent disk .
there is a good match between the fit parameters obtained in the @xmath0 and @xmath40 images , both indicating that the host is almost 2 magnitudes fainter ( @xmath73 , @xmath74 ) than the central nucleus . * * * + as the name indicates , this quasar is a strong radio source .
the host is smooth and elongated , with 3 small apparent companions ( projected distances are 11 kpc ( 25 ) , 12 kpc ( 28 ) and 20 kpc ( 47 ) ) within 5 of the qso .
the companion 12 kpc to the south is of stellar appearance .
our image suggests that , in the central part of the host ( at about 2 from the qso ) , the southeast side of the galaxy is brighter than the northwest side .
however , the effect is subtle and may be due to a psf artifact ; we therefore have chosen to classify this host as undisturbed .
the 2-d analysis suggests that the host is an elliptical .
the classification by @xcite is ambiguous , whereas @xcite classify the host as an elliptical from an _ hst_/wfpc2 image .
spectroscopy of the host @xcite also support classification as an elliptical galaxy . * * * + this host has a very peculiar radial profile , although the 2-d analysis suggests that it is an elliptical .
sharp concentric rings surround the the central qso at radii ranging from 2.5 kpc to 5 kpc , and are especially obvious in the north - west part of the host .
these shells are probably tidal debris from a recent merger . * * * + the quasar is a 4c and parkes radio source .
the smooth host is moderately elongated in the east - west direction , and it has a non - stellar apparent companion 43 to the north .
the 2-d analysis shows a slight preference for a disk , but the classification is to be considered somewhat unreliable due to poor psf subtraction
. it would be quite unusual for a radio source of this power to have a host dominated by a disk , and @xcite report an @xmath48 profile for it . * * * + this highly disturbed host shows 3 tidal arms and a close companion at 5 kpc ( 17 ) projected distance .
the small companion is composed of a bright compact nucleus and two fainter knots about 03 on either side of it , to the east and west .
spectroscopy @xcite has confirmed that the companion and qso redshifts match .
this host was identified as a elliptical undergoing strong interaction from _ hst_/wfpc2 observations @xcite , which is in agreement with our 2-d fit and our image . due to the features visible in our image , we chose to classify this host as strongly disturbed .
the quasar itself is a strong parkes radio source .
table [ tab:2dfitresult ] lists the value of the @xmath46-model fit parameters ( for the adopted host " type listed in col .
8 of table [ tab : hostclass ] ) according for the following host decomposition : @xmath75 where @xmath76 for all objects , we verified that the minimization algorithm gives results independent of the starting point , and that @xmath49 is smaller than @xmath50 and @xmath51 . for e " and d@xmath56 " types in col . 8 of table
[ tab : hostclass ] only the first or second term , respectively , on the right - hand side of equation [ equ : hostfit ] is relevant , while for type b+d " all 9 parameters are used in the fit . when the host classification is uncertain ( ? " in col .
8 of table [ tab : hostclass ] ) , the 9 parameters of the bulge+disk fit decompositions are listed in table [ tab:2dfitresult ] .
the results presented in the previous section illustrate both the overall success of our survey in being able to detect the great majority of the host galaxies of nearby pgqsos , as well as the difficulties encountered when attempting to fit model radial surface brightness profiles to individual hosts . the two year timeframe of the survey provided sufficient opportunity to revisit objects where the initial data had been compromised by poor seeing and/or problems with the stability of the psf . in the end
we were able to reliably detect the host galaxies in all but four of our targets ( 28/32 ) , with sufficient signal - to - noise to allow fitting of the radial surface brightness profiles out to radii typically @xmath710 kpc a tribute in large part to the light gathering power of 8-m telescopes .
however , despite extreme attention to detail , the stability of the psf was still a limiting factor in preventing accurate subtraction of the central point source , thus compromising our ability to determine the host light profile a radii @xmath61 - 3kpc , thus making the fit result rather sensitive to the choice of inner radius limit .
our model fits also proved to be somewhat sensitive to the exact choice of outer radius limit , thus in the end our refusal to use @xmath46-fits alone , and our attempt to give at least equal weight to visual inspection of the images , in the final determination of host type " . despite these caveats ,
there are still several important statistical results and trends that clearly emerge from the data .
twenty - eight of our 32 objects had detected hosts for which we attempted make a classification , and we discuss the distribution of host type " and host absolute magnitude for these 28 objects below .
the remaining 4 objects were not classified due either to having hosts which were too faint or having data which suffered from a bad psf ( see footnote to table [ tab : hostclassstatnir ] ) .
table [ tab : hostclassstatnir ] summarizes the number of objects and their mean h - band absolute magnitudes for each host type " e " ,
d@xmath56 " , b+d " as well as for those objects where a host type " could not be reliably determined ( i.e. unknown " ) , and also by degree of disturbance " non - disturbed " , disturbed " , strongly disturbed " .
these classifications follow the category listings as used in column `` dist . '' and column `` adopted type '' , respectively of table [ tab : hostclass ] .
the percentage distributions of host type " and the relationship between degree of disturbance " and the host type " can be seen from the graph in fig .
[ fig : hosttype ] .
it is immediately apparent that the hosts of the sample of pgqsos in our survey are not characterized by a single dominant type . for the subsample of 19 objects with classified host type ( e " , d@xmath56 " ,
b+d " in table [ tab : hostclass ] ) , there appears to be an almost equal split between objects with no detected disk component ( e " ) and objects with detected disks ( d@xmath56 " and b+d " ) .
it is also apparent that a significant fraction , ( 9/28 = 30% ) , of the hosts of pgqsos show clear signs of disturbance and/or are strongly disturbed .
if we then make the reasonable assumption that the disturbed " , and even more likely the strongly disturbed " , hosts also have a disk component , then the percentage of objects with a detectable disk component could be as large as 57% ( 16/28 ) to 68% ( 19/28 ) .
table [ tab : hostclassstatnir ] also lists the median and average @xmath77 values , and the 1@xmath44 uncertainty in the average for both the host and nucleus , for the total sample ( 28 objects ) and for each host type " subsample .
these values have been computed directly from the adopted fits for each object as listed previously in table [ tab:2dfitresult ] .
the median and mean host magnitudes typically differ by less than 0.3mag , with no obvious systematic offset . in what follows we will use the mean values to characterize the properties of the total sample and the host type " subsamples . for our entire sample ( 28 ) ,
the mean host absolute @xmath0-band magnitude is @xmath78 , which is equivalent to @xmath79 [ assuming @xmath80 ] .
our value of @xmath81 for the mean host luminosity is almost identical to the values of @xmath82 and 2.3@xmath4 found respectively by @xcite , and @xcite from their earlier ground - based observations of samples of qsos selected from the same bqs sample of pgqsos , and using the same @xmath83 redshift cutoff as our current sample .
the latter study also finds a range of 0.57.5@xmath4 for the host galaxies , very similar to our measured range of @xmath20.6310.0@xmath4 .
mean @xmath0-band host magnitudes and luminosities for samples of pgqsos at @xmath1 , and for quasars in general , are not that abundant in the literature , but the few reported measurements all seem to fall generally in the range @xmath213@xmath4 , with the lower value being found for low - luminosity qsos and the higher value being found for more luminous qsos ( e.g. , @xcite ) .
it is interesting to note that this range of host @xmath77 is _ nearly identical _ to the range in the average @xmath77 found for our different host type " _ sub - samples_. table [ tab : hostclassstatnir ] shows that the lowest mean value@xmath84 ( @xmath85)is found for hosts of type d@xmath56 , i.e. , hosts which have a dominant disk component , while the mean value for the ellipticals ( type e ) in our sample is @xmath86 ( @xmath87 ) .
the similarity in range and mean _ h_-band luminosities for different host type with previous mean values versus qso luminosity could be explained , for example , if there were a clear correlation of host type with qso luminosity , and indeed we do see such a correlation , as discussed below . a more complete representation of the range of properties of our sample of pgqsos is given by the data shown in both fig .
[ fig : mhhmb_class ] and fig .
[ fig : mbvsmh_class ]
. the color coding and shading of the symbols used to represent the individual objects is the same in each of the figure panels , and provides an aid to easily see the mean and range of host and nuclear properties for objects with different host type and degree of disturbance .
figure [ fig : mhhmb_class ] ( top panel ) clearly shows a distinct change in host type with increasing @xmath77 of the host .
hosts with dominant disk components ( i.e. `` d@xmath56 '' ) are preferentially clustered at the low end ( i.e. , @xmath2sub-@xmath4 to @xmath4 ) of the observed range , whereas e - type hosts dominate at above-@xmath4 .
objects which have both a relatively large bulge component and a detectable disk ( i.e b+d " ) are distributed more like e - type hosts , although the relatively small number ( 3 ) of b+d " hosts limits our ability to say much more , other than that the they cluster near @xmath88 . in general ,
[ fig : mhhmb_class ] ( top panel ) , shows that disturbed and strongly disturbed hosts are found at all values of host @xmath77 , although there seems to be a tendency for the strongly disturbed hosts to be found at the upper end of our observed range of @xmath77 . finally , we comment on the apparent bimodal distribution of host @xmath77 in figure [ fig : mhhmb_class ] , simply by noting that with just a relatively small shift in the computed host @xmath77 for a few objects , this could just as easily be a single distribution with a high luminosity tail .
figure [ fig : mhhmb_class ] ( bottom panel ) also shows a change in the distribution of host type with increasing @xmath89 of the qso , similar to what is seen with increasing host @xmath77 ( top panel)e - type hosts dominate at @xmath34 while at @xmath29 hosts with prominent disks ( d@xmath56 " and b+d " ) appear to become an increasing dominant fraction of all objects with decreasing b - band luminosity of the qso .
several earlier imaging studies of low - redshift qsos ( i.e. , @xmath1 ) have also commented on the fact that lower luminosity qsos appear to have a larger fraction of hosts with a distinguishable disk component , while high luminosity qsos seem to reside primarily in elliptical hosts .
the collective series of papers by mcleod and coworkers @xcite used both ground - based and _ hst _ near - infrared imaging data of a pg sample that is most similar to ours to conclude that high luminosity qsos seem to predominantly live in elliptical hosts while the host types of low - luminosity qsos are more likely to show evidence for disks . independently , dunlop and coworkers @xcite , using a combination of ground and space - based , near - infrared and optical imaging data for a mostly different sample of qsos , reached generally similar conclusions when they emphasized the almost exclusive presence of elliptical hosts for the higher luminosity qso in their sample , while also pointing out the increasing number of disks found in the hosts of the lower - luminosity , radio - quiet qso in their sample .
and recently , @xcite provided a comprehensive summary of their _ hst _ archival study of 71 qso host galaxies with wfpc2 images in f606w or redder filters , which concluded that
elliptical hosts are typically twice as luminous as spiral hosts " , although they also noted that the luminosity distributions of elliptical and spiral hosts for the radio - quiet qsos were generally compatible " . a more comprehensive presentation of our data is given in fig .
[ fig : mbvsmh_class ] , which shows the relationship of the computed 2-d fit value of @xmath77 ( column 5 of table [ tab:2dfitresult ] ) and the observed @xmath89 ( column 5 of table [ tab : sourcelist ] ) for each qso , along with the symbol for host type , the measured redshift and an indicator for radio loudness ( rlq " ) .
in addition to showing the trends of host type versus @xmath77 and @xmath89 discussed above , fig . [
fig : mbvsmh_class ] also shows how the ratio of these two quantities ( which was summarized by host type in cols .
9 - 11 of table [ tab : hostclassstatnir ] ) varies over the sample .
one interesting question is whether there is any evidence for a correlation between qso luminosity and the luminosity of the host galaxy .
the results shown in fig . [ fig : mbvsmh_class ] suggest not , at least for hosts with @xmath90 ( i.e. @xmath20.6 - 4 @xmath4(host ) ) ; if anything , the mean ratio @xmath91 versus @xmath77(host ) , in this range of @xmath77(host ) , seems to _ decrease _ with increasing @xmath77(host ) , but this result is not significant because of small number statistics .
however , we also note that that all of our objects with @xmath92 are at @xmath93 , and the stated @xmath10-band limit ( @xmath94 ) of the bqs @xcite implies that we should still be complete out to this redshift for qsos with @xmath95 . for hosts with @xmath96 , fig .
[ fig : mbvsmh_class ] does seem to imply that they produce more luminous qsos .
however , this result is clearly compromised by incompleteness due to the @xmath10-band limit of the bqs ; all of the sources with @xmath96 lie at redshift @xmath97 , which excludes detection of qsos with @xmath29 . as was pointed out by the referee ,
what does seem to be significant from fig .
[ fig : mbvsmh_class ] is that even the most luminous galaxies with disks do not seem to harbor really luminous qsos .
it would seem that it may be necessary to have a merger violent enough to completely destroy the disks of the progenitor galaxies to produce the most luminous qsos .
the number of radio - loud quasars ( rlqs ) in our sample is relatively small ( 5/28 = 18% ) , but it is representative of the fraction found in previous studies of large samples of qsos , which show that rlqs are only @xmath215% of the total population of optically - selected qsos . although we did not specifically distinguish rlqs and rqqs when computing mean host @xmath0-band luminosities for different host types " in table [ tab : hostclassstatnir ] , the rlq hosts are explicitly identified in fig .
[ fig : hosttype][fig : mbvsmh_class ] .
it is immediately obvious that rlqs live in the most luminous qsos ( i.e. , @xmath98 ) with the most luminous hosts ( i.e. , @xmath99 corresponding to @xmath100@xmath4 ) .
these facts , coupled with our previously discussed trends of host type with qso @xmath10-band and host @xmath0-band luminosity , are consistent with previous published results which claim that rlqs are typically 1 - 2 mags more luminous than rqqs in the optical , and that rlqs are exclusively found in elliptical hosts , which in turn are usually much more luminous typically 1 - 2 magnitudes , depending on whether the observations are in the optical or near - infrared , respectively than the mean values found for the hosts of rqqs ( e.g. , @xcite ) .
although the origin of the observed dichotomy in the mean properties of rlq and rqq hosts is somewhat beyond the scope of our current study , it is nevertheless worth restating the obvious that qso surveys biased toward radio - selected objects will necessarily find more luminous , elliptical hosts when compared with pure optically selected samples of qsos .
one of the expected strengths of ao observations of qsos using 8 m telescopes was the ability to better characterize the host type " , and in addition to detect faint extended structure in the host which may have escaped detection in previous imaging observations with smaller telescopes .
although the majority of our host classifications are in general agreement with previous results ( see [ sec : notes ] ) , our images were more likely to allow us to distinguish asymmetric features ( e.g. putative tidal debris , central bars , etc . ) from , for example , more symmetric spiral structure , and to better separate out compact , circumnuclear structures from the bright qso nucleus .
we note that since we rely on morphological features such as tidal arms to classify hosts as `` not disturbed '' , `` disturbed '' or `` strongly disturbed '' , this classification is relatively insensitive to usually very symmetrical psf subtraction residuals .
for example , the host type of pg 1626 + 554 and pg 2251 + 113 are unknown due to poor psf subtraction , but the smooth host images obtained lead us to classify them as `` not disturbed '' . according to our analysis , 30% ( 9/30 ) of the classified qso hosts
were categorized as being disturbed " or strongly disturbed " in table [ tab : hostclass ] .
the 6 hosts classified as strongly disturbed " show asymmetric features reminiscent of tidal structure , as might be due , for example , to the advanced merger of two spiral disks .
putative tidal arms " are clearly seen in pg0157 + 001 , pg 1411 + 442 , pg 1613 + 658 , pg 1700 + 518 and pg 2349@xmath5014 , while pg 1426 + 215 has a clearly asymmetric disk .
the three additional objects listed as disturbed " generally show fainter asymmetric features , ( plausibly faint tidal debris ) whose effect on the 1-d radial profile is small enough to still allow us to distinguish between different host types .
the 4 disturbed " hosts include two `` d@xmath56 '' and one `` e '' . in addition to the 9 hosts classified as either disturbed " or strongly disturbed " , there are 3 hosts that were found to have a prominent well - defined bar , whose size in each case was @xmath25kpc radius . in two of these cases
the bar actually dominated the integrated luminosity of the host .
the hosts of two of these barred galaxies ( pg0804 + 761 , pg1001 + 054 ) , were classified as having a dominant disk profile ( d@xmath56 " ) , while the third ( pg1229 + 204 ) , was classified as b+d " .
it is unclear whether these bars have any relation to the processes that produced the putative tidal features in the disturbed " and strongly disturbed " hosts , but if they do , then the percentage of such hosts rises to 40% ( 12/30 ) .
our finding of 30% for the fraction of strongly disturbed " + disturbed " hosts ( 40% if well - defined bars " are included ) is similar to the range of values , typically 25%-50% , quoted for morphologically peculiar " , and/or interacting " hosts in previous surveys of low - z qsos ( e.g. , @xcite ) .
although it is not easy to determine if the definitions used to define degree of disturbance in previous surveys are the same as used here , a comparison of those objects from our current survey which have been included in previous surveys suggests that we are in fact using similar criteria . the breakdown by host type " of disturbed " , strongly disturbed " and well defined bar " systems is given in fig . [ fig : hosttype ] .
only one such object ( pg 2214 + 139 ) with obvious concentric shells " is classified as an elliptical .
the remainder of the disturbed " and well defined bar " systems have obvious disks , as presumably do the strongly disturbed " objects ( see above ) .
taken together , 11 of these 17 systems ( 65% ) have features which are interpreted as being indicative of obviously disturbed , non - axisymmetric structure .
such structure has in the past been seen as evidence for interaction / mergers as a possible trigger for producing the qso .
the breakdown by @xmath77(host ) and @xmath89(total ) of disturbed " , strongly disturbed " and well defined bar " systems is given in fig .
[ fig : mhhmb_class ] . in sum
, there appears to be no clear evidence for a correlation with either @xmath77(host ) or @xmath89(total ) when these three classes are considered together .
however , if we consider only the strongly disturbed " systems , there _ is _ a clear preference for these to be found in the most luminous ( i.e. , @xmath101 ) hosts .
in addition to asymmetric structures , we also found a small , bright knot in the inner disks of 9 hosts ( see notes in table [ tab : hostclass ] ) .
these knots were typically within a few kpc ( projected distance ) of the qso .
they are being discussed here since we believe that the evidence points to the knots being associated directly with the host galaxy , as opposed , for example , to being projected foreground or background objects at much larger distances from the qso .
the terms companion " or close companion " have been widely used in the literature to describe a range of distance scales around qsos where _ galaxies _ at the same redshift as the qso are found .
for example comparative studies of the environment around qsos typically find that , at low redshifts , qsos live in regions of moderate galaxy overdensity , typical of compact or small groups ( e.g. , @xcite ) .
closer in , typically within a projected nuclear separation of @xmath102kpc between the companion and qso nuclei , are numerous reports of close companion galaxies , clearly distinguishable as a separate galaxy , but often with reported evidence of a tidal interaction with the qso host ( e.g. , @xcite ) . at the closest distances ( typically nuclear separations @xmath103kpc ) , where the putative close companion galaxy strongly overlaps the qso host , as is the case in most of the strongly disturbed systems , including our own , the close companion is then plausibly the second object involved in an advanced merger with the qso host .
this last case is similar to what is meant here by the term close companion " , except in our case we also clearly distinguish a relatively bright compact knot " which could either be the nucleus of the companion galaxy , _ or _ possibly a super star - cluster , perhaps triggered by the merger .
we have referred to these objects ( the ones within 10kpc projected separation ) as close compact companions " even though redshift is available for only one of the 9 knots " , since it is likely that these knots " are indeed at the same redshift as the qso rather than simply being unrelated foreground or background objects .
five of the 9 qso hosts with
close compact companions " are the same 5 hosts classified as being strongly disturbed " , and one of the remaining four is classified as being disturbed " .
several of the companions are also observed to be strongly elongated ( e.g. , in pg 1119 + 120 , pg 1411 + 441 , pg 1426 + 215 , pg 2349@xmath5014 ) .
these facts strongly suggest that the knots are directly related to a merger event .
however we remain somewhat agnostic whether these objects represent the second nucleus of a nearly merged galaxy pair , or , instead , a super star cluster perhaps triggered by the merger .
although , the knots " appear to be too bright to be normal " star - formation knots , and are therefore good candidates to be distinct galaxy nuclei , they are in fact similar in brightness to the putative super star cluster " knots that have often been found in the circumnuclear regions of infrared - excess qsos " ( e.g. , @xcite ) and that appear to be nearly ubiquitous in the inner disks of warm ultraluminous infrared galaxies " ( e.g. @xcite ) .
all but one of the knots " seen in our current qso sample have measured @xmath77 in the range @xmath104 to @xmath105 , and in each case this translates into an absolute @xmath0-band luminosity in the range 3.55.5mag fainter than the host galaxy , and the most luminous knot ( @xmath106 ) , is still 3.3mag fainter than its host ( pg1302@xmath5102 ) , and is still within the upper range of luminosities observed for the star forming knots in warm uligs " .
in addition to trying to better categorize the host types for our sample of qsos , we were also interested in investigating their mid- and far - infrared emission properties , and in particular to see if a relationship exists between the infrared luminosity of the qso and morphology of the host .
interest in the infrared properties of qsos has been fueled by suggestions of an evolutionary connection between qsos and ultraluminous infrared galaxies ( ulirgs ) , where both types of objects represent different phases in the end stage of the merger of two relatively equal mass gas - rich spirals ( see review by @xcite ) .
key to such a connection has been the suggestion that those qsos with the strongest mid / far-infrared excess " emission seem to have disturbed " hosts ( e.g. , @xcite ) , whose properties are somewhat similar to the hosts of warm " ulirgs . for the qsos , infrared - excess "
has been defined using the mean radio - to - xray spectral energy distributions ( seds ) of qsos @xcite .
the data have been interpreted as showing that the sed is dominated by two thermal emission bumps " .
in addition to the well - known blue bump " ( bb ) of optical / uv emission at wavelengths @xmath107@xmath39 m , thought to be associated with thermal emission from an accretion disk , there also exists an infrared / submillimeter bump " at wavelengths @xmath108@xmath39 m , which is interpreted as being primarily due to dust re - radiation of emission from either the accretion disk or from embedded circumnuclear star formation .
the infrared excess " is then defined to be the ratio of luminosities in the two bumps " , @xmath15 .
the recent availability of reprocessed iso observations of qsos @xcite , plus the release of the _ iras _ faint source catalog @xcite , have now made it possible to more accurately estimate the strength as well as the color temperature of the `` infrared / submillimeter bump '' for the bqs qsos .
in particular , we have been able for the first time to compile fluxes in four wavebands ( 12@xmath39 m , 25@xmath39 m , 60@xmath39 m and 100@xmath39 m ) for the majority ( 18/32 ) of our bqs sources , and in 2 - 3 of these wavebands for an additional 10 sources .
three sources have only one flux measurement , but with sensitive upper limits in the remaining 3 wavebands .
only one source has just upper limits in all four wavebands .
these measurements are listed here in table [ tab : rad_prop ] where they have been used to compute an infrared luminosity , @xmath109 m , and the luminosity ratio , @xmath110 , for each source following the definitions and prescriptions found in table 1 of @xcite . using the data in table [ tab : rad_prop ] , we compare the infrared / optical ratio , @xmath111 , of the qso hosts versus host type in fig .
[ fig : lirlbb_class ] , and tabulate the mean and median values of @xmath112 and median @xmath111 , for different host type and degree of disturbance in table [ tab : hostclassstat ] .
figure [ fig : lirlbb_class ] suggests a fairly strong correlation between @xmath15 and host type , with nearly all ellipticals ( e " ) having the lowest observed values of @xmath111 .
table [ tab : hostclassstat ] shows that es have median values of @xmath111 slightly less than half that of hosts with detectable disks ( d@xmath56 " + b+d " ) .
the previously observed correlation between non - disturbed hosts and e host type then implies ( as shown in table [ tab : hostclassstat ] ) that the median @xmath111 ratio for disturbed " and strongly disturbed " hosts is also approximately twice the median ratio for non - disturbed " hosts .
a strong correlation between ir luminosity and host morphology had previously been suggested by @xcite , although a close comparison between his
quiet / disturbed " classification and the images presented here shows that the images he used lacked the angular resolution and/or depth to reliably classify the host galaxies : for example , pg1411 + 442 and pg1700 + 518 , two of the most disturbed objects in our sample , were classified as `` quiet '' , while pg0923 + 201 and pg0844 + 349 were classified as disturbed .
another method of characterizing the mid / far - infrared emission of qsos has been to use the mid / far - infrared color temperature ( e.g. , @xcite ) . the color - color plot shown in fig .
[ fig : ircolcol ] adopts the same definitions used in @xcite , who first pointed out that nearly all ( 8/9 ) of the transition qso " they studied
, objects with an intermediate position " in the far - infrared color - color diagram , in - between the empirical location of optically - selected qsos and far - infrared - selected ulirgs were undergoing strong tidal interactions indicative of major mergers .
our new results are similar , in that all 8 of our disturbed " and strongly disturbed " qsos for which we could reliably compute far - infrared colors , fall in the transition region " between the two limits marked by the lines denoting power - law " and black - body " emission .
additionally , we find only objects we classified as non - disturbed " above the power - law " continuum line . we interpret this as additional evidence for a positive correlation between fading evidence of tidal structure ( i.e. , increasing age of the merger ) , and decreasing @xmath15 , since the latter is correlated with increasing far - infrared color temperature .
we have used ao imaging on the 8.1 m gemini - north and 8.2 m subaru telescopes to acquire near - infrared ( mostly @xmath0-band ) images of an unbiased sample of 32 nearby ( @xmath1 ) optically selected qsos from the pgbqs in order to investigate the luminosity and morphology of the host galaxies , and to compare the properties of the host with properties of the qso .
the @xmath10-band absolute magnitudes of our qso targets span a factor of @xmath223 , from @xmath113 ( the minimum threshold for bona - fide qsos " according to the definition given by @xcite ) to @xmath114 .
three sources have absolute luminosities slightly below the minimum threshold and are traditionally classified as seyfert1 nuclei " .
given that this threshold is somewhat arbitrary , we have included these three objects in the current study .
following the definitions of @xcite , twenty - seven of our sources were radio - quiet and 5 were radio - loud . 1 . the mean @xmath0-band magnitude ( and corresponding @xmath4 ) of the host galaxies for our sample of qsos was @xmath3 ( @xmath115 ) , with a range of @xmath116 to @xmath117 ( @xmath118 ) .
2 . for the sample of 28 classified hosts ,
the distribution of best - fit " host types was 10 ( 36% ) ellipticals ( e " ) , 3 ( 11% ) bulge+disk ( b+d " ) , 8 ( 29% ) disk dominated ( d@xmath56 " ) , and 7 ( 25% ) of indeterminate type .
3 . there appears to be a strong correlation between host type and the @xmath0-band absolute magnitude of the host .
sub-@xmath4 hosts all have a dominant ( d@xmath56 " ) or strong disk component ( b+d " ) ; @xmath119 hosts appear to be equally divided between ellipticals ( e " ) and disks ( d@xmath56 " or b+d " ) ; @xmath120 hosts are mostly ellipticals ( e " ) . a similar , but
somewhat weaker trend is found for the @xmath10-band absolute magnitude of the qso .
4 . thirty percent ( 9/30 ) of our detected hosts were classified as either being disturbed " ( 3 ) or strongly disturbed " ( 6 ) .
an additional 3 objects were found to have strong well defined bars , bringing the total fraction of objects with obvious disturbed and/or non - axisymmetric structure up to 40% ( 12/30 ) . in total , these objects are relatively equally distributed over all values of @xmath8 , and @xmath89(total ) , although in detail , the strongly disturbed " objects are much more likely to be found among the most luminous hosts ( i.e. @xmath101 ) , with disturbed " and well defined bars " more likely to be associated with less luminous hosts ( i.e. @xmath121 ) .
5 . for the subsample of 18 qsos with redshifts @xmath9 , which is the redshift completeness limit corresponding to the @xmath122 limit of the pgbqs @xcite
, we find _ no _ obvious correlation between the absolute magnitude of the qso , @xmath89(total ) , and the absolute magnitude of the host galaxy , @xmath8 .
however at @xmath12 , all 8 sources in our sample have the largest host luminosities , @xmath123 ( @xmath124 ) , _ and _ the largest qso luminosities , @xmath125 , suggesting that the most luminous qsos are found among the most luminous hosts . 6 .
the mean properties of the hosts of radio - loud quasars ( rlqs ) are clearly different from the mean properties of the hosts of radio - quiet qsos ( rqqs ) .
all 5 of the rlqs are among the 10 sources in our sample at @xmath126 .
all 5 rlqs are among the 10 most luminous qsos ( i.e. @xmath98 ) , and for the 4 rlqs where we were able to reliable measure the host , all 4 are among the 5 most luminous hosts ( i.e. @xmath127 ) .
three of the 5 rlq hosts are classified as ellipticals , one host suffers from a bad psf but otherwise shows no visible sign of disk structure , and one is strongly disturbed and can not be classified .
these results are in keeping with the paradigm which suggest that rlqs are preferentially found in giant elliptical hosts with the most luminous qsos . 7 .
we find a strong correlation between the infrared excess , @xmath15 , of the qso and its host type and host degree of disturbance .
this ratio is twice as large , on average , in hosts with strong disk components as in elliptical hosts , and likewise is twice as large , on average , in disturbed " + strongly disturbed " hosts as in non - disturbed " hosts .
we also find that disturbed " and strongly disturbed " hosts have mid / far - infrared colors that place them in a region of the far - infrared color - color plane in between the power - law that is characteristic of the majority of qsos , and the black - body that is characteristic of the majority of ulirgs .
our results concerning the mean and range of host luminosities , the general breakdown of host types , the fraction of objects with disturbed features , and the differences in the mean properties of the hosts of rlqs versus rqqs , are in general agreement with previous studies of the host properties of low - z qsos .
our mean and range of @xmath0-band host luminosities are almost identical to the values reported earlier by @xcite and @xcite , respectively , in their studies of a similar sample of pgqsos selected from the bqs @xcite .
one advantage of our ao observations with 8m - class telescopes is that we have perhaps been able to give more emphasis to trying to accurately characterize the host properties of individual sources , and to search for systematic trends in host type versus other properties of the qso .
perhaps the most interesting new results from our survey are those concerning the detailed distribution of disturbed hosts versus the luminosity of the host and the luminosity of the qso , and the apparent correlation of infrared excess with host type .
our results suggest that the percentage of disturbed hosts is significant ( 15 - 30% ) across the range of observed host luminosities ( @xmath128 ) , even though the main host type clearly changes from hosts with prominent disks to hosts which appear to be luminous ellipticals .
whether disturbed hosts represent a particular phase is the lifetime of virtually all qsos or whether they represent an inherently separate class of qsos is somewhat unclear .
the answer to this question depends on the answers to two subsidiary questions : ( 1 ) what is the time scale for qso activity relative to that for the persistence of recognizable signs of disturbance due to a strong interaction or merger ? and ( 2 ) what is the true space density of qsos with bolometric luminosities above some specific value and the fraction of these that show some level of disturbance indicating a recent merger ?
our study has some bearing on this second question .
however , it now seems likely that uv - selected optical qso samples , such as the one studied here , may represent considerably less than half of the total qso population at a comparable bolometric luminosity .
in addition , recent work by @xcite comparing qsos found in the sloan digital sky survey with the pg sample shows that the pg sample is seriously incomplete even in terms of its own specified selection criteria .
this incompleteness suggests two imperatives for future work .
first , it seems essential to develop a sample that is substantially complete to a specific bolometric qso luminosity over some large volume .
this effort will necessarily require complementary techniques covering wavelength regimes at least from x - rays to the far ir .
second , it will be necessary to carry out imaging studies similar to those we have done for the pgbqs on representative subsamples of qsos identified by these other techniques , for it is quite likely that , for example , the fraction showing signs of mergers may vary for samples selected by different criteria . a recent step in this direction
has been taken by @xcite for qsos with @xmath129 found in the 2mass survey , for which they find an extraordinarily high fraction ( @xmath17% ) show disturbances indicating tidal interactions .
we thank the referee for a careful reading of the paper and useful comments to help improve both its content and its presentation .
we also thank gabriela canalizo for a helpful conversation .
the hokupaa ao observations were supported by members of the university of hawaii ao group : p. baudoz , o. guyon and d. potter .
support for hokupaa came from the national science foundation under grant no .
ast 96 - 18852 .
our results were based in part on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the particle physics and astronomy research council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , cnpq ( brazil ) and conicet ( argentina ) .
the results were also based in part on data collected at subaru telescope , which is operated by the national astronomical observatory of japan .
andreani , p. , franceschini , a. , & granato , g. 1999 , , 306 , 161 bahcall , j.n . , kirhakos , s. , saxe , d. h. , & schneider , d. p. 1997
, , 479 , 642 balzano , v. a. , & weedman , d. w. , , 243 , 756 barthel , p. b. , & sanders , d. b. eds .
, 2006 , qso host galaxies : evolution and environment , , in press bothun , g. d. , romanishin , w. , strom , s. e. , & strom , k. m. 1984 , , 89 , 1300 boyce , p. j. , disney , m. j. , & bleaken , d. g. 1999 , , 302 , l39 boyce , p. j. et al .
1996 , , 473 , 760 canalizo , g. , & stockton , a. 2000 , , 120 , 1750 canalizo , g. , & stockton , a. 2001 , , 555 , 719 clements , d.l . 2000 , , 311 , 833 clements , d. l. , & prez - fournon , eds .
2001 , quasar hosts , ( berlin : springer ) , 336pp cutri , r. m. , nelson , b. o. , francis , p. j. , & smith , p. s. 2002 , in agn surveys , eds .
r. green , e. khachikian , d. sanders , ( san francisco : asp conf . ser . ) , 284 , 127 disney , m. j. , et al .
1995 , , 376 , 150 dunlop , j. s. , taylor , g. l. , hughes , d. h. , & robson , e. i. 1993 , , 264 , 455 dunlop , j. s. , mclure , r. j. , kukula , m. j. , baum , s. a. , odea , c. p. , & hughes , d. h. 2003 , , 340 , 1095 elvis , m. et al .
1994 , , 95 , 1 giveon , u. , maoz , d. , kaspi , s. , netzer , h. , & smith , p. s. 1999 , , 306 , 637 graves , j. e. , northcott , m. j. , roddier , f. j. , roddier , c. a. , close , l. m. 1998 , , 3353 , 34 haas , m. , klaas , u. , coulson , i. , thommes , e. , & xu , c. 2000 , , 356 , 83 haas , m. , et al . 2003 , , 402 , 87 hamilton , t. s. , casertano , s. , & turnshek , d. a. 2002 , , 576 , 61 heckman , t. m. , bothun , g. d. , balick , b. , & smith , e. p. 1984
, , 89 , 958 hines , d. c. , low , f. j. , thompson , r. i. , weymann , r. j. , & storrie - lombardi , l. j. 1999 , , 512 , 140 hutchings , j. b. , crampton , d. , & campbell , b. 1984 , , 280 , 41 hutchings , j. b. , janson , t. , & neff , s. g. 1989 , , 342 , 660 hutchings , j. b. , & crampton , d. 1990 , , 99 , 37 hutchings , j. b. , & neff , s. g. 1991 , , 101 , 434 hutchings , j. b. , & neff , s. g. 1992 , , 104 , 1 hutchings , j. b. , & neff , s. g. 1996 , , 113 , 550 hutchings , j. b. , maddox , s. , cutri , r. m. , & nelson , b. o. 2003 , , 126 , 63 hyland , a. r. , & allen , d. a. 1982 , , 199 , 943 jahnke , k. , & wisotzki , l. 2003 , , 346 , 304 jester , s. , et al .
2005 , , 130 , 873 kellermann , k. i. , sramek , r. a. , schmidt , m. , green , r. f. , & shaffer , d. b. 1994 , , 108 , 1163 lipari , s. 1994 , , 436 , 102 low , f. j. , cutri , r. m. , huchra , j. p. , & kleinmann , s. g. 1988 , , 327 , l41 mackenty , j. w. , & stockton , a. 1984 , , 283 , 64 mrquez , i. , masegosa , j. , del olmo , a. , lara , l. , garca , e. , & molina , j. eds . , 2001 , qso hosts and their environments , ( new york : kluwer ) , 379pp mrquez , i. , petitjean , p. , thodore , b. , bremer , m , monnet , g. , & beuzit , j .- l .
, , 371 , 97 mcalary , c. w. , mclaren , r. a. , mcgonegal , & r. j. , maza , j. 1983 , , 52 , 341 mcleod k. k. , & mcleod b. a. 2001 , , 546 , 782 mcleod , k. k. , & rieke , g. h. 1994a , , 420 , 58 mcleod , k. k. , & rieke , g. h. 1994b , , 431 , 137 mcleod , k. k. , & rieke , g. h. 1995 , , 454 , l77 mcleod , k. k. , rieke , g. h. , & storrie - lombardi , l. j. 1999 , , 511 , l67 mclure , r. j. , kukula , m. j. , dunlop , j. s. , baum , s. a. , odea , c. p. , & hughes , d. h. 1999 , , 308 , 377 miller , j. s. , & sheinis , a. i. 2002 , , 588 , l9 mobasher , b. , sharples , r. m. , & ellis , r. s. 1993 , , 263 , 560 moshir , m. , et al .
1992 , explanatory supplement to the _ iras _ faint source catalog , version 2 ( jpl d-10015 ) ( pasadena : jpl ) neugebauer , g. , oke , j. b. , becklin , e. e. , & matthews , k. 1979 , , 230 , 79 neugebauer , g. , green , r. f. , matthews , k. , schmidt , m. , soifer , b. t. , & bennett , j. 1987 , , 63 , 615 neugebauer , g. , soifer , b. t. , matthews , k. , & elias , j.h .
1989 , , 97 , 957 percival , w. j. , miller , l. , mclure , r. j. , & dunlop , j. s. 2001 , , 322 , 843 polletta , m. , courvoisier , t. j .-
l . , hooper , e. j. , & wilkes , b. j. 2000 , , 362 , 75 rudy , r. j. , levan , p. d. , & rodriguez - espinosa , j. m. 1982 , , 87 , 598 sanders , d. b. , soifer , b. t. , elias , j. h. , madore , b. f. , matthews , k. , neugebauer , g. , & scoville , n. z. 1988 , , 325 , 74 sanders , d. b. , phinney , e. s. , neugebauer , g. , soifer , b. t. , & matthews , k. 1989 , , 347 , 29 sanders , d. b. , & mirabel , i. f. 1996 , , 34 , 749 schmidt , m. , & green , r. f. 1983 , , 269 , 352 scoville , n. z. , frayer , d. t. , schinnerer , e. , & christopher , m. 2003 , , 585 , l105 sitko , m. l. , stein , w. a. , zhang , y .- x .
, & wisniewski , w. z. 1982 , , 259 , 486 smith , e. p. , heckman , t. m. , bothun , g. d. , romanishin , w. , & balick , b. 1986 , , 306 , 64 staguhn , j. g. , schinnerer , e. , eckart , a. , & scharw " achter , j. 2004 , , 609 , 85 stockton , a. 1982 , , 257 , 33 stockton , a. , canalizo , g. , & close , l. 1998 , , 500 , l121 stockton , a. , & mackenty , j. w. 1983 , , 305 , 678 stockton , a. , & mackenty , j. w. 1987 , , 316 , 584 surace , j. a. , sanders , d. b. , vacca , w. d. , veilleux , s. , & mazzarella , j. m. 1998 , , 492 , 116 surace , j. a. , & sanders , d. b. 1999 , , 512 , 162 surace , j. a. , & sanders , d. b. 2000 , , 120 , 604 surace , j. a. , sanders , d. b. , & evans , a. s. 2001 , , 122 , 2791 takami , h. , et al . 2004 , , 56 , 225 taylor , g. l. , dunlop , j. s. , hughes , d. h. , & robson , e. i. 1996 , , 283 , 930 tokunaga , a. t. , simons , d. a. , & vacca , w. d. 2002 , , 114 , 180 veilleux , s. , et al .
2005 , apj , submitted veron - cetty , m .-
p . , & woltjer , l. 1990 , , 236 , 69 yee , h. k. c. , & green , r. f. 1984 , , 280 , 79 zheng , z. , wu , h. , mao , s. , xia , x .- y . ,
deng , z .- g .
, & zou , z .-
1999 , , 349 , 735 l c c c c c pg0026 + 129 & 00 29 13.81 + 13 16 04.5 & 0.142 & @xmath130 & @xmath131 & q + pg0050 + 124 & 00 53 35.08 + 12 41 34.4 & 0.061 & @xmath132 & @xmath133 & q + pg0157 + 001 & 01 59 50.24 + 00 23 40.9 & 0.163 & @xmath134 & @xmath135 & q + pg0804 + 761 & 08 10 58.55 + 76 02 41.8 & 0.100 & @xmath136 & @xmath137 & q + pg0838 + 770 & 08 44 45.63 + 76 53 09.4 & 0.131 & @xmath138 & @xmath139 & q + pg0844 + 349 & 08 47 42.51 + 34 45 04.6 & 0.064 & @xmath140 & @xmath141 & q + pg0923 + 201 & 09 25 54.87 + 19 54 06.7 & 0.190 & @xmath142 & @xmath143 & q + pg0953 + 414 & 09 56 52.46 + 41 15 22.0 & 0.234 & @xmath144 & @xmath145 & q + pg1001 + 054 & 10 04 20.15 + 05 13 00.3 & 0.160 & @xmath146 & @xmath147 & q + pg1004 + 130 & 10 07 26.11 + 12 48 56.2 & 0.240 & @xmath148 & @xmath149 & l + pg1116 + 215 & 11 19 08.70 + 21 19 18.0 & 0.177 & @xmath150 & @xmath151 & q + pg1119 + 120 & 11 21 47.17 + 11 44 18.1 & 0.050 & @xmath152 & @xmath153 & q + pg1126@xmath5041 & 11 29 16.73 @xmath504 24 07.8 & 0.060 & @xmath154 & @xmath155 & q + pg1211 + 143 & 12 14 17.67 + 14 03 12.4 & 0.081 & @xmath156 & @xmath157 & q + pg1229 + 204 & 12 32 03.67 + 20 09 29.1 & 0.063 & @xmath158 & @xmath159 & q + pg1302@xmath5102 & 13 05 33.01 @xmath510 33 19.4 & 0.278 & @xmath160 & @xmath161 & l + pg1307 + 085 & 13 09 47.03 + 08 19 49.3 & 0.155 & @xmath142 & @xmath162 & q + pg1309 + 355 & 13 12 17.74 + 35 15 21.3 & 0.184 & @xmath158 & @xmath163 & q + pg1351 + 640 & 13 53 15.85 + 63 45 44.9 & 0.088 & @xmath164 & @xmath165 & q + pg1411 + 442 & 14 13 48.47 + 44 00 13.6 & 0.090 & @xmath166 & @xmath167 & q + pg1426 + 015 & 14 29 06.67 + 01 17 06.3 & 0.086 & @xmath168 & @xmath169 & q + pg1435@xmath5067 & 14 38 16.25 @xmath506 58 21.5 & 0.126 & @xmath170 & @xmath165 & q + pg1440 + 356 & 14 42 07.54 + 35 26 22.9 & 0.079 & @xmath171 & @xmath172 & q + pg1613 + 658 & 16 13 57.27 + 65 43 09.7 & 0.129 & @xmath173 & @xmath174 & q + pg1617 + 175 & 16 20 11.29 + 17 24 27.6 & 0.112 & @xmath158 & @xmath175 & q + pg1626 + 554 & 16 27 56.03 + 55 22 31.2 & 0.133 & @xmath168 & @xmath176 & q + pg1700 + 518 & 17 01 24.87 + 51 49 21.0 & 0.292 & @xmath144 & @xmath177 & q + pg2130 + 099 & 21 32 27.92 + 10 08 18.7 & 0.063 & @xmath164 & @xmath178 & q + 4c 31.63 & 22 03 15.98 + 31 45 38.3 & 0.295 & @xmath179 & @xmath180 & l + pg2214 + 139 & 22 17 12.35 + 14 14 21.2 & 0.066 & @xmath181 & @xmath182 & q + pg2251 + 113 & 22 54 10.50 + 11 36 39.0 & 0.326 & @xmath183 & @xmath184 & l + pg2349@xmath5014 & 23 51 56.13 @xmath501 09 13.4 & 0.174 & @xmath185 & @xmath186 & l + l c c c c c pg0026 + 129 & 2001 jun 27 & 2670 & @xmath187 & 0.13 & gemini + pg0050 + 124 & 2000 nov 17 & ( 1200 ) & @xmath187 & 0.32 & gemini + & 2000 nov 17 & ( 1620 ) & @xmath40 & 0.23 & gemini + & 2000 nov 19 & ( 1560 ) & @xmath187 & 0.16 & gemini + & 2000 nov 19 & ( 1200 ) & @xmath40 & 0.15 & gemini + & 2000 nov 20 & ( 2300 ) & @xmath187 & 0.23 & gemini + & 2000 nov 20 & ( 1130 ) & @xmath40 & 0.21 & gemini + pg0157 + 001 & 2001 sep 14 & 4050 & @xmath187 & 0.19 & gemini + pg0804 + 761 & 2000 dec 21 & 2910 & @xmath187 & 0.15 & gemini + & 2000 dec 21 & 2620 & @xmath40 & 0.15 & gemini + pg0838 + 770 & 2003 feb 12 & 1460 & @xmath187 & 0.32 & subaru + pg0844 + 349 & 2001 oct 10 & 600 & @xmath187 & 0.38 & gemini + pg0923 + 201 & 2001 feb 21 & 3930 & @xmath187 & 0.20 & gemini + & 2001 feb 23 & ( 5700 ) & @xmath188 & 0.17 & gemini + pg0953 + 414 & 2000 dec 20 & 4590 & @xmath40 & 0.13 & gemini + & 2001 feb 23 & ( 1380 ) & @xmath188 & 0.23 & gemini + pg1001 + 054 & 2000 dec 26 & 4140 & @xmath187 & 0.20 & gemini + pg1004 + 130 & 2003 feb 12 & 2400 & @xmath187 & 0.16 & subaru + pg1116 + 215 & 2002 apr 21 & ( 3030 ) & @xmath187 & 0.18 & gemini + pg1119 + 120 & 2001 feb 23 & ( 3960 ) & @xmath187 & 0.15 & gemini + pg1126@xmath5041 & 2001 feb 21 & 4950 & @xmath187 & 0.15 & gemini + pg1211 + 143 & 2002 apr 21 & ( 2970 ) & @xmath187 & 0.17 & gemini + pg1229 + 204 & 2002 apr 23 & ( 630 ) & @xmath187 & 0.24 & gemini + pg1302@xmath5102 & 2003 feb 12 & 1620 & @xmath187 & 0.15 & subaru + pg1307 + 085 & 2003 feb 12 & 2420 & @xmath187 & 0.15 & subaru + pg1309 + 355 & 2002 apr 23 & ( 2490 ) & @xmath187 & 0.22 & gemini + pg1351 + 640 & 2002 apr 21 & ( 2100 ) & @xmath40 & 0.23 & gemini + pg1411 + 442 & 2001 feb 23 & ( 1800 ) & @xmath187 & 0.13 & gemini + & 2001 feb 23 & ( 1620 ) & @xmath188 & 0.15 & gemini + & 2001 feb 25 & 3870 & @xmath187 & 0.17 & gemini + & 2001 apr 25 & 2730 & @xmath40 & 0.13 & gemini + pg1426 + 015 & 2001 apr 24 & ( 4080 ) & @xmath40 & 0.18 & gemini + pg1435@xmath5067 & 2001 feb 23 & ( 2100 ) & @xmath188 & 0.21 & gemini + pg1440 + 356 & 2001 jun 28 & ( 960 ) & @xmath40 & 0.16 & gemini + pg1613 + 658 & 2001 feb 25 & 4500 & @xmath40 & 0.30 & gemini + & 2001 apr 25 & 3420 & @xmath188 & 0.18 & gemini + & 2001 apr 27 & 1380 & @xmath187 & 0.18 & gemini + & 2001 apr 27 & 1110 & @xmath40 & 0.28 & gemini + pg1617 + 175 & 2002 apr 21 & ( 4140 ) & @xmath40 & 0.23 & gemini + pg1626 + 554 & 2001 apr 27 & 1170 & @xmath40 & 0.18 & gemini + & 2001 jun 27 & 4530 & @xmath187 & 0.16 & gemini + pg1700 + 518 & 2001 sep 12 & 990 & @xmath187 & 0.13 & gemini + & 2001 sep 14 & 3270 & @xmath187 & 0.14 & gemini + & 2001 oct 10 & 1350 & @xmath40 & 0.15 & gemini + pg2130 + 099 & 2001 jun 28 & ( 3180 ) & @xmath40 & 0.18 & gemini + & 2001 oct 10 & 1080 & @xmath187 & 0.15 & gemini + 4c 31.63 & 2001 jun 27 & 3540 & @xmath187 & 0.13 & gemini + pg2214 + 139 & 2001 sep 12 & 2070 & @xmath187 & 0.17 & gemini + pg2251 + 113 & 2001 sep 14 & 2910 & @xmath187 & 0.14 & gemini + pg2349@xmath5014 & 2001 sep 12 & 3540 & @xmath187 & 0.13 & gemini + .
l c c c c c c pg0026 + 129 & @xmath187 & @xmath189 & @xmath190 & & @xmath191 & @xmath191 + pg0050 + 124 & @xmath188 & @xmath192 & & & & + pg0050 + 124 & @xmath187 & @xmath193 & & & & + pg0050 + 124 & @xmath40 & @xmath194 & & & & + pg0157 + 001 & @xmath187 & @xmath195 & @xmath196 & & @xmath197 & @xmath197 + pg0804 + 761 & @xmath187 & @xmath198 & @xmath199 & & @xmath200 & @xmath200 + pg0804 + 761 & @xmath40 & @xmath201 & @xmath202 & & @xmath203 & @xmath203 + pg0838 + 770 & @xmath187 & @xmath204 & @xmath205 & & @xmath206 & @xmath206 + pg0844 + 349 & @xmath187 & @xmath207 & @xmath208 & & @xmath209 & @xmath209 + pg0923 + 201 & @xmath188 & @xmath210 & & & @xmath211 & @xmath212 + pg0923 + 201 & @xmath187 & @xmath213 & & & @xmath214 & @xmath214 + pg0953 + 414 & @xmath188 & @xmath215 & @xmath216 & & @xmath217 & @xmath218 + pg0953 + 414 & @xmath40 & @xmath219 & @xmath220 & & @xmath221 & @xmath221 + pg1001 + 054 & @xmath187 & @xmath222 & @xmath220 & @xmath223 & @xmath224 & @xmath224 + pg1004 + 130 & @xmath187 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath228 + pg1116 + 215 & @xmath187 & @xmath229 & & & @xmath230 & @xmath231 + pg1119 + 120 & @xmath187 & @xmath232 & @xmath233 & & @xmath234 & @xmath235 + pg1126@xmath5041 & @xmath187 & @xmath236 & & & @xmath237 & @xmath237 + pg1211 + 143 & @xmath187 & @xmath238 & @xmath239 & & @xmath240 & @xmath241 + pg1229 + 204 & @xmath187 & @xmath242 & @xmath243 & & @xmath244 & @xmath245 + pg1302@xmath5102 & @xmath187 & @xmath246 & & @xmath247 & @xmath248 & @xmath248 + pg1307 + 085 & @xmath187 & @xmath249 & @xmath250 & & @xmath251 & @xmath251 + pg1309 + 355 & @xmath187 & @xmath252 & & & @xmath253 & @xmath254 + pg1351 + 640 & @xmath40 & @xmath255 & @xmath256 & & @xmath257 & @xmath258 + pg1411 + 442 & @xmath188 & @xmath259 & @xmath260 & & @xmath261 & @xmath262 + pg1411 + 442 & @xmath187 & @xmath263 & @xmath264 & & @xmath241 & @xmath241 + pg1411 + 442 & @xmath40 & @xmath265 & @xmath266 & & @xmath267 & @xmath267 + pg1426 + 015 & @xmath40 & @xmath268 & @xmath269 & & @xmath270 & @xmath271 + pg1435@xmath5067 & @xmath188 & @xmath272 & & & @xmath273 & @xmath274 + pg1440 + 356 & @xmath40 & @xmath275 & @xmath276 & & @xmath277 & @xmath278 + pg1613 + 658 & @xmath188 & @xmath279 & @xmath280 & & @xmath281 & @xmath281 + pg1613 + 658 & @xmath187 & @xmath282 & @xmath283 & & @xmath284 & @xmath284 + pg1613 + 658 & @xmath40 & @xmath285 & @xmath286 & & @xmath287 & @xmath287 + pg1617 + 175 & @xmath40 & @xmath268 & @xmath288 & & @xmath289 & @xmath271 + pg1626 + 554 & @xmath187 & @xmath290 & & & @xmath291 & @xmath291 + pg1626 + 554 & @xmath40 & @xmath292 & & & @xmath293 & @xmath293 + pg1700 + 518 & @xmath187 & @xmath294 & @xmath295 & & @xmath296 & @xmath296 + pg1700 + 518 & @xmath40 & @xmath297 & @xmath298 & & @xmath299 & @xmath299 + pg2130 + 099 & @xmath187 & @xmath300 & @xmath301 & & @xmath302 & @xmath302 + pg2130 + 099 & @xmath40 & @xmath303 & @xmath304 & & @xmath305 & @xmath306 + 4c 31.63 & @xmath187 & @xmath307 & @xmath308 & & @xmath309 & @xmath309 + pg2214 + 139 & @xmath187 & @xmath310 & & @xmath311 & @xmath312 & @xmath312 + pg2251 + 113 & @xmath187 & @xmath313 & @xmath280 & & @xmath314 & @xmath314 + pg2349@xmath5014 & @xmath187 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & @xmath318 + lccc pg0026 + 129 @xmath0 & 1.05 & 1.08 & 1.04 [ 0.96/0.04 ] + pg0157 + 001 @xmath0 & 3.79 & 4.26 & 3.12 [ 0.64/0.36 ] + pg0804 + 761 @xmath0 & 1.54 & 1.90 & 1.05 [ 0.82/0.18 ] + pg0804 + 761 @xmath40 & 1.01 & 1.10 & 0.95 [ 0.68/0.32 ] + pg0838 + 770 @xmath0 & 4.43 & 5.05 & 3.58 [ 0.78/0.22 ] + pg0844 + 349 @xmath0 & 2.33 & 3.07 & 2.31 [ 0.69/0.31 ] + pg0923 + 201 @xmath41 & 1.12 & 1.24 & 1.15 [ 0.89/0.11 ] + pg0923 + 201 @xmath0 & 1.07 & 1.24 & 1.07 [ 0.94/0.06 ] + pg0953 + 414 @xmath41 & 0.94 & 0.90 & 0.88 [ 0.32/0.68 ] + pg0953 + 414 @xmath40 & 0.92 & 0.93 & 0.91 [ 0.98/0.02 ] + pg1001 + 054 @xmath0 & 0.97 & 1.03 & 0.93 [ 0.48/0.52 ] + pg1004 + 130 @xmath0 & 1.13 & 1.79 & 1.06 [ 0.65/0.35 ] + pg1116 + 215 @xmath0 & 1.20 & 1.13 & 1.13 [ 0.00/1.00 ] + pg1119 + 120 @xmath0 & 11.38 & 10.86 & 6.57 [ 0.70/0.30 ] + pg1126@xmath5041 @xmath0 & 11.61 & 17.00 & 9.66 [ 0.64/0.36 ] + pg1211 + 143 @xmath0 & 1.28 & 1.28 & 1.30 [ 0.46/0.54 ] + pg1229 + 204 @xmath0 & 1.97 & 2.46 & 1.41 [ 0.65/0.35 ]
+ pg1302@xmath5102 @xmath0 & 2.20 & 2.46 & 2.00 [ 0.74/0.26 ] + pg1307 + 085 @xmath0 & 1.42 & 2.15 & 1.19 [ 0.74/0.26 ] + pg1309 + 355 @xmath0 & 0.98 & 1.28 & 0.97 [ 0.92/0.08 ] + pg1351 + 640 @xmath40 & 0.88 & 0.88 & 0.87 [ 0.48/0.52 ] + pg1411 + 442 @xmath41 & 2.08 & 2.05 & 2.00 [ 0.86/0.14 ] + pg1411 + 442 @xmath0 & 9.20 & 8.77 & 8.57 [ 0.61/0.39 ] + pg1411 + 442 @xmath40 & 1.48 & 2.26 & 1.47 [ 0.96/0.04 ] + pg1426 + 015 @xmath40 & 1.53 & 1.52 & 1.37 [ 0.91/0.09 ] + pg1435@xmath5067 @xmath41 & 0.98 & 1.04 & 0.96 [ 0.92/0.08 ] + pg1440 + 356 @xmath40 & 2.14 & 1.93 & 1.65 [ 0.84/0.16 ] + pg1613 + 658 @xmath0 & 4.30 & 6.67 & 3.95 [ 0.90/0.10 ] + pg1613 + 658 @xmath41 & 5.14 & 5.61 & 5.15 [ 0.99/0.01 ] + pg1613 + 658 @xmath40 & 6.15 & 13.17 & 5.97 [ 0.98/0.02 ] + pg1617 + 175 @xmath40 & 1.05 & 1.62 & 1.06 [ 0.98/0.02 ] + pg1626 + 554 @xmath0 & 0.89 & 0.86 & 0.84 [ 0.52/0.48 ] + pg1626 + 554 @xmath40 & 0.93 & 0.92 & 0.92 [ 0.79/0.21 ] + pg1700 + 518 @xmath0 & 5.48 & 5.52 & 5.38 [ 0.80/0.20 ] + pg1700 + 518 @xmath40 & 2.53 & 2.34 & 2.34 [ 0.00/1.00 ] + pg2130 + 099 @xmath0 & 4.57 & 3.13 & 3.28 [ 0.00/1.00 ] + pg2130 + 099 @xmath40 & 5.25 & 3.98 & 3.87 [ 0.42/0.58 ] +
4c 31.63 @xmath0 & 1.07 & 1.49 & 1.06 [ 0.84/0.16 ] + pg2214 + 139 @xmath0 & 1.79 & 3.26 & 1.49 [ 0.76/0.24 ] + pg2251 + 113 @xmath0 & 1.04 & 1.02 & 0.95 [ 0.53/0.47 ] + pg2349@xmath5014 @xmath0 & 4.31 & 5.70 & 4.13 [ 0.82/0.18 ] + l c c c c c c c l pg0026 + 129 & & ( @xmath48 ) & & no & no & no & e & + pg0050 + 124 & & & & no & yes & no & d@xmath56 & + pg0157 + 001 & & @xmath48+exp .
& & no & yes & str & ? & host asymmetry + pg0804 + 761 & & @xmath48 & @xmath48 & yes & no & no & d@xmath56 & well - defined bar + pg0838 + 770 & & @xmath48 & & yes & yes & yes & d@xmath56 & twisted bar+ring , well - defined bar + pg0844 + 349 & & @xmath48 & & no & yes & no & d@xmath56 & + pg0923 + 201 & @xmath48 & @xmath48 & & no & no & no & e & + pg0953 + 414 & ( exp ) & & ?
& no & no & & & host too faint to be classified + pg1001 + 054 & & ( @xmath48 ) & & yes & no & no & d@xmath56 & well - defined bar + pg1004 + 130 & & @xmath48 & & no & no & no & e & + pg1116 + 215 & & exp & & no & no & yes & d@xmath56 & cc , host asymmetry + pg1119 + 120 & & @xmath48+exp .
& & no & no & no & b+d & cc + pg1126@xmath5041 & & @xmath48+exp .
& & no & yes & no & b+d & cc + pg1211 + 143 & & ?
& & no & no & & & host too faint to be classified + pg1229 + 204 & & @xmath48+exp .
& & yes & yes & no & b+d & well - defined bar + pg1302@xmath5102 & & @xmath48 & & no & no & no & e & cc + pg1307 + 085 & & @xmath48 & & no & no & no & e & + pg1309 + 355 & & @xmath48 & & no & no & no & e & + pg1351 + 640 & & & ?
& no & no & no & ?
& + pg1411 + 442 & exp & @xmath48+exp & @xmath48 & no & ? & str & ? & cc + pg1426 + 015 & & & @xmath48 & no & ? & str & ? & cc + pg1435@xmath5067 & @xmath48 & & & no & no & no & e & + pg1440 + 356 & & & exp & no & yes & no & d@xmath56 & + pg1613 + 658 & @xmath48 & @xmath48 & @xmath48 & no & ? & str & ? & cc + pg1617 + 175 & & & @xmath48 & no & no & no & e & + pg1626 + 554 & & ( exp ) & ?
& no & no & no & & bad psf ; unreliable 2-d fit + pg1700 + 518 & & @xmath48 & exp & no & ? & str & ? & cc + pg2130 + 099 & & exp & @xmath48+exp & no & yes & no & d@xmath56 & + 4c 31.63 & & @xmath48 & & no & no & no & e & + pg2214 + 139 & & @xmath48 & & no & no & yes & e & concentric shells + pg2251 + 113 & & ? & & no & no & no & & bad psf ; unreliable 2-d fit + pg2349@xmath5014 & & @xmath48 & & no & ? & str & ?
& cc + lccccccccccccc pg0026 + 129 @xmath0 & 13.05 & @xmath525.93 & 14.46 & @xmath524.52 & 22.16 & 10.62 & 0.21 & 0.01 & & & & + pg0157 + 001 @xmath0 & 13.41 & @xmath525.90 & 13.30 & @xmath526.01 & 18.79 & 1.94 & 0.43 & 0.60 & 18.55 & 3.75 & 0.43 & 1.73 + pg0804 + 761 @xmath0 & 12.22 & @xmath525.94 & 14.24 & @xmath523.92 & & & & & 17.08 & 1.74 & 0.28 & 0.82 + pg0804 + 761 @xmath40 & 10.67 & @xmath527.49 & 14.44 & @xmath523.72 & & & & & 18.02 & 3.39 & 0.61 & 0.82 + pg0838 + 770 @xmath0 & 14.31 & @xmath524.48 & 13.89 & @xmath524.90 & & & & & 17.26 & 2.97 & 0.59 & 2.85 + pg0844 + 349 @xmath0 & 13.35 & @xmath523.82 & 13.65 & @xmath523.52 & & & & & 16.04 & 1.38 & 0.25 & 0.53 + pg0923 + 201 @xmath41 & 14.52 & @xmath525.16 & 15.46 & @xmath524.22 & 22.25 & 5.73 & 0.14 & 1.18 & & & & + pg0923 + 201 @xmath0 & 13.53 & @xmath526.15 & 14.73 & @xmath524.95 & 21.20 & 4.53 & 0.07 & 0.79 & & & & + pg0953 + 414 @xmath41 & 14.39 & @xmath525.80 & 15.25 & @xmath527.77 & 22.25 & 3.58 & 0.42 & 3.07 & 20.95 & 7.91 & 0.40 & 1.62 + pg0953 + 414 @xmath40 & 12.42 & @xmath524.94 & 13.86 & @xmath526.33 & 18.29 & 0.95 & 0.90 & 1.55 & 22.02 & 8.75 & 0.90 & 1.98 + pg1001 + 054 @xmath0 & 14.21 & @xmath525.06 & 16.68 & @xmath522.59 & & & & & 19.59 & 1.61 & 0.12 & 0.56 + pg1004 + 130 @xmath0 & 13.84 & @xmath526.41 & 14.12 & @xmath526.13 & 22.00 & 10.04 & 0.10 & 2.94 & & & & + pg1116 + 215 @xmath0 & 12.61 & @xmath526.90 & 15.96 & @xmath523.55 & & & & & 17.50 & 0.99 & 0.33 & 0.04 + pg1119 + 120 @xmath0 & 13.30 & @xmath523.32 & 12.39 & @xmath524.23 & 17.22 & 0.91 & 0.43 & 2.37 & 16.07 & 1.49 & 0.37 & 1.38 + pg1126@xmath5041 @xmath0 & 12.49 & @xmath524.53 & 12.63 & @xmath524.39 & 17.94 & 2.65 & 0.79 & 1.21 & 16.05 & 1.65 & 0.52 & 1.00 + pg1211 + 143 @xmath0 & 12.37 & @xmath525.33 & 14.44 & @xmath523.26 & 21.98 & 4.99 & 0.00 & 2.41 & 20.19 & 7.49 & 0.51 & 0.85 + pg1229 + 204 @xmath0 & 13.56 & @xmath523.57 & 12.18 & @xmath524.95 & 17.59 & 1.57 & 0.24 & 1.51 & 16.94 & 3.38 & 0.59 & 2.67 + pg1302@xmath5102 @xmath0 & 13.74 & @xmath526.88 & 14.25 & @xmath526.37 & 19.89 & 3.14 & 0.46 & 0.81 & & & & + pg1307 + 085 @xmath0 & 13.97 & @xmath525.22 & 14.71 & @xmath524.48 & 21.38 & 4.70 & 0.00 & 0.64 & & & & + pg1309 + 355 @xmath0 & 14.03 & @xmath525.57 & 13.71 & @xmath525.89 & 19.53 & 3.15 & 0.18 & 1.46 & & & & + pg1351 + 640 @xmath40 & 12.02 & @xmath525.84 & 13.77 & @xmath524.09 & 20.83 & 3.82 & 0.00 & 0.64 & 15.48 & 0.63 & 0.00 & 3.04 + pg1411 + 442 @xmath41 & 13.48 & @xmath524.43 & 14.39 & @xmath523.52 & 21.01 & 6.32 & 0.52 & 1.28 & 17.20 & 0.82 & 0.55 & 2.36 + pg1411 + 442 @xmath0 & 12.45 & @xmath525.46 & 13.45 & @xmath524.46 & 18.82 & 1.30 & 0.15 & 1.09 & 17.83 & 2.85 & 0.56 & 1.35 + pg1411 + 442 @xmath40 & 11.52 & @xmath526.39 & 13.54 & @xmath524.37 & 19.84 & 5.90 & 0.57 & 1.28 & 15.99 & 0.62 & 0.83 & 2.50 + pg1426 + 015 @xmath40 & 11.37 & @xmath526.46 & 12.17 & @xmath525.66 & 18.61 & 5.33 & 0.42 & 2.45 & 14.88 & 0.65 & 0.60 & 0.36 + pg1435@xmath5067 @xmath41 & 14.72 & @xmath523.98 & 15.25 & @xmath523.45 & 21.13 & 2.97 & 0.00 & 2.55 & & & & + pg1440 + 356 @xmath40 & 11.51 & @xmath526.12 & 12.80 & @xmath524.84 & & & & & 14.63 & 1.02 & 0.18 & 2.35 + pg1613 + 658 @xmath41 & 13.56 & @xmath525.19 & 13.76 & @xmath524.99 & 20.98 & 7.90 & 0.22 & 0.82 & 15.99 & 0.30 & 0.90 & 2.14 + pg1613 + 658 @xmath0 & 12.90 & @xmath525.85 & 12.88 & @xmath525.87 & 19.98 & 6.94 & 0.25 & 0.95 & 15.91 & 0.68 & 0.46 & 2.91 + pg1613 + 658 @xmath40 & 11.88 & @xmath526.99 & 12.11 & @xmath526.64 & 18.34 & 4.00 & 0.14 & 0.88 & 17.08 & 0.86 & 0.61 & 2.77 + pg1617 + 175 @xmath40 & 11.12 & @xmath527.30 & 13.40 & @xmath525.02 & 21.73 & 14.34 & 0.00 & 0.08 & & & & + pg1626 + 554 @xmath0 & 13.71 & @xmath525.12 & 14.33 & @xmath525.87 & 20.34 & 2.08 & 0.13 & 1.69 & 18.53 & 1.99 & 0.07 & 0.32 + pg1626 + 554 @xmath40 & 12.96 & @xmath524.50 & 13.11 & @xmath525.72 & 17.76 & 0.30 & 0.01 & 1.13 & 15.61 & 0.59 & 0.03 & 1.30 + pg1700 + 518 @xmath0 & 13.28 & @xmath527.47 & 14.62 & @xmath526.13 & 20.85 & 4.29 & 0.44 & 1.98 & 15.80 & 0.60 & 0.74 & 1.59 + pg1700 + 518 @xmath40 & 12.04 & @xmath528.71 & 15.56 & @xmath525.19 & inf & 2.11 & 0.54 & 1.24 & 15.87 & 1.04 & 0.81 & 1.70 + pg2130 + 099 @xmath0 & 12.05 & @xmath525.08 & 13.60 & @xmath523.53 & & & & & 15.54 & 1.25 & 0.40 & 2.27 + pg2130 + 099 @xmath40 & 10.76 & @xmath526.37 & 12.72 & @xmath524.41 & & & & & 14.42 & 1.08 & 0.35 & 2.24 + 4c 31.63 @xmath0 & 13.68 & @xmath527.09 & 14.25 & @xmath526.52 & 20.97 & 6.47 & 0.37 & 0.64 & & & & + pg2214 + 139 @xmath0 & 12.41 & @xmath524.83 & 12.37 & @xmath524.87 & 18.07 & 2.72 & 0.07 & 3.02 & & & & + pg2251 + 113 @xmath0 & 13.51 & @xmath527.51 & 14.27 & @xmath526.75 & 19.77 & 1.43 & 0.34 & 0.21 & 20.79 & 10.00 & 0.16 & 0.56 + pg2349@xmath5014 @xmath0 & 13.54 & @xmath525.93 & 13.57 & @xmath525.90 & 20.41 & 6.19 & 0.38 & 0.22 & 19.32 & 3.60 & 0.51 & 1.67 + l c c c c c c c c c c elliptical & 10 & @xmath526.04 & @xmath525.93 & 0.24 & @xmath524.91 & @xmath525.24 & 0.27 & 1.83 & 2.21 & 0.40 + disk present & 7 & @xmath525.08 & @xmath525.20 & 0.35 & @xmath523.55 & @xmath523.78 & 0.26 & 4.17 & 6.58 & 2.61 + bulge+disk & 3 & @xmath523.57 & @xmath523.81 & 0.30 & @xmath524.39 & @xmath524.52 & 0.18 & 0.43 & 0.62 & 0.22 + unknown & 7 & @xmath525.85 & @xmath525.99 & 0.24 & @xmath525.87 & @xmath525.39 & 0.28 & 1.23 & 2.16 & 0.54 + non - disturbed & 18 & @xmath525.40 & @xmath525.38 & 0.25 & @xmath524.47 & @xmath524.62 & 0.24 & 1.93 & 2.86 & 0.56 + disturbed & 3 & @xmath524.83 & @xmath525.40 & 0.62 & @xmath524.87 & @xmath524.44 & 0.36 & 0.96 & 7.84 & 5.73 + strongly disturbed & 6 & @xmath525.88 & @xmath526.02 & 0.27 & @xmath525.88 & @xmath525.61 & 0.26 & 1.13 & 1.68 & 0.39 + total & 27 & @xmath525.57 & @xmath525.52 & 0.20 & @xmath524.58 & @xmath524.82 & 0.19 & 1.69 & 3.15 & 0.81 + l c c c c c c pg0026 + 129 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath324 + pg0050 + 124 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 + pg0157 + 001 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 + pg0804 + 761 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 + pg0838 + 770 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + pg0844 + 349 & @xmath349 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & @xmath354 + pg0923 + 201 & @xmath355 & @xmath356 & @xmath357 & @xmath358 & @xmath359 & @xmath360 + pg0953 + 414 & @xmath361 & @xmath362 & @xmath363 & @xmath364 & @xmath365 & @xmath366 + pg1001 + 054 & @xmath367 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath372 + pg1004 + 130 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 + pg1116 + 215 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 + pg1119 + 120 & @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 + pg1126@xmath5041 & @xmath391 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 + pg1211 + 143 & @xmath397 & @xmath398 & @xmath399 & @xmath400 & @xmath401 & @xmath402 + pg1229 + 204 & @xmath403 & @xmath404 & @xmath405 & @xmath406 & @xmath407 & @xmath408 + pg1302@xmath5102 & @xmath409 & @xmath410 & @xmath411 & @xmath411 & @xmath412 & @xmath413 + pg1307 + 085 & @xmath414 & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath419 + pg1309 + 355 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath424 & @xmath425 + pg1351 + 640 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 & @xmath431 + pg1411 + 442 & @xmath432 & @xmath433 & @xmath434 & @xmath435 & @xmath436 & @xmath437 + pg1426 + 015 & @xmath438 & @xmath439 & @xmath440 & @xmath441 & @xmath442 & @xmath443 + pg1435@xmath5067 & @xmath444 & @xmath445 & @xmath446 & @xmath364 & @xmath447 & @xmath448 + pg1440 + 356 & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 & @xmath454 + pg1613 + 658 & @xmath455 & @xmath456 & @xmath457 & @xmath458 & @xmath459 & @xmath460 + pg1617 + 175 & @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath465 & @xmath342 + pg1626 + 554 & @xmath466 & @xmath467 & @xmath468 & @xmath469 & @xmath470 & @xmath471 + pg1700 + 518 & @xmath472 & @xmath473 & @xmath474 & @xmath475 & @xmath476 & @xmath477 + pg2130 + 099 & @xmath478 & @xmath479 & @xmath480 & @xmath481 & @xmath482 & @xmath483 + 4c 31.63 & @xmath484 & @xmath485 & @xmath486 & @xmath487 & @xmath488 & @xmath372 + pg2214 + 139 & @xmath489 & @xmath490 & @xmath491 & @xmath492 & @xmath493 & @xmath494 + pg2251 + 113 & @xmath495 & @xmath496 & @xmath497 & @xmath498 & @xmath499 & @xmath500 + pg2349@xmath5014 & @xmath501 & @xmath410 & @xmath502 & @xmath503 & @xmath504 & @xmath505 + l c c c c c c elliptical & 10 & @xmath523.88 & 11.61 & 0.27 & 0.29 [ 0.14 ] + disk present & 8 & @xmath522.52 & 11.49 & 0.53 & 0.81 [ 0.56 ] + bulge+disk & 3 & @xmath521.63 & 11.12 & 0.68 & 0.66 [ 0.04 ] + unknown & 11 & @xmath523.85 & 11.77 & 0.38 & 0.54 [ 0.39 ] + non - disturbed & 21 & @xmath523.16 & 11.42 & 0.42 & 0.47 [ 0.35 ] + disturbed & 3 & @xmath522.19 & 11.47 & 0.46 & 0.85 [ 0.67 ] + strongly disturbed & 6 & @xmath523.93 & 11.85 & 0.67 & 0.73 [ 0.39 ] + unknown & 2 & @xmath524.04 & 11.67 & 0.27 & 0.27 [ 0.10 ] + | we report near - infrared ( primarily @xmath0-band ) adaptive optics ( ao ) imaging with the gemini - n and subaru telescopes , of a representative sample of 32 nearby ( @xmath1 ) qsos selected from the palomar - green ( pg ) bright quasar survey ( bqs ) , in order to investigate the properties of the host galaxies .
two - dimensional modeling and visual inspection of the images shows that @xmath236% of the hosts are ellipticals , @xmath239% contain a prominent disk component , and @xmath225% are of undetermined type .
thirty percent show obvious signs of disturbance .
the mean @xmath3 ( 2.1@xmath4 ) , with a range @xmath523.5 to @xmath526.5 ( @xmath20.63@xmath4 to 10@xmath4 ) . at @xmath6@xmath4 ,
all hosts have a dominant disk component , while at @xmath72@xmath4 most are ellipticals .
disturbed " hosts are found at all @xmath8 , while strongly disturbed " hosts appear to favor the more luminous hosts .
hosts with prominent disks have less luminous qsos , while the most luminous qsos are almost exclusively in ellipticals or in mergers ( which presumably shortly will be ellipticals ) . at @xmath9 , where our sample is complete at @xmath10-band
, we find no clear correlation between @xmath11 and @xmath8 .
however , at @xmath12 , the more luminous qsos ( @xmath13 ) , and 4/5 of the radio - loud qsos , have the most luminous @xmath0-band hosts ( @xmath14 ) , most of which are ellipticals .
finally , we find a strong correlation between the infrared - excess " , @xmath15 , of qsos with host type and degree of disturbance .
disturbed and strongly disturbed hosts and hosts with dominant disks have @xmath15 twice that of non - disturbed and elliptical hosts , respectively .
qsos with disturbed " and strongly - disturbed " hosts are also found to have morphologies and mid / far - infrared colors that are similar to what is found for warm " ultraluminous infrared galaxies , providing further evidence for a possible evolutionary connection between both classes of objects . |
close to the kinematical threshold the total cross section for the meson production via the nucleon - nucleon interaction grows rapidly with increasing excess energy q. it is well established @xcite that for the @xmath0 @xcite and @xmath3 @xcite reactions this total cross section changes by about two orders of magnitude within a q range of about ten mev .
the shape of the excitation function is predominantly determined by the changes of the phase space volume and by the final state interaction among the produced particles .
the precision of the experiments performed at the cooler synchrotrons allows to distinguish the subtle effects originating from the meson - nucleon interaction .
a quantitative derivation of the @xmath4 and @xmath5 hadronic potentials requires , however , a sophisticated theoretical treatment since the distortion caused by the nucleons is by orders of magnitude larger than that due to the meson nucleon forces , and even small fractional inaccuracies in the description of nucleon nucleon effects may obscure the inference on the meson nucleon interaction . to minimize the ambiguities which may result from these discrepancies at least for the qualitative estimation of the effects of the unknown meson nucleon interaction one can compare the spectra from the production of a meson under investigation to the spectra determined for the production of a meson whose interaction with nucleons is established . to visualize the influences of the @xmath4 and @xmath5 interaction on the energy dependence of the total cross section we have compared the modulus of the primary transition amplitude @xmath6 of the @xmath0 and @xmath3 reactions to the one extracted from the data on the @xmath7 reaction @xcite
-proton interaction is negligibly weak in comparison to the proton - proton one .
the real part of the @xmath8 scattering length ( @xmath9 fm @xcite ) is more than a factor of 50 smaller than @xmath10 fm @xcite . ] .
figures [ ratio]a and [ ratio]b show the dependence of @xmath6 on the phase
space volume for @xmath1 and @xmath2 production normalized to @xmath11 .
the values of @xmath6 were extracted from the experimental data disregarding the proton meson interaction . if the influence of the neglected interactions were the same in the case of the @xmath12 and @xmath13 production the points would be consistent with the solid line .
this holds in case of the @xmath14 reaction indicating the weakness of the proton@xmath2 interaction independently of the prescription used for the proton proton fsi @xcite . in case of the @xmath2 meson
energy interaction with the nucleons was expected to be very weak since there exists no baryonic resonance which would decay into @xmath15 channel @xcite . in contrary
, the existence of the @xmath16 resonance , which decays significantly into nucleon and the @xmath1 meson , indicates that the @xmath17 interaction is much stronger than the @xmath18 one , and indeed as depicted in figure [ ratio]a the strong effects of the @xmath19 fsi at low @xmath20 are visible . as a next step for a quantitative understanding of the @xmath21 dynamics a full three - body description of the system with the complex hadronic potentials
is required as well as an exact determination of the magnitudes of the contributing partial waves .
some aspects of the latter issue are discussed in the next section .
on previous conferences we have already reported on the phase space density distribution determined for the @xmath0 reaction at an excess energy of q = 15.5 mev @xcite . the obtained spectrum
shown here in figure [ figure2 ] revealed a strong deviation from the expectation based on the factorization of the transition amplitude into the constant primary production and the on - shell incoherent pairwise interaction among the produced particles ( see solid and dashed lines ) .
the experimental data obtained independently by tof @xcite and cosy-11 @xcite using different detection systems agree perfectly with each other and make possible systematical errors rather improbable .
the data show also a fully isotropic distribution over the polar emission angle of the @xmath1 meson in the center - of - mass frame @xcite , and are consistent with an isotropic angular distribution of the relative momentum of the protons seen in the center - of - mass system @xcite ( angle @xmath22 on the middle side of figure [ figure2 ] ) . according to preliminary investigations based on the meson exchange model @xcite ,
the observed distribution of the proton - proton invariant mass can not be explained assuming that the production of the outgoing particles takes place exclusively with the relative angular momentum equal to zero .
the discrepancy between the solid line of figure [ figure2 ] and the data is also too large to be explained by the underestimation of the s - wave proton-@xmath1 interaction .
inspired by that difficulty we checked the partial wave distribution in the proton - proton system deriving from the data the angular distribution of the proton momentum in the rest frame of proton - proton system ( angle @xmath23 on the right side of figure [ figure2 ] ) .
a possible non - zero angular momentum between outgoing protons should manifest itself in an unisotropic population of the angle between the relative proton - proton momentum and the recoil particle ( @xmath1 ) seen from the di - proton rest system @xcite .
the distributions determined for three intervals of @xmath24 are shown in figure [ figure3 ] .
as a first step we restricted the analysis assuming that only s and p waves contribute , and we fit the data by the linear combination of the legendre polynomials up to the second degree , which in case of two identical particles reads : @xmath25 . in this representation
the parameter @xmath26 is a measure of the relative amplitude of the p and s wave contributions . as a result we obtained that @xmath26 is consistent with zero for @xmath27 gev@xmath28/c@xmath29 and amounts to @xmath30 for the middle and upper ranges of @xmath24 .
this value supports the hypothesis that an admixture of p - waves in the proton - proton subsystem is not negligible already at an excess energy of q = 15.5 mev .
9 p. moskal , m. wolke , a. khoukaz , w. oelert , prog .
* 49 * ( 2002 ) 1 . h. caln et al . , phys . lett . *
b 366 * ( 1996 ) 39 . ; j. smyrski et al .
* b 474 * ( 2000 ) 182 .
; e. chiavassa et al . , phys . lett . *
b 322 * ( 1994 ) 270 . ; a. m. bergdolt et al .
, phys . rev .
* d 48 * ( 1993 ) r2969 .
; h. caln et al . , phys .
* 79 * ( 1997 ) 2642 .
f. balestra et al . , phys . lett . * b 491 * ( 2000 ) 29 . ; p. moskal et al . , phys . lett . * b 474 * ( 2000 ) 416 . ; p. moskal et al . , phys .
lett . * 80 * ( 1998 ) 3202 . ; f. hibou et al .
, phys . lett .
* b 438 * ( 1998 ) 41 .
p. moskal et al . , phys . lett . * b 482 * ( 2000 ) 356 .
h. o. meyer et al . ,
. phys . * a 539 * ( 1992 ) 633 . and phys
* 65 * ( 1990 ) 2846 . ;
a. bondar et al .
* b 356 * ( 1995 ) 8 .
d. sigg et al . , nucl .
phys . * a 609 * ( 1996 ) 269 ; nucl . phys . * a 617 * ( 1997 ) 526
. j. p. naisse , nucl . phys . * a 278 * ( 1977 ) 506 .
groom et al .
j. * c 15 * ( 2000 ) 1 .
p. moskal et al . , e print archive : nucl - ex/0110018 .
p. moskal et al . , e - print archive : nucl - ex/0208004 ; and nucl - ex/0210019 .
m. abdel - bary et al . , e - print archive : nucl - ex/0205016 .
k. nakayama , ch .
hanhart , private communication ( 2002 ) .
d. grzonka , k. kilian , schriften des fz jlich matter and materials * 11 * ( 2002 ) 100 . | due to their short life - time , flavour - neutral mesons can not be utilized as free secondary beams or targets , and therefore a study of their interaction with nucleons is not possible via direct scattering experiments .
this interaction is , however , accessible via its influence on the energy dependence and on the phase space distributions of the cross sections for reactions in which these mesons are produced .
in case of the @xmath0 reaction the experimentally determined distributions of the differential cross sections close to the production threshold can not be described by taking into account the s - wave proton - proton and proton-@xmath1 interaction only . here
we show that the angular distributions determined at the cosy-11 facility reveal some evidence for p - wave admixture in the proton - proton subsystem already at an excess energy as low as q = 15.5 mev .
we also present that one can estimate the relative strength of the @xmath1-nucleon and @xmath2-nucleon interactions by comparison of the @xmath1 and @xmath2 production yield . |
in this section , we are going to show how the drinfeld double @xmath3 structure appears in the context of three dimensional riemannian gravity coupled to massive point particles : the partial physical hilbert space is in fact isomorphic as a hilbert space to a tensor product of simple representations of @xmath3 . in that picture , the partial physical scalar product between two states is an intertwining coefficient between simple representations .
thus , the so - called barrett - crane ( bc ) intertwiner @xcite plays a crucial role in three dimensional quantum gravity whereas it was introduced as a model of four dimensional gravity .
this section is organized as follows .
first , we recall basic facts concerning the drinfeld double @xmath3 : definition , representation theory .
we also introduce the notion of simple representation and symmetric ( or bc ) intertwiner .
then , we show the role of @xmath3 in the case of a sphere with @xmath4 particles : @xmath3 appears as the symmetry group of the quantum theory and the non - trivial braiding is interpreted in the language of lqg .
finally , we generalize the previous results in the case of any riemann surface @xmath1 .
the general definition of the quantum double ( or the drinfeld double ) of a hopf algebra @xmath24 is briefly recalled in the appendix a. the general definition is then illustrated in the generic case of a finite group @xcite . in this section ,
we consider the case of the compact group @xmath0 , i.e. the case where @xmath313 $ ] @xcite .
the drinfeld double @xmath314 \otimes f(su(2))^{op}$ ] is a hopf algebra whose definition is precisely given in the appendix .
in particular , it is quasi - triangular and admits the group algebra @xmath315 $ ] and the algebra of functions @xmath316 ( with opposite co - product ) as sub - hopf algebra . _
hopf algebra structure _ the hopf algebra structure of @xmath315 $ ]
is defined by the group law and the following co - algebra relations : @xmath317 note that there exists a right and left invariant haar measure on @xmath315 $ ] given by : @xmath318 \ ; \longrightarrow \ ; \mathbb c \;\;\ ; \text{and } \;\;\ ; h(x)=\delta(x)\;.\end{aligned}\ ] ] the algebra of functions @xmath319 is commutative and its coalgebra struture is given by : @xmath320 where @xmath141 is a function and @xmath321 . we know that @xmath319 admits a haar measure that we have denoted @xmath322 as usual .
finally , the drinfeld double is defined by the previous formulae and the following ones ( which are a direct application of the defintion given in the appendix ) : @xmath323 where @xmath324 , @xmath325 .
the action of the antipode reads @xmath326 for any @xmath327 and the co - unit is simply given by @xmath328 . _
representation theory _
unitary irreducible representations of @xmath3 have been studied in @xcite and are classified by a couple @xmath329 : @xmath330 $ ] labels a conjugacy class of @xmath0 , denoted @xmath331 , whose representative is still chosen to be @xmath332 ; @xmath333 is an integer which labels irreducible representations of the centralizer @xmath334 of the element @xmath332 .
note that @xmath335 for @xmath336 ; otherwise @xmath334 is obviously the group @xmath18 itself . in the generic case @xmath337 , the vector space of a representation @xmath329
consists in the subspace of functions on @xmath0 defined by : @xmath338 the representation of any element @xmath339 on the above vector space is defined by : @xmath340 the vector space @xmath341 inherits a natural hilbert space structure .
the particular case @xmath342 has been considered in @xcite . in the sequel
, we use the notation @xmath343 . _
simple representations and symmetric intertwiners _ representations of the form @xmath344 are called simple representations @xcite and are the building blocks of the barrett - crane model @xcite .
the vector space of a simple representation is isomorphic to the space of functions on the sphere @xmath128 .
any simple representation admits a normalized @xmath0-invariant vector denoted @xmath345 .
the vector space associated to a simple representation @xmath346 will be denoted @xmath347 ( we omit the subscript @xmath348 for the spin ) .
it is obvious that @xmath349 and we will identify in the following the spaces @xmath347 and the conjugacy class @xmath331 which contains the information on the mass .
we will use the same notation @xmath350 or @xmath351 for the state @xmath103 when viewed as an element of @xmath352 or an element of @xmath331 .
the notion of symmetric intertwiner ( or the barrett - crane intertwiner ) has been introduced a long time ago @xcite and is defined as an intertwiner between simple representations whose decomposition into three - valent intertwiners introduces only simple representations in the intermediate channel . up to a normalization
, the simple intertwiner is unique @xcite .
there exists an integral formulation of the symmetric intertwiner . to present this formulation ,
we need to recall the definition of the symmetric propagator @xcite . given a simple representation labelled by @xmath346 , the symmetric kernel @xmath353 is an element of @xmath354 defined by @xmath355 it is convenient and equivalent to view ( by duality ) the symmetric kernel as an element of @xmath315 $ ] as follows : @xmath356 note that @xmath357 is the characteristic function of the conjugacy class @xmath331 and can be decomposed into fourier modes as follows : @xmath358 if @xmath353 had been a function on a `` classical '' group @xmath6 , we would have defined the `` classical '' propagator as the function @xmath359 . in our case ,
the propagator , denoted @xmath360 , is an element of @xmath315^{\otimes 2}$ ] defined by : @xmath361 this propagator is the building block of the symmetric intertwiner whose construction is given next @xcite .
let @xmath362=(p_1,\cdots , p_a)$ ] and @xmath363=(q_1,\cdots , q_b)$ ] two ordered families of simple representations .
one defines the symmetric intertwiner between the representations @xmath362 $ ] and @xmath363 $ ] , @xmath364;[q ] ) \ ; : \ ; \bigotimes_{i=1}^a { \cal v}_{p_i } \ ; \longrightarrow \ ; \bigotimes_{j=1}^b { \cal v}_{q_j } \;,\end{aligned}\ ] ] by its matrix elements given by : @xmath365;[q ] ) \bigotimes_{k=1}^a
\psi_k > \ ; = \ ; h \left ( \prod_{j=1}^b ( \overline{\varphi_j } \otimes 1 ) k_{q_j}(x_j , g ) \ ; \prod_{i = a}^1 ( \psi_i \otimes 1)k_{\overline{p_i}}(y_i , g)\right)\;,\end{aligned}\ ] ] where @xmath366 and @xmath367 for any @xmath368 or @xmath235 .
note that @xmath369 is the haar measure on the group algebra @xmath315 $ ] and the products in the formulae ( [ matelements ] ) are ordered such that when one develops the expression , one obtains : @xmath370;[q ] ) \bigotimes_{k=1}^a \psi_k > = \int \ !
\prod_{i , j } dx_j \ ; dy_i \ ; \mathbb i_{q_j}(x_j)\overline{\varphi_j(x_j)}\ ; \delta(\prod_{\ell=1}^b x_\ell^{-1 } \prod_{k = a}^1 y_k)\;\mathbb i_{p_i}(y_i)\psi_i(y_i ) \;.\end{aligned}\ ] ] let us propose several remarks concerning the symmetric intertwiner .
if one views the states @xmath371 and @xmath372 as functions on the 2-sphere , the previous formulae reduces to the following one : @xmath370;[q ] ) \bigotimes_{k=1}^a \psi_k > \ ; = \ ; \int \ ; \prod_{j=1}^b \overline{\varphi_j(\tilde{x}_j ) } \ ; d\tilde{x}_j \ ; \prod_{i=1}^a \psi_i(\tilde{y}_i ) \ ; d\tilde{y}_i \ ; \delta(\prod_{l=1}^b x_l h(\overline{q_l})x_l^{-1 } \prod_{k = a}^1 y_k h(p_k ) y_k^{-1})\;\nonumber\end{aligned}\ ] ] in the particular ( and singular ) case where @xmath373 and @xmath374 are picked at the values @xmath375 and @xmath376 on the sphere , the symmetric intertwiner is a distribution given by : @xmath377;[q ] ) \bigotimes_{k=1}^a \tilde{y}_k > \ ; = \ ; \delta(\prod_{\ell=1}^b x_\ell h(\overline{q_\ell } ) x_\ell^{-1 } \prod_{k = a}^1 y_k h(p_k ) y_k^{-1})\;.\end{aligned}\ ] ] remark 2 .
we easily show that the decomposition of the symmetric intertwiner introduces only simple representations in the intermediate channel . using the previous notations
, we have for instance : @xmath364;[q ] ) = \frac{1}{\pi}\int dr \ ; p(r ) \ ;
\iota_s(p_1,p_2;r ) \ ; \circ \;\iota_s(\overline{r},p_3,\cdots , p_a;[q])\;,\end{aligned}\ ] ] where @xmath378 denotes the usual composition between intertwiners and @xmath379 is the plancherel measure .
therefore , we concentrate for the moment on the three valent intertwiner @xmath380 .
we start by saying that it is indeed an intertwiner , i.e. it satisfies the following identity : @xmath381 for any @xmath382 .
the symmetric intertwiner is not normalized and we define the coefficient @xmath383 such that the intertwiner @xmath384 is normalized , i.e : @xmath385 a straightforward calculation shows that : @xmath386;[q ] ) \vert 1 > \ ; = \ ; \frac{\pi}{4 } \frac{y(p , q_1,q_2 ) } { \sin p \ ; \sin q_1 \ ;
\sin q_2 } \;\end{aligned}\ ] ] where @xmath387 is the characteristic function of the set @xmath388 .
remark 4 . from the previous remarks ,
it is easy to show that @xmath389;[q])$ ] is an intertwiner , i.e. it satisfies : @xmath364;[q ] ) \ ; ( \otimes_{i=1}^a \pi_{p_i})\delta^{(a)}(z \otimes f ) \ ; = \ ; ( \otimes_{i=1}^b \pi_{q_i})\delta^{(b)}(z \otimes f)\ ; \iota_s([p];[q])\;.\end{aligned}\ ] ] we used the notation @xmath390 defined by @xmath391 and @xmath392 .
it is straightforward to extend the action of the symmetric intertwiner @xmath389;[q])$ ] to the space @xmath393 . _
r - matrix and braiding _
another important property of the quantum double is its quasitriangularity .
indeed , the quantum double is , by construction , a quasitriangular hopf - algebra and therefore admits a @xmath394-matrix .
the universal expression of @xmath394 is given by the formula : @xmath395 then , we can evaluate the @xmath394-matrix on any pair of representations .
we are particularly interested in the case of simple representations .
let @xmath396 be a pair of simple representations , the evaluation of @xmath394 in these representations is given by : @xmath397 such that , for any @xmath398 , then : @xmath399 where @xmath400 means the action of @xmath182 on the point @xmath401 .
the evaluation of the inverse @xmath394-matrix on simple representations is given by : @xmath402 it is convenient to write the action of the r - matrix and of its inverse on the states viewed as functions on the conjugacy class .
in that context , we do not need to specify the evaluation representation of the @xmath394 matrix ( and its inverse @xmath403 ) and we have : @xmath404 it is then straightforward to extend the action of the @xmath394-matrix ( and its inverse ) to any function on @xmath405 .
note the important property that the symmetric intertwiner is invariant under braidings , i.e. : @xmath364;[q ] ) \ ; r^{\varepsilon}_{p_i , p_{i+1 } } \ ; = \ ; r^{\varepsilon}_{q_j , q_{j+1 } } \ ; \iota_s([p];[q ] ) \ ; = \ ; \iota_s([p];[q]),\end{aligned}\ ] ] for any @xmath406 $ ] , @xmath407 $ ] ( using the notations of the previous sections ) and @xmath408 .
this is known as the pivotal symmetry of the bc intertwiner .
we are going to show how the drinfeld double appears in the context of 3d lqg coupled to point particles .
this is in fact immediate and we have the following theorem : [ doubleinlqg ] let @xmath285=(m_0,\cdots , m_{n-1})$ ] the masses of @xmath4 particles ; @xmath285 $ ] labels also a family of simple representations of @xmath3 .
let us define the trivial map between the partial physical hilbert space @xmath409;g)$ ] and the tensor product @xmath410 of simple representations as follows : @xmath411;g ) \longrightarrow \bigotimes_{i=0}^{n-1 } { \cal v}_{\overline{m_i } } \;,\;\;\ ;
\varphi \longmapsto { \cal f}(\varphi ) : ( \tilde{x}_{0},\cdots,\tilde{x}_{n-1 } ) \mapsto \mathbb i(\tilde{x}_0 ) \varphi(\tilde{x}_1,\cdots,\tilde{x}_{n-1})\;. \end{aligned}\ ] ] this map is obviously linear and is trivially extended to the algebra of functions .
moreover , the partial physical hilbert space between two states @xmath305 and @xmath103 is given by : @xmath412;0 ) \vert { \cal f}(\overline{\varphi}\psi ) > \;.\end{aligned}\ ] ] _ proof .
_ this theorem is an immediate consequence of the results of the previous section .
@xmath218 as a consequence , a ( partial ) physical state is a tensor product of vectors of simple representations of the quantum double @xmath3 : each particle is represented by a vector of a simple representation .
therefore , the quantum double is the symmetry group of the quantum theory : given an element @xmath413 , @xmath153 is a rotation and @xmath141 is a translation acting on a one - particle state .
it is immediate to understand that the @xmath0 part of the double generates rotations ; to see that the @xmath319 part generates translations , we take @xmath414 and we consider the element @xmath415 defined by @xmath416 where @xmath417 with @xmath418 ( @xmath419 a unitary vector ) .
it is clear that @xmath415 acts on a state by translation .
let us discuss some consequences of this theorem .
first of all , we note that physical processes conserves the momenta of each particle : a pure momenta state has a non - trivial physical amplitude with the same pure momenta state .
next , we say some words concerning computation of expectation value of operators .
an operator can be viewed as ( a product of ) the evaluation ( on simple representations ) of elements of @xmath3 : such elements act naturally on @xmath420;g)$ ] .
particular examples of operators are the multiplicative @xmath421 and the derivative operator @xmath422 respectively associated to a function @xmath141 and a group element @xmath153 acting on a given particle @xmath423 $ ] as follows : @xmath424 note that the multiplicative operator can be naturally extended to the case where @xmath141 is a distribution .
in fact , any operator ( in our sense ) on the partial physical hilbert space can be decomposed into a sum of products of these basic operators .
we introduce a star ( antilinear involutive ) operator in the algebra of operator generated by @xmath421 and @xmath425 as follows : @xmath426 and @xmath427 .
an operator @xmath428 is unitary in the partial physical hilbert space if it satisfies the identity @xmath429 for any state @xmath305 and @xmath103 .
note that @xmath98 is trivially an unitary operator in the partial physical hilbert space whereas @xmath430 is not in general .
both operators are unitary in the partial kinematical hilbert space . to construct ( non trivial ) unitary operator in the partial physical hilbert space
, one needs to consider braiding operators that are a particular combinaison of multiplicative and derivative operators .
we will see their definition in the sequel .
there is a natural adjoint action of @xmath3 on the set of operators defined by : @xmath431 where we have used the sweedler notation and @xmath428 is an operator .
this action is well - defined . among the operators ,
one distinguishes particularly the so - called observables , i.e.the operators that are invariant under the action of the quantum double ( [ actionofd ] ) .
given a physical process involving in and out states denoted @xmath432 and @xmath433 , the expectation value of any operator @xmath428 during this process is given by : @xmath434;0 ) \vert { \cal f}(\overline{\varphi_{out } } { \cal o } \varphi_{in})>\;.\end{aligned}\ ] ] the mean value of any operator @xmath428 is trivially given by : @xmath435;0 ) \vert { \cal f}({\cal o } ) > } { \text{tr } ( 1 ) } \;.\end{aligned}\ ] ] mean values of any operator are therefore given by coefficients of simple intertwining coefficients .
interesting operators are the monodromy around the particles and their trace are quite easy to compute . given a finite dimensional representation @xmath436 of @xmath0 and a loop @xmath143 around @xmath437 particles among the @xmath438 particles ( let us denote by @xmath439 the set of particles ) , a monodromy @xmath440 is defined by @xmath441 where the expression of the holonomy @xmath442 depends on the choice of the loop @xmath143 . if @xmath443 is a loop around the particle @xmath368 , then @xmath444 is obviously a diagonal operator whose mean value is given by the character @xmath445 .
in the case where @xmath143 is any loop , the computation of the mean value is more involved but the result does not depend on the homotopy class of the loop @xmath143 and is given by : @xmath446 where @xmath447 if @xmath448 satisfy triangular inequalities and @xmath449 if not .
one can compute the mean value of more general operators if one knows explicitely simple intertwining coefficients .
the expression of a simple intertwining coefficient in the basis of functions @xmath450 is explicitly given by the formula : @xmath451 in term of the following coefficients : @xmath452 which can easily expressed in term of the coefficients of @xmath453 symbols of @xmath0 ( see figure ( [ simplecoefficient ] ) for an illustration ) . ) .
the dot colored with the mass @xmath346 denotes insertion of the element @xmath332 ; any trivalent vertex denotes @xmath453 coefficients of @xmath0 ; each edge is colored with an irrep of @xmath0 and can end up with a magnetic number like @xmath454 or @xmath348 . _
_ ] now , let us discuss some aspects concerning the braiding .
first of all , we remark that the @xmath394 matrix , as introduced in the previous section ( [ rmatrix ] ) , defines an isomorphism of ( partial physical ) hilbert spaces . indeed ,
if one considers the particle labelled by @xmath455 $ ] , we define the following map : @xmath456;g ) \ ; \longrightarrow \ ; { \cal h}_{pphys}^{0,n}(p_i[m];g)\end{aligned}\ ] ] where @xmath457=(m_0,\cdots , m_{i+1},m_i , \cdots , m_{n-1})$ ] permutes the particles @xmath368 and @xmath458 .
the fact that the previous map is an isometry is a consequence of the invariance of the simple intertwiner under braidings .
let us remark that the operator ( [ braidoperator ] ) can be written in term of multiplicative and derivative basic operators as follows : @xmath459 then , it is easy to compute the action of the star operator on @xmath394 and we show that it is an unitary operator
. physically , the braiding operator ( [ braidoperator ] ) corresponds to a braiding between the two spin - network edges that link the observer with the particles @xmath368 and @xmath458 .
in fact , there is two ways to braid the edges : one is associated to the @xmath394 matrix and the other to its inverse .
this point is illustrated in the picture ( [ braid ] ) . and @xmath458 .
on the left , the braiding is associated to the @xmath394 matrix whereas it is associated to its inverse @xmath460 on the right . _ _ ] this braiding property means that the partial physical hilbert space does not depend on the order on the particles .
one can generalize the braiding operator to any pair of particles @xmath461 ( let us assume that @xmath462 ) .
there are many ways to do so and a simple example of such an operator is given by : @xmath463 one can as well defines an operator which involves not only @xmath394 matrices but also its inverse .
the operator ( [ generalbraid ] ) defines an isometry between the hilbert spaces @xmath409,g)$ ] and @xmath464,g)$ ] where @xmath465 permutes the particles @xmath368 and @xmath235 .
the generalization of the theorem ( [ doubleinlqg ] ) to the case where @xmath1 is any riemann surface ( of genus @xmath20 ) is quite simple .
to do so , we first need the following proposition : let @xmath437 be a simple representation .
we define the map @xmath466 as follows : @xmath467;g ) & \longrightarrow & { \cal h}_{pphys}^{g-1,n+2}([m],p,\overline{p};g ) \\
\psi & \longmapsto & \upsilon_p(\psi)(a^{g-1,n},\tilde{x},\tilde{y } ) = \int da_1 \ ; db_1 \ ; \delta(b_1 y h(p ) y^{-1 } ) \ ; \delta(ayx^{-1 } ) \ ; \psi(a^{g , n } ) \nonumber\end{aligned}\ ] ] where we have used the notations of the proposition ( [ inclusions ] ) .
this map satisfies the following property : @xmath468 where the partial physical scalar products are respectively those of @xmath284;g)$ ] and @xmath469,p,\overline{p};g)$ ] .
_ given two states @xmath103 and @xmath305 in @xmath284;g)$ ] , we start by computing the partial physical scalar product in @xmath469,p,\overline{p};g)$ ] and a direct calculation shows that : @xmath470 ) \end{aligned}\ ] ] a straightforward application of the kirillov formula , @xmath471 where @xmath472 , leads to ( [ upsilon ] ) .
@xmath218 an immediate consequence of this proposition is that one can reduce the case of a riemann surface to the case of the sphere ( regarding the computation of the partial physical scalar product ) and we have the following theorem : [ theo2 ] let @xmath362=(p_1,\cdots , p_g)$ ] a family of @xmath20 simple representations .
we define the map @xmath473}=\upsilon_{p_g } \circ \cdots \circ \upsilon_{p_1}$ ] and the partial physical scalar product between two states @xmath474;g)$ ] is expressed in term of symmetric intertwiner as follows : @xmath475,[p,\overline{p}];0 ) \vert { \cal f}\circ \upsilon_{[p]}(\overline{\varphi } \psi)>\end{aligned}\ ] ] where @xmath476=(p_1,\overline{p}_1,\cdots , p_g,\overline{p}_g)$ ] and @xmath477 has been introduced in the theorem ( [ doubleinlqg ] ) .
_ let @xmath474;g)$ ] . by a recursive use of the proposition ( [ upsilon ] )
, we immediately show that : @xmath478}(\varphi),\upsilon_{[p]}(\psi)>\end{aligned}\ ] ] where the scalar product on the r.h.s . is the partial physical scalar product in the case of the sphere .
then , we conclude immediately thanks to the theorem ( [ doubleinlqg ] ) .
@xmath218 let us discuss some consequences of this theorem .
first of all , the quantum group structure still holds when @xmath1 is any riemann surface .
however , to make the contact between the hilbert structure and the simple intertwiner is more involved than the case of the sphere . anyway , the quantum double is manisfestly the symmetry group of the quantum theory .
\2 . what about the braiding ?
we have chosen an order to define the partial physical hilbert space : this order is illustrated in the picture ( [ braidgeneral ] ) and consists in putting the particles variables first and the handledodies variables next .
denotes the particles and @xmath479 the handles . _ _ ] we have shown previously that permuting two particles variables corresponds to acting with a @xmath394 matrix .
in fact the same thing happens when one permutes particles with handles or handles with handles .
indeed , in the theorem ( [ theo2 ] ) , we see that one handle is decomposed into a pair of particles whose masses are @xmath346 and @xmath480 ; therefore permuting one particle with one handle corresponds to permuting this particle with the two particles associated to the handle , as a consequence this operation corresponds to acting twice with a @xmath394 matrix . to be more concrete , let us consider the example of the picture ( [ braidgeneral ] ) .
the hilbert spaces @xmath284,\gamma_1;g)$ ] and @xmath284,\gamma_2;g)$ ] , where @xmath71 and @xmath72 are the minimal graphes associated , are related by the following isometry : @xmath481 where the @xmath394 matrix @xmath482 acts on the particles variables @xmath368 and @xmath458 .
we have explicitely shown the representation for the evaluation of the r matrices in order to be clear .
on can easily show that this map is an isometry .
it is easy to generalize this map to the case where @xmath72 is any minimal graph .
the next remark concerns some factorization properties of the partial physical scalar product . to clarify what we mean by that we use the notation @xmath483 $ ] for the partial physical scalar product within this remark and we have : @xmath484 ) \ ; = \ ; \int d\mu(p ) \ ; < ; > _ { pphys}^{g-1,n+1}([m],p ) \ ; \otimes \ ; < ; > _ { pphys}^{1,1}(\overline{p } ) \;. \end{aligned}\ ] ] therefore , the physical scalar product can be factorized into scalar product on the torus .
this property is well - known and is illustrated in the picture of the figure ( [ factorization ] ) .
the last remark concerns the evaluation of geometric observables .
indeed , in the case of a non - trivial topology , there exists non - trivial `` geometric '' operators which corresponds to monodromies around handles .
for instance , the trace of @xmath485 where @xmath454 is a loop around one handle is given by : @xmath486 } \frac{\prod_{i=0}^{n-1 } \chi_j(h(m_i))}{d_j^{2g+n-2 } } \end{aligned}\ ] ] in the case of a genus @xmath20 surface with @xmath4 particles .
@xmath487 $ ] denotes the entire part of @xmath153 .
there is at least two different ways to proove this identity : one can use the last theorem ( [ theo2 ] ) to come back to the case of the sphere or one can make a direct calculation . as a consequence
, we have now a complete description la lqg of three dimensional gravity coupled to massive particles .
in this paper , we have performed the hamiltonian quantization of three dimensional riemannian gravity ( with no cosmological constant ) coupled with a finite number of massive spinless point particles using loop quantum gravity techniques .
the work we have proposed contains : a complete description of the ( partial ) physical hilbert space in the general case of any riemann surface ( i.e. states and scalar product ) and a discussion concerning the question of the observables ( i.e. their definition and the computation of their expectation values in some examples ) .
we have emphasized the crucial role of the drinfeld double which is clearly the symmetry group of the quantum theory , as expected . in the lqg approach
, the drinfeld double appears as a result of the dynamics in the sense that its role becomes obvious when one imposes the three dimensional analog of the `` hamiltonian '' constraint .
this point of view is quite interesting because it is a non - trivial example where one sees the emergence of a non - compact quantum group ( at the level of the physical hibert space ) starting from a classical compact group ( at the level of the kinematical hilbert space ) .
our construction is in that sense very different , at the conceptual level , from the combinatorial quantization , the group field theory quantization or the spin - foam quantization where the quantum group is put by hand as the basic starting block of these methods . from the point of view of a particles physicist ,
our approach might seem more satisfying for we describe particles as simple representations of the `` poincare '' ( in riemannian space ) group and we notice that , using lqg techniques , scattering amplitudes of these particles in a quantum background are described in terms of intertwiners of a quantum group .
in fact , in three dimensions , we see that the hamiltonian constraint ( in the presence of particles ) is in fact a simple intertwiners !
thus , physical states are constructed from simple representations of the drinfeld double and the physical scalar product is in fact given by a symmetric ( or barrett - crane -bc ) intertwiner between the simple representations defining the states .
the ressemblance with the usual bc model is obvious and asks the question whether the bc model is really a model of four dimensional quantum gravity or a model of three dimensional gravity coupled to particles .
this paper opens new insights in lqg techniques .
the first one would be to see the emergence of the quantum group @xmath488 when one imposes the hamiltonian constraint with a cosmological constant ( the quantum parameter @xmath489 is related to @xmath490 in the usual way ) .
we are currently working in that direction .
it seems also possible to generalize our construction to the lorentzian case because we finally do not need to make sense of the whole space of cylindrical functions but only of the space of cylindrical functions on one graph , the minimal graph .
therefore , the obstruction raised in @xcite and in @xcite should not be problematic in our context . going from one minimal graph to another
reduces to a non - trivial braiding which is still well defined in the lorentzian regime .
moreover , one could introduce new types of particles : lightlike , spacelike or timelike particles .
we could also think of introducing particles that would describe a black hole in the presence of a negative cosmological constant .
another issue to be solved is to extend this construction to the case of spinning particles .
this case has been considered briefly in the spin - foam approach by @xcite . in the lqg point of view
, the description of the kinematical hilbert space has been performed completely in @xcite but the way that the physical scalar product is related to the quantum double remains to be understood .
we expect that this link works in the same way as the non - spinning case with the difference that the physical scalar product ( or spinning particles scattering amplitudes in a quantum background ) is no longer given by a simple intertwiner but a more general @xmath3 intertwiner .
in our point of view , the case of spinning particles is not a conceptual issue but only a technical issue . in fact
, the original motivation for this work , is the construction of a self - gravitating quantum field theory in the haniltonian framework .
this means a precise description of the fock space associated to its creation and annihilation operators .
we are finishing a work in that direction @xcite and we have shown in particular that the quantum field theory so obtained is closely related to the one constructed in @xcite in the context of spin - foam models . finally , the last but not the least is to make use of this construction to describe a coupling of quantum gravity to matter field in four dimensions .
it seems quite obvious how to generalize , at the level of the kinematical hilbert space , the coupling of spin - networks to some matter fields .
the more interesting question would be to understand what kind of matter that represents and what is the dynamics of such a coupling ( see @xcite for ideas in that direction ) .
i would like to greatly thank a. perez .
i would also like to thank d. oriti , j. ryan and k. krasnov for stimulating and interesting discussions .
i would also like to thank l. freidel and p. roche for many many discussions concerning quantum groups and the quantum double .
let @xmath24 be a hopf algebra , the quantum double of @xmath24 is denoted @xmath491 ( or @xmath492 ) , is a quasitriangular hopf algebra containing @xmath24 as a sub - hopf algebra and has different equivalent definitions .
if @xmath24 is a hopf algebra , we denote @xmath346 the multiplication and @xmath69 the coproduct ; it will be useful to introduce a permutation operator @xmath493 . from @xmath24
, we can construct different hopf algebra : @xmath494 is the hopf algebra with multiplication @xmath495 and coproduct @xmath69 ; @xmath496 is the hopf algebra with multiplication @xmath346 and coproduct @xmath497 ; the dual hopf algebra @xmath498 construted from the duality bracket @xmath499 between @xmath24 and @xmath500 .
let us illustrate this construction in the case of a finite group to be more concrete .
let @xmath6 be a finite group and we assume that @xmath24 is the group algebra of @xmath6 i.e. @xmath508 $ ] . as a result @xmath509 ,
the algebra of functions on @xmath6 , and the quantum double @xmath510)=\mathbb c[g ] \otimes f(g)^{op}$ ] is usually denoted @xmath511 .
a basis of @xmath511 is @xmath512 and the hopf algebra and coalgebra structures are respectively given by : @xmath513 the action of the antipode on the previous basis is given by @xmath514 and the co - unit is defined by @xmath515 .
the @xmath394-matrix is given by @xmath516 .
this construction can be directly generalized to the case where @xmath6 is locally compact @xcite .
we introduce the group @xmath517 which acts on @xmath284;g)$ ] : it acts by right @xmath6-action of each variable @xmath518 and by right and left @xmath6-action on each variables @xmath454 or @xmath519 . given a partial physical state @xmath520 and an element @xmath521 , we will use the notation @xmath522 for the action of @xmath523 on @xmath520 .
there exists many couples @xmath524 such that the state @xmath520 can be viewed as an element of @xmath284;g)$ ] and we denote @xmath525 the minimal couple .
then , we define the invariance group transformations of @xmath520 by @xmath526\equiv { \cal b}^{g_{min},n_{min}}$ ] . to be more concrete , @xmath526 $ ] transforms only the variables which appear explicitely in the state @xmath520 .
there is a natural action of @xmath526 $ ] on @xmath527 $ ] and we define the sub - hilbert space @xmath528 \subset { \cal h}^{n , g}[m]$ ] as follows : @xmath529 \ ; \equiv \ ; \ { f \in { \cal h}^{g , n}[m ] \ ; \vert \ ; \forall \ ; y \in { \cal b}[s ] \;,\ ; \ ; b_y(f)=f\}\;.\end{aligned}\ ] ] [ inclusions ] let @xmath530;g)$ ] defined by @xmath531)$ ] and @xmath532 where @xmath533 denotes the trace in the fundamental representation of @xmath0 .
we have the following inclusions of hilbert spaces : @xmath534 & \hookrightarrow & \int^{\oplus } \!\ !
d\mu(p ) \ ; { \cal h}_{pphys}^{g-1,n+2}([m],p;\overline{p};g ) \label{inclu1}\;,\\ c^{g , n}[v ] & \hookrightarrow & \int^{\oplus } \!\ !
d\mu(q ) \ ; { \cal h}_{pphys}^{g , n-1}(m_0,\cdots , m_{n-3 } , \overline{q};g ) \otimes { \cal h}_{pphys}^{0,3}(q , m_{n-2},m_{n-1};g)\;.\label{inclu2}\end{aligned}\ ] ] the measure are given by @xmath535 where @xmath536 is the measure on the circle @xmath171 such that @xmath537 .
we have introduced the notation @xmath538 .
we concentrate on the first map ( [ inclu1 ] ) and we show that an explicit map is given as follows : @xmath539 where the family @xmath540 contains the @xmath541 arguments of the function @xmath141 and @xmath542 contains @xmath543 arguments .
the norm of the function @xmath544 viewed as an element of the space @xmath545,p,\overline{p};g)$ ] is given by : @xmath546)\;.\end{aligned}\ ] ] it is then straightforward to verify that : @xmath547)\\ & & \int dx \ ; \delta([x , a_1 ] ) \ ; \overline{f(a_1,b_1;\cdots ) } f(a_1,b_1x;\cdots ) \ ; .\end{aligned}\ ] ] as @xmath548 $ ] and @xmath549 if @xmath550=1 $ ] then we conclude immediately that : @xmath551 therefore , the first inclusion ( [ inclu1 ] ) is proven . to proove the second one ( [ inclu2 ] ) , we proceed in the same way .
we start by claiming that the following map trivially realises the injection ( [ inclu2 ] ) : @xmath552 note that @xmath103 is viewed as an element of @xmath553 and we have introduced the notation @xmath554 .
the norm of @xmath103 is trivially given by : @xmath555 ) \\ & & \ ; \delta(h(\overline{q})z_{n-2 } h(m_{n-2 } ) z_{n-2}^{-1 } z_{n-1 } h(m_{n-1 } ) z_{n-1}^{-1 } ) \;.\end{aligned}\ ] ] therefore , we verify easily that : @xmath556)\;.\end{aligned}\ ] ] finally , as @xmath557 $ ] , then we conclude that : @xmath558 and the hilbert spaces inclusion ( [ inclu2 ] ) is proven .
@xmath218 it is clear that the unit element @xmath309 \in c^{n , g}[u]$ ] for any function @xmath559 . therefore , we can decompose the unit as a tensor product of unit elements of partial hilbert spaces associated to the sphere with three particles .
this is a direct consequence of the previous proposition and the explicit decomposition of @xmath309 $ ] is given by : @xmath560 & = & \int \prod_{i=0}^{g-1 } d\mu(p_i ) \prod_{j=0}^{g-1 } d\mu(m_{n+i } ) \prod_{k=1}^{n+g-2 } d\mu(q_k ) \nonumber \\ & & \bigotimes_{a=0}^{g-1 } e^{0,3}(p_a,\overline{p_a},m_{n+a } ) \ ; \bigotimes_{b=1}^{n+g-2}e^{0,3}(\overline{r_{b-1}},m_b , r_b)\end{aligned}\ ] ] with the conventions @xmath561 and @xmath562 . as a consequence
, one can compute the norm of the identity in two ways and therefore one can relate the coefficients @xmath308)$ ] to the coefficients @xmath312)$ ] .
let us assume that the coefficients @xmath312)=\lambda$ ] do not depend on the values of the masses , then the relation ( [ decompunit ] ) implies that : @xmath563 \vert \!\vert^2_{pphys } & = & \lambda^{n+2g-2 } \int \prod_{i=0}^{g-1 } d\mu(p_i ) \prod_{j=0}^{g-1 } d\mu(m_{n+i } ) \prod_{k=1}^{n+g-2 } d\mu(q_k ) \;\nonumber \\ & & \prod_{a=0}^{g-1 } \vert\!\vert e^{0,3}(p_a,\overline{p_a},m_{n+a } ) \vert \!\vert^2_{pphys } \prod_{b=1}^{n+g-2}\vert\!\vert e^{0,3}(\overline{r_{b-1}},m_b , r_b)\vert\!\vert ^2_{pphys}\end{aligned}\ ] ] after some direct calculations , we show that : @xmath563 \vert \!\vert^2_{pphys } \ ; = \ ; \lambda^{n+2g-2 } \ ; \sum_{k=1}^{+\infty}\ ; { k^{1 - 2g - n}}\ ; \prod_{e=0}^{n-1 } \frac{\sin(km_e ) } { \sin(m_e ) } \ ; .\end{aligned}\ ] ] then , by comparison with the expression ( [ unitnorm ] ) , we conclude that @xmath564)$ ] is also independent of the values of the masses and we have : @xmath565 ) \ ; \equiv \;\lambda(g , n ) \ ; = \;\lambda^{n+2g-2}\;.\end{aligned}\ ] ] this fixes the normalization of the partial physical scalar product . | it is well known that the quantum double structure plays an important role in three dimensional quantum gravity coupled to matter field . in this paper , we show how this algebraic structure emerges in the context of three dimensional riemannian loop quantum gravity ( lqg ) coupled to a finite number of massive spinless point particles . in lqg ,
physical states are usually constructed from the notion of @xmath0 cylindrical functions on a riemann surface @xmath1 and the hilbert structure is defined by the ashtekar - lewandowski measure . in the case where @xmath1 is the sphere @xmath2 , we show that the physical hilbert space is in fact isomorphic to a tensor product of simple unitary representations of the drinfeld double @xmath3 : the masses of the particles label the simple representations , the physical states are tensor products of vectors of simple representations and the physical scalar product is given by intertwining coefficients between simple representations .
this result is generalized to the case of any riemann surface @xmath1 .
most of the articles dealing with any aspect of three dimensional gravity ( 3d ) are introduced and motivated by the well - known fact that 3d gravity would provide a nice framework to answer some questions that are too cumbersome to solve in the four dimensional ( 4d ) context .
it is true that 3d ( pure ) gravity is a topological theory which admits only a finite number of degrees of freedom : this makes the classical theory exactly solvable and different quantization schemes have been deeply studied in the literature @xcite .
the theory is nevertheless far from being trivial and exhibits very rich mathematical structures that are still extensively studied : topological invariants , knots invariants , quantum groups , moduli spaces , teichmuller spaces and so on ... the main objection is that it is not true that all these techniques will shed light on the problem of quantizing four dimensional general relativity .
the interest of 3d gravity is widely limited by the fact that there is no local degrees of freedom in the theory and therefore there are many aspects of 4d gravity that will never be clarified thanks to this toy - model . in order to make 3d gravity a richer tool ,
one should try to introduce local degrees of freedom in the theory .
coupling 3d gravity to an external ( non - trivial ) field is a natural extension of the model that contains local degrees of freedom .
furthermore , it makes the model physically more interesting .
but , the problem of defining a consistent ( full ) quantization of a self - gravitating field theory is very difficult and impossible to solve by making use of standard perturbative quantum field theory techniques , which are so successful for describing the physics of elementary particles .
recently , freidel and livine @xcite have proposed a solution of the problem in the context of 3d spin - foam models : they have shown that ( under certain hypothesis ) a quantum self - gravitating field theory is equivalent to a non - self - gravitating but non - commutative quantum field theory .
the authors claim that the non - commutativity encodes the quantum gravity effects on the field .
even if 3d gravity is physically irrelevant , it is rather interesting to have a model at hand where one has complete control of the quantum gravity effects . for these reasons
, the article @xcite certainly deserves attention for it is a first step in the understanding of a theory of quantum gravity with local degrees of freedom . since
, other interesting articles dealing with different aspects of this problem have been proposed @xcite .
this paper is a first step in the understanding and the analysis of the work of @xcite in the hamiltonian context .
the hamiltonian quantization is a very interesting quantization scheme because the notions of quantum states and quantum operators are very well understood in that framework .
therefore , one can compute explicit transition amplitudes whereas spin - foam models give a priori only the partition function of the theory .
the main problems with the hamiltonian quantization of gravity is that it breaks the covariance of general relativity : there is an explicit choice of time slicing .
topology changings are not allowed because the topology of the spacelike surface is a priori fixed once and for all . on the contrary ,
spin - foam models propose a covariant quantization of general relativity ; this makes this approach so attractive . in this article
, we reformulate the loop quantum gravity ( lqg ) description of 3d gravity coupled to massive point particles where we make clear the crucial role of quantum groups in the quantization .
therefore , we end up with a clear description of the @xmath4-particles states evolving in a quantum background which is an essential step toward the hamiltonian description of a self - gravitating quantum scalar field theory . the last point is presented in a companion paper @xcite .
the paper is organized as follows .
after the introduction , we start by recalling the techniques used in @xcite to perform the hamiltonian quantization la lqg of three dimensional gravity ( pure or coupled to particles ) . in particular , we clarified the link between the canonical quantization and the covariant spin - foam quantization .
however , in this previous works , we never underline the role of the quantum double @xmath3 in our construction whereas it is clear that it should play a central role @xcite .
we fill this gap within the present article . to that aim ,
we first define in the second section the notions of partial kinematical and physical hilbert space .
these spaces are constructed from a particular observer : this observer is a given particle which defines a reference frame where the momenta of the other particles and the energy of the global system are defined .
this decription will appear very important above all in the construction of the fock space @xcite . in the third section ,
we show how the quantum double @xmath3 is naturally present in the ( partial ) physical hilbert space : ( partial ) physical states are defined as vectors of a tensor product of simple representations of @xmath3 and the ( partial ) physical scalar product is nothing but an intertwining coefficients between simple representations of @xmath3 . in this section , we also discuss some aspects concerning the observables : definition , computation of expectation values etc ... we present the construction first in the case of the sphere then we generalize to the case of any riemann surface .
we conclude with some perspectives and we postpone the construction of the fock space to thr companion paper @xcite . in this paper , we will work with dimensionless quantities , in particular for the masses of the particles .
but , the natural mass unit is the planck mass @xmath5 defined from the newton constant @xmath6 by @xmath7 .
then , any mass value has to be understood in terms of the plank unit .
loop quantum gravity is presented as a two steps quantization of general relativity : the construction of the kinematical hilbert space and the construction of the physical hilbert space .
the first point is completely under control and gave rise to what is commonly called quantum geometry ( for review see @xcite , @xcite , @xcite or @xcite ) : the kinematical states ( or states of quantum geometry ) are described in term of spin - networks and their set is endowed with a hilbert structure defined by the ashtekar - lewandowski measure .
but the construction of the physical hilbert space is one the biggest difficulties and one of the most interesting problem of lqg .
different ways have been explored to attack the problem : construction of the extractor @xmath8 by using a regularization of the hamiltonian constraint @xcite , the master constraint program @xcite , spin - foam models @xcite etc ... very recently , thiemann @xcite and han @xmath9 ma @xcite proved the existence of the physical hilbert space in the context of the master program using spectral techniques .
nevertheless , no consistent solution have been found so far ! in a couple of papers @xcite , we have successfully studied 3d general relativity ( pure gravity and gravity coupled to particles ) as a toy - model for lqg .
in particular , we have constructed the extractor @xmath8 and have shown that the physical scalar product between states has a spin - foam representation .
thus , we made the explicit relation between lqg and spin - foam models .
this section is devoted to recall the construction in the pure gravity case and in the case where gravity is coupled to massive spinless point particles .
then , we propose an alternative description for the physical hilbert space from where the quantum double structure emerges naturally .
let @xmath10 be a three dimensional manifold .
in the first order formalism , the degrees of freedom of the gravitational field in @xmath10 are encoded in a triad @xmath11 and in a spin - connection @xmath12 where @xmath13 are the generators of the lie algebra @xmath14 satisfying the lie - algebra relation @xmath15=\varepsilon_{ij}{}^l j_l$ ] .
indices are lowered and raised by the kronecker symbols @xmath16 .
the dynamics is governed by the @xmath17-action for the group @xmath18 . when the manifold has the topology @xmath19 , @xmath1 being a riemann surface of genus @xmath20 and @xmath21 a compact part of the real line @xmath22 , then gravity admits a hamiltonian formulation .
the non - reduced phase space is ( schematically ) defined by : @xmath23 the variables @xmath24 and @xmath25 form a canonical variables pair and are respectively defined by the pullbacks of the spin - connection and of the triad on the surface @xmath1 .
the theory admits first class constraints that generate infinitesimal symmetries in the non - reduced phase space @xmath26 .
an element @xmath27 is called @xmath28-invariant when it poisson - commutes with the constraints .
the symmetry group is given by @xmath29 where @xmath30 is the ( universal cover of the ) group of euclidean transformations .
it is a semi - direct group : @xmath31 note that the symmetry group @xmath28 contains ( on - shell ) the diffeomorphisms group on @xmath1 but is not the diffeomorphism group .
important discrepencies between gravity and @xmath17 theory appears because of this fact @xcite .
the physical phase space @xmath32 is obtained by applying an infinite dimensional version of the dirac reduction to the space @xmath26 and schematically reads : @xmath33 the space @xmath32 is a finite dimensional symplectic manifold and it is isomorphic to the moduli space of flat @xmath30-connections on the surface @xmath1 .
functions on @xmath32 are called observables and , by definition , are invariant under the ( induced ) action of the symmetry group @xmath28 .
different strategies have been developed the last fifteen years to quantize the theory ( see @xcite and references therein ) .
most of them ( combinatorial quantization @xcite , fonctional quantization @xcite , nelson - regge quantization @xcite etc ... ) are proper to 3d gravity and can not be generalized to higher dimensions .
what makes the loop and the spin - foam quantizations very attractive is precisely that these schemes can , in principle , be applied in 4d . even if neither lqg nor spin - foam models provide so far a complete and consistent quantization of 4d gravity .
the loop quantization program consists in first quantizing the non - physical phase space @xmath26 , then promoting the first class constraints into quantum operators and then finding the kernel of these quantum operators . in 4d gravity , all the constraints but
one can be solved in the quantum theory .
the remaining and problematic constraint is known as the hamiltonian constraint . in the context of 3d pure gravity
, this program can be completely achieved and one can find solutions of all the constraints @xcite .
we proceed as follows . 1
. * space of cylindrical functions on @xmath1 : @xmath34 . * + we start by choosing the connection @xmath24 to be the configuration variable ( choice of polarization ) and we denote @xmath35 the space of @xmath6-connections on @xmath1 .
we introduce the space of discrete connections associated to a graph @xmath36 ( with @xmath25 edges and @xmath37 vertices ) , denoted @xmath38 by , rougthly speaking , replacing local connections @xmath24 with holonomies @xmath39 of the connection along the edges @xmath40 of the graph
. a cylindrical function on the graph @xmath36 is an element @xmath41 such that there exists a function @xmath42 and @xmath43 .
the space of cylindrical functions is denoted @xmath44 and is naturally endowed with a measure @xmath45 where @xmath46 is the normalized @xmath6 haar measure associated to the edge @xmath40 .
finally , we define the space of cylindrical functions on the surface @xmath1 as the following union : @xmath47 this space inherits a natural measure , known as the ashtekar - lewandowski measure , and the completion of @xmath48 with respect to this measure is the hilbert space of non - physical states , denoted @xmath49 .
the action of the symmetry group @xmath28 on the connections induces a ( co- ) action on @xmath48 and on @xmath49 and we will use the same notation @xmath28 to denote the symmetry group acting on @xmath48 .
physical states are those functions which are left invariant under this symmetry .
2 .
* invariance under the symmetry group @xmath6 . *
+ the action of the symmetry sub - group @xmath50 reduces to an action on the vertices of any element of @xmath38 for any graph @xmath36 .
invariant states under this action are called kinematical states .
the set of kinematical states is a vector subspace of @xmath48 and , due to left and right invariance of the @xmath6 haar measure , inherits the pre - hilbert structure of @xmath48 .
after completion , we obtain the kinematical hilbert space , denoted @xmath51 .
spin - network states provide an orthonormal basis of @xmath51 .
3 .
* invariance under the translations @xmath52 . * + physical states are a priori kinematical states that are invariant under the action of the residual symmetry subgroup
. however , there is no physical states in the kinematical hilbert space .
this is a well - known fact that is a consequence of the non - compactness of the sub - group @xmath52 .
physical states are in fact `` distributional '' , in the sense that they are elements of @xmath53 , the topological dual of @xmath48 .
finding physical states is equivalent to finding a physical extractor ( abusively called a projector ) , i.e. an operator : @xmath54 that satisfies the property @xmath55 for any cylindrical functions @xmath56 and @xmath57 and any element @xmath58 .
we have denoted by @xmath59 the action of @xmath28 on @xmath48 .
thus the physical hilbert space @xmath60 is defined as the image of @xmath48 by @xmath8 .
given two physical states @xmath61 and @xmath62 , there exists @xmath63 such that @xmath64 and @xmath65 , and the physical scalar product is defined by : @xmath66 the physical scalar product does not depend on the choice of the functions @xmath56 and @xmath57 . generally , we omit to mention the elements @xmath61 and @xmath62 , and we write the physical scalar product between cylindrical functions .
the operator @xmath8 `` extracts '' the physical component of any cylindrical function .
the explicit construction of @xmath8 has been done in @xcite .
in fact , the extractor has been constructed by imposing locally the flatness condition @xmath67 after a choice of regularization .
the physical scalar product so obtained is well - defined ( i.e. convergent ) , is positive and satisfies hermiticity condition .
moreover , the relation to spin - foam model has been established and is breefly recalled in the following . given two spin - network states @xmath56 and @xmath57 respectively associated to the colored graphs @xmath68 , one associates a cellular decomposition @xmath69 of the three - dimensional manifold @xmath70 which interpolates the graphs @xmath71 and @xmath72 at the boundaries .
moreover , @xmath69 is colored in the usual way : faces are associated with irreps of @xmath0 , edges with intertwiners and the colors are compatible with the boundaries colors . at this point ,
one introduces the dual graph @xmath73 whose edges are now colored with @xmath0 irreps .
the spin - foam amplitude associated to the graph @xmath69 ( or in an equivalent way to the dual graph @xmath73 ) gives the physical scalar product between the two spin - network states if and only if @xmath69 is free of bubles , i.e. @xmath69 has a tree structure .
the last condition makes the amplitude convergent and appears naturally in the hamiltonian framework as a consequence of the regularization of the constraint @xmath67 .
therefore , in the case where @xmath73 is a triangulation of @xmath70 we have : @xmath74 where the sum runs over @xmath0 irreps coloring the edges @xmath40 of @xmath73 , the weight @xmath75 is generically given by @xmath76 and is @xmath77 if the edge @xmath40 is dual to a face whose boundary contains an edge of @xmath71 or @xmath72 colored by @xmath78 ; @xmath79 is the normalized symmetric @xmath80 symbols defined by the six representations coloring the edges of the tetrahedron @xmath81 .
this property is illustrated by the following example : @xmath82 where @xmath83 is a the norm of three - valent intertwiners ( evaluation of @xmath84 spin - network ) taken to be 1 .
the state @xmath85 is the spin - network state defined on the sphere and represented in the picture ( [ startriangle ] ) .
and vertices with normalized @xmath0 intertwiners . on the right side
, we have drawn the spin - foam picture illustrating the transition amplitude between the no - state and the state on the left . the amplitude associated to this
spin - foam is a @xmath80 symbol . _
_ ] let us finish with a remark .
we consider two cylindrical functions @xmath56 and @xmath57 which differ only by the fact that their associated graphes are different but related by a diffeomorphism .
therefore , the state @xmath86 is a null - vector in the physical hilbert space and we can identify the spaces @xmath87 and @xmath88 . this is a consequence of the invariance of the theory under the diffeomorphisms group .
furthermore , the identification of the spaces still holds when the graphs @xmath36 and @xmath89 are homotopic and not necessary related by a diffeormorphism .
thus , we identify states that are not related by a diffeomorphism when we impose the flatness condition .
we have to be aware of this fact which is a consequence of the fact that we are working with first order gravity and then we include degenerate metrics in the model .
how to avoid degenerate metrics in the quantum theory is still an open question . in order to couple particles to the gravitational field
, we now consider a surface @xmath1 with boundary : the genus of @xmath1 is still denoted @xmath20 and we assume the boundary is a disjoint union of @xmath4 boundaries whose topology is a circle , i.e. @xmath90 . each boundary @xmath91 is associated to one particle @xmath92 whose degrees of freedom are encoded in an element @xmath93 : @xmath94 is the position of the particle and @xmath95 is related to the momemtum by @xmath96 where @xmath97 .
the mass @xmath98 of each particle @xmath92 is fixed by appropriate boundary conditions of the gravitational field on @xmath91 @xcite .
the action of the coupled system is obtained via a minimal coupling and takes the form : @xmath99 \ ; + \ ; \sum_{i=0}^{n-1 } s_{p_i}[q_i,\lambda_i ] \ ; + \ ; s_c[e,\omega;(q_i,\lambda_i)_i ] \;.\end{aligned}\ ] ] the explicit expression of the action is given in @xcite : @xmath100 is the usual @xmath0 @xmath17-action , @xmath101 is the ( first order ) action of free relativistic particles and @xmath102 represents the minimal coupling term . note that @xmath102 enforces the appropriate boundary conditions which fixes the masses of the particles .
the loop quantization of the action ( [ couplingaction ] ) is conceptually similar to the pure gravity case and has been performed in @xcite .
let us recall the main results . 1 .
* particles - cylindrical functions . *
+ particles - cylindrical functions are a direct generalization of the usual notion of cylindrical functions .
a particles - cylindrical function @xmath103 is now a function of the connection @xmath24 and of the momenta @xmath104 .
such a function is defined by a graph @xmath105 which can admit open ends at the boundaries and by a function @xmath106 where @xmath25 is the number of edges and @xmath107 the number of open ends such that : @xmath108 remind that @xmath39 is the holonomy of the connection along the edge @xmath40 and @xmath109 if the open end @xmath110 .
therefore , the vector space of particles - cylindrical functions @xmath111 is given by : @xmath112 @xmath113 labels the set of the particles .
this space is naturally endowed with a measure ( constructed from the al measure and the @xmath0 normalized haar measure ) which makes it a pre - hilbert space .
its completion is denoted @xmath114 .
2 .
* symmetries of the system . *
+ the system admits first class constraints that generates infinitesimal symmetries .
the symmetry group of the hamiltonian theory is a cartesian product @xmath115 .
any element of @xmath116 is an element of @xmath117 which is constant at each boundary @xmath91 . as in the pure gravity case ,
we distinguish the `` compact '' subgroup @xmath118 from the non - compact one @xmath119 .
the compact part generates usual gauge transformations which read on the variables @xmath24 and @xmath104 : @xmath120 the non - compact part is related to the diffeomorphisms group of the manifold @xmath70 and acts non - trivially on the momenta canonically conjugated to the connection @xmath24 .
+ the group @xmath121 is the set of functions on @xmath122 which are constant on each components @xmath91 and take value in the cartan torus @xmath73 of @xmath30 ( note that @xmath73 is the two - dimensional group @xmath123 ) .
therefore @xmath124 .
it acts trivially on the connection @xmath24 and by right multiplication on the momenta as shown in the following : @xmath125 } \in { \cal c } \;\ ; \text{s.t . } \ ;
\lambda(i ) \in su(2 ) \ ; , \ ; \tau(i ) \in \mathbb r^3 \;\;\;\;\;\;\;\;\ ; \lambda_i \ ; \mapsto \ ; \lambda_i h(i ) \;.\end{aligned}\ ] ] one can show that @xmath126 is the subgroup of time reparatrizations of the worldline of each particle whereas @xmath127 is the subgroup of internal frame rotations which preserves the momentum of each particle .
it is then natural that only the @xmath0 cartan subgroup has a non - trivial action on the momenta .
+ the action of @xmath28 induces a co - action on @xmath111 and particles - physical states are the elements of @xmath111 which are co - invariant under @xmath28 .
3 .
* imposing the constraints and the particles - physical hilbert space . * + invariance under reparametrization is trivial because we work in the momentum representation and invariance under internal frame rotations imposes that states are functions on the sphere @xmath128 instead of a function on the whole group @xmath0 ( see @xcite for more details ) .
therefore , the set of particles - cylindrical functions satisfying these invariances reads : @xmath129 due to the left and right invariance of the @xmath6 haar measure , the space @xmath130 inherits the measure of @xmath111 .
to find particles - physical states , we have to impose the invariance under @xmath28 . as in the pure gravity case , this is a two - steps procedure .
first , we impose the invariance under the sub - group of gauge transformations : this is immediate to do and we obtain the particles - kinematical hilbert space @xmath131 .
particles - spin - network states provide an orthonormal basis of @xmath131 .
note that particles - spin - network states are associated to graphs with ( eventually ) open ends at the location of the particles .
+ imposing the remaining constraint ( which generates the non - compact symmetry group ) is done in the same way as in the pure gravity case .
this means finding an extractor @xmath8 : @xmath132 such that @xmath55 for any particles - cylindrical functions @xmath56 and @xmath57 and any @xmath58 .
the image of @xmath130 by the extractor is a vector space .
we endow this space with a ( ashtekar - lewandowski ) measure and after completion we end up with the particles - physical hilbert space @xmath133 .
the explicit construction of @xmath8 has been done in @xcite .
the relation to spin - foam models has been unraveled : the particles - physical scalar product reproduces spin - foam amplitudes of @xcite as shown in the following section .
it would be very interesting to ask the question of the unicity of the extractor @xmath8 in the context of three dimensional gravity because it seems that the same kinematical space is the starting point of the construction of different physical states depending whether we are dealing with pure gravity , gravity coupled to particles or in the presence of a cosmological constant ... as presented before , any physical state can be viewed as an equivalent class of a particles - cylindrical function ( two states are equivalent if their difference has a non - zero physical norm ) or equivalently as the image of a particles - cylindrical function by the extractor @xmath8 defined above ( [ pparticles ] ) . a particles - physical state @xmath103
is said to be explicitely dependent of the particles degrees of freedom if any of its representative is a particles - cylindrical function defined on a graph which has at least one open vertex ( at a boundary ) .
otherwise , the state does not depend explicitely on the particles - degrees of freedom .
states which do not depend explicitely on the particles degrees of freedom caracterize the three dimensional space - time geometry with conical singularities but do not contain any information ( a part from the masses ) concerning the particles .
the physical scalar product between two such ( spin - network ) states @xmath56 and @xmath57 has a regularized well - defined spin - foam representation of the type @xcite .
indeed , one fixes a tridimensional graph @xmath134 with no - buble interpolating the flat graphes @xmath135 and @xmath136 associated to the cylindrical functions @xmath56 and @xmath57 and the dual graph @xmath73 , and we still have the identity ( [ prmodel ] ) . in that case , edges @xmath40 of @xmath73 are still colored with a representation @xmath137 .
the weight associated to each tetrahedron is still given by @xmath138 but the weight @xmath75 associated to each edge is slightly modified : @xmath139 where @xmath140 if the boundary associated to the particle of mass @xmath98 crosses the dual face @xmath141 of the edge @xmath40 ; @xmath142 where @xmath78 is the representation coloring one edge @xmath143 of the graphes @xmath144 or @xmath145 if the boundary of the dual face of @xmath40 contains the edge @xmath143 ; @xmath146 otherwise . note that , in the construction of @xcite , one has to fix a maximal tree @xmath147 of @xmath73 and to impose the condition @xmath148 to each edge @xmath40 of @xmath147 in order to make the sum ( [ prmodel ] ) convergent . in our construction
, we do not have such a condition because the graph @xmath149 is chosen to be free of bubbles .
the two conditions are in fact equivalent .
the relation ( [ prmodel ] ) is illustrated in the example of the picture ( fig .
[ example ] ) . .
on the r.h.s . the dual picture : the circle around the edge @xmath150 reminds the presence of the boundary .
_ _ ] the spin - foam amplitude associated to the spin - foam on the l.h.s .
is given by : @xmath151 reciprocally , the amplitude of any spin - foam of the type @xcite can be interpreted as the physical scalar product between two states which do not depend explicitely on the particles degrees of freedom . of particular interest
is the physical scalar product between states which depend explicitely on the particles degrees of freedom .
this kind of physical scalar product does not admit a spin - foam representation of the previous type @xcite . in that case , the cellular decomposition of the manifold @xmath70 , defining the eventual spin - foam model , involves faces whose edges belong to the boundaries ( see figure [ example2 ] as an example ) whereas the faces where crossed by the boundaries in the previous case .
it would be very interesting to generalize the spin - foam model developed in @xcite in order to include transitions of the type ( figure [ example2 ] ) .
we let this issue for future investigations . in the sequel ,
we concentrate on the physical scalar product between particles - physical states and we do not ask for the moment the question of its spin - foam representation . then , we want to chose a basis of physical states and we are going to compute the physical scalar product between the elements of the basis . in other words
we are going to compute explicitely the matrix elements of the extractor @xmath8 whose definition and properties were briefly recalled above ( [ pparticles ] ) . for that purpose , we start by chosing a minimal graph @xmath36 on the riemann surface with boundaries @xmath1 : @xmath20 and @xmath4 denote respectively the genus of @xmath1 and the number of connected components of @xmath122 . a minimal graph ( figure [ minimalgraph ] )
consists in a choice of a base point @xmath152 , a choice of @xmath4 oriented edges linking @xmath153 and a point on each boundary @xmath91 ( let @xmath25 be the set of these edges ) and a choice of @xmath154 oriented non - contractible loops around each handles of the surface ( let @xmath155 be the set of these loops that we separate as usual into two sets @xmath24 and @xmath156 ) . the space of discrete connections on the minimal graph is denoted @xmath38 : any discrete connection is a family @xmath157 where @xmath158 $ ] labels the elements of @xmath25 , @xmath159 $ ] the elements of @xmath155 and @xmath160 $ ] the particles .
the infinite dimensional symmetry group @xmath161 ( [ calg],[calc ] ) acting on the space of connections reduces to a finite dimensional group when acting on @xmath38 .
the compact part of this group is trivially isomorphic to @xmath162 .
its action on @xmath38 is given by : @xmath163 the non - compact subgroup acts trivially on the variables @xmath24 and @xmath104 and has a non - trivial action on the variables canonically conjugated . imposing invariance under the non - compact subgroup is done by finding the appropriate extractor @xmath8 .
the functions on discrete connections form the set @xmath44 of cylindrical functions on the minimal graph @xmath36 .
recall that @xmath164 if there exists a function @xmath165 such that given @xmath166 we have : @xmath167 due to the compactness of the gauge group @xmath18 , it is clear that @xmath44 is isomorphic to the space @xmath168 .
then , the space of cylindrical functions is naturally endowed with a pre - hilbert structure defined by the haar measure on the group @xmath6 and its completion , denoted @xmath169 , is called the space of non - physical states .
this space is essential in the constructions of the kinematical hilbert space and the physical hilbert space .
obviously , these constructions do not depend on the choice of the base point @xmath153 on the surface : indeed , the base point has no physical meanning and it is natural that `` physical '' quantities do not depend on it . technically , it is the adjoint action of the gauge group @xmath6 which makes the physical hilbert space independant of the base point .
even the kinematical hilbert space does not depend on that choice .
therefore , one can choose @xmath153 to be a point of a given boundary , let us say @xmath170 . yet
, there exists now two types of symmetries acting on the variables defined on @xmath170 : the action of @xmath6 and the action of @xmath171 .
the former is the usual action of the gauge group @xmath0 whereas @xmath171 is the group of internal frame symmetries ( associated to the particle @xmath172 ) . to construct kinematical states
, one has to consider both symmetries and find invariant functions under these two symmetries .
however , it is interesting to forget the @xmath171 invariance .
this is equivalent to choose a particular internal frame for the particle @xmath172 .
thus , one can capture interesting informations concerning the particles : in particular one can define the momenta of each particles in the internal frame of the particle @xmath172 .
this is obviously impossible if we do not choose a particular frame .
thus , one can interpret the particle @xmath172 as an observer sitting somewhere in the surface @xmath1 .
this observer is measuring the physical caracteristics of each particles and also the properties of the spacelike geometries in its reference frame .
@xmath6-invariant states defined in this reference frame are not kinematical states in the sense of lqg ; they are called in the sequel partial kinematical states .
the set of partial kinematical states is naturally endowed with a pre - hilbert structure .
its completion is called the partial kinematical hilbert space and is denoted @xmath173 .
note that one can extend the space @xmath173 to distributions in the sense that one can choose the element @xmath103 defining the spin - network to be a distribution rather than a function . in that case
, one can define pure momenta states for instance as `` delta '' functions . as the manifold in consideration ( @xmath6 or @xmath2 )
are compact , we will identify the space of function with the space of distributions on this manifold .
we keep this remark in mind for the following .
[ pkin ] the hilbert space of partial kinematical states @xmath173 is isomorphic to the space @xmath174 endowed with the measure @xmath175 .
indeed , a function @xmath176 is a partial kinematical states if and only if there exists a function @xmath177 such that : @xmath178 where @xmath179 and @xmath180 is the measure on the cartan torus of @xmath6 .
we have used the notation @xmath181 for the action of the group element @xmath182 on a point @xmath183 . _ proof .
_ first , let us impose the invariance under the action of @xmath6 on the boundary @xmath184 associated to any particle @xmath185 but the observer @xmath172 .
the group @xmath6 acts non - trivially on the degrees of freedom @xmath186 and @xmath39 according to the following maps ( [ calg ] ) : @xmath187 this action induces a co - action on @xmath188 and it is obvious that any function @xmath189 is co - invariant if and only if it is a function of @xmath190 .
this results holds for any particle @xmath191 .
the action of @xmath6 at the boundary @xmath172 is more involved because there are many edges ending and starting at @xmath172 .
but the conclusion is similar because any function @xmath103 which is @xmath6 invariant at @xmath172 can be written as a sum of functions of the variables @xmath192 , @xmath193 and @xmath194 . to proove this point
, we decompose the function @xmath103 into ( tensor product of ) irreducible unitary finite dimensional representations ( irreps ) of @xmath6 using the plancherel formula and we obtain : @xmath195,[j],[k ] } \stackrel{[i]}{f}{}\!\!_{[j]}^{[k]}(\lambda_e ) \prod_{\ell } \stackrel{i_\ell}{\pi}{}\!\!^{j_\ell}_{k_\ell } ( u_\ell ) \times \stackrel{i_0}{\pi}{}\!\!^{j_0}_{k_0 } ( \lambda_0 ) \;.\end{aligned}\ ] ] @xmath196 \in ( \frac{1}{2}\mathbb n)^{2g+n}$ ] are @xmath0 irreps labelling the edges @xmath143 ( for the links @xmath40 and the non - contractible loops @xmath197 ) and the reference particle @xmath172 ; @xmath198 $ ] and @xmath199 $ ] are families of magnetic numbers . we have denoted @xmath200 the matrix elements of irreps of @xmath0 , @xmath201 is the vector space associated to the spin @xmath21 representation and @xmath202 ( resp .
@xmath203 ) the ( resp .
dual ) basis of @xmath201 ( resp .
@xmath204 ) .
the dependence in the variables @xmath186 , @xmath191 , is contained in the definitions of the fourier components @xmath205}{f}{}\!\!_{[j]}^{[k]}(\lambda_e)$ ] .
the decomposition ( [ decomposition ] ) is nothing else but the spin - network decomposition of the state .
the @xmath6 invariance at @xmath172 implies that there exist a @xmath6 intertwiner @xmath206 defined by : @xmath207}{f}{}\!\!_{[j]}^{[k ] } \ ; = \ ; < \bigotimes_{a , b=1}^g \stackrel{i_a}{e}{}\!\!^{k_a } \otimes \stackrel{i_b}{e}{}\!\!^{k_b } \vert \iota(\stackrel{i_0}{e}{}\!\!^{k_0 } \otimes \bigotimes_{e=1}^{n-1}\stackrel{i_e}{e}{}\!\ ! ^{k_e } ) \vert \stackrel{i_0}{e}_{i_0 } \otimes \bigotimes_{\ell=1}^{2g+n-1 } \stackrel{i_\ell}{e}_{j_\ell } > \;.\end{aligned}\ ] ] in the language of lqg , we color the vertex @xmath172 with the intertwiner @xmath206 . due to the invariance of the intertwiner @xmath206 , we have : @xmath208,[j],[k ] } \stackrel{[i]}{f}{}\!\!_{[j]}^{[k]}(\lambda_e)\ ; \delta^{j_0}_{k_0 } \ ; \prod_{a } \stackrel{i_a}{\pi}{}\!\!^{j_a}_{k_a } ( \lambda_0^{-1 } u_a \lambda_0)\ ; \prod_{b } \stackrel{i_b}{\pi}{}\!\!^{j_b}_{k_b } ( \lambda_0^{-1 } u_b \lambda_0)\ ; \prod_{e } \stackrel{i_e}{\pi}{}\!\!^{j_e}_{k_e } ( \lambda_0^{-1 } u_e)\;.\end{aligned}\ ] ] then , we impose the @xmath6 invariance at the location of each particle and we obtain in the same way that : @xmath208,[j],[k ] } \stackrel{[i]}{f}{}\!\!_{[j]}^{[k]}(1)\ ; \delta^{j_0}_{k_0 } \ ; \prod_{a } \stackrel{i_a}{\pi}{}\!\!^{j_a}_{k_a } ( \lambda_0^{-1 } u_a \lambda_0)\ ; \prod_{b } \stackrel{i_b}{\pi}{}\!\!^{j_b}_{k_b } ( \lambda_0^{-1 } u_b \lambda_0)\ ; \prod_{e } \stackrel{i_e}{\pi}{}\!\!^{j_e}_{k_e } ( \lambda_0^{-1 } u_e \lambda_e)\;.\end{aligned}\ ] ] finally , the invariance under the action of @xmath209 on each ( spinless ) particle but the observer implies directly that @xmath210 for all @xmath211 .
yet @xmath212 are functions on the sphere @xmath213 ( spherical harmonic functions ) and we can write them as follows : @xmath214 the space of such invariant functions will be denoted @xmath215 .
it inherits the measure from @xmath44 due to left and right invariance of the haar measure and then it possesses a natural pre - hilbert structure .
its completion is the partial kinematical hilbert space @xmath216 and it is a sub - hilbert space of @xmath169 .
moreover , it is a straightforward to show that any element of @xmath217 is a kinematical partial hilbert space .
finally , the proposition ( [ pkin ] ) follows immediately .
@xmath218 let us propose several remarks concerning the partial kinematical hilbert space . \1 .
first of all , note that one can precise the proposition in the sense that the isomorphism holds when the partial kinematical hilbert space is defined with elements @xmath103 which are distributional for the observer variable and functional for the others . to clarify this claiming ,
let us illustrate it in the case of the sphere with particles .
indeed , the fact that the observer is colored with an intertwiner means that any spin - network states can be written as : @xmath219 it is then clear that to make @xmath220 a function on @xmath221 , it is enought to require that @xmath141 is a distribution on the first argument and a function for the others .
we made implicitely this asumption in the proof of the proposition .
\2 . as a secong remark
, we underline the fact that the proposition ( [ pkin ] ) is valid whatever the choice of the observer among the different particles . indeed ,
each partial kinematical hilbert space are isomorphic one to the other .
in fact , the previous proposition is still valid when the base point @xmath153 is not a point of @xmath122 . to understand this point , we recall that @xmath216 is a sub - hilbert space of @xmath169 and there exists a projector @xmath222 given by a product @xmath223 defined by : @xmath224 for any @xmath176 associated to a function @xmath165 .
note that @xmath225 or @xmath226 is the haar measure on the group @xmath6 and @xmath227 the haar measure on the group @xmath171 .
thus , any state @xmath228 is a partial kinematical state if there exists a state @xmath176 associated to a function @xmath165 such that : @xmath229 therefore , after some calculations it is direct to see that there exists @xmath230 $ ] such that : @xmath231 where the function @xmath232 is related to the function @xmath141 by the expression : @xmath233 @xmath234
the choice of @xmath235 corresponds to the choice of the observer .
therefore , the proposition ( [ pkin ] ) is valid for any choice of the base point . in the sequel
, we will assume that the observer is the particle @xmath172 .
\3 .
the final remark is that one can recover the kinematical hilbert space from the partial kinematical hilbert space as follows : @xmath236 where @xmath171 is the symmetry group which acts by right multiplication on the variable @xmath237 ( [ calc ] ) .
the next step is the construction of the partial physical hilbert space @xmath238 . for that purpose ,
we start by defining the extractor @xmath239 formally defined by the expression : @xmath240 \ ; \overline{\psi(a ) } \ ; { \cal k}(a ) \ ; \varphi(a ) \;.\end{aligned}\ ] ] the measure on the space of discrete connections @xmath38 is constructed from the @xmath0 haar measure @xmath241 via the relation @xmath242 = d\mu^{\otimes 2(g+n)}$ ] for @xmath38 is isomorphic to @xmath243 copies of @xmath0 .
the kernel @xmath244 is a distribution defined by : @xmath245 \ ; \equiv \ ; \delta(\lambda_0 h(m_0 ) \lambda_0^{-1 } \prod_{e=1}^{n-1 } u_e \lambda_e h(m_e ) \lambda_e^{-1 } u_e^{-1 } \prod_{(a , b)}[u_{a},u_{b } ] ) \;.\end{aligned}\ ] ] we have introduced the usual notation @xmath246=u_au_bu_a^{-1}u_b^{-1}$ ] ; @xmath247 denotes the set of the pairs of non - contractible loops ; and the elements @xmath248 are in the cartan ( diagonal ) torus of @xmath0 fixed by the mass @xmath249 by @xmath250 in the @xmath0 fundamental representation .
the extractor @xmath8 defines a bilinear form on the space @xmath251 ; it is denoted @xmath252 and is given by : @xmath253 in fact , this bilinear form defines a scalar product for it satisfies the following properties : 1 .
it is definite positive : the kernel @xmath254 is a delta distribution on the group and therefore can be obtained as a limit of positive functions on @xmath6 .
2 . it is convergent as soon as the stability condition @xmath255 is satisfied ; we have the following bound : @xmath256 we have introduced the notation @xmath257 for the @xmath258 norm on the group @xmath0 .
the bound ( [ bound ] ) is a direct consequence of the plancherel theorem for @xmath0 and of the expression of the @xmath0 character in any @xmath259 irrep , i.e. @xmath260 .
the vector space of partial physical states is defined by @xmath261 .
it is clear that the kernel @xmath244 is in fact a distribution on the space @xmath215 and therefore @xmath262 . in particular ,
the kernel @xmath263 is co - invariant under the transformations @xmath264 for any @xmath40 and @xmath265 .
thus , @xmath263 depends only on the equivalent classes @xmath266 where @xmath267 is the @xmath268 haar measure . for convenience
, it will be useful to introduce the notation @xmath269 for the normalized measure on the 2-sphere .
the bilinear form ( [ hform ] ) defines a pre - hilbert structure on @xmath261 whose completion ( up to null - vectors ) is the partial physical hilbert space @xmath238 . when restricted on @xmath238 , the scalar product ( [ hform ] )
is called the partial physical scalar product and is denoted @xmath270 .
[ philbert ] let @xmath271 .
one defines a partial physical states @xmath272 where @xmath273 was introduced in the proposition ( [ pkin ] ) .
therefore , there is a map @xmath274 when @xmath217 is endowed with the following hilbert structure : @xmath275 @xmath276 ) \ ; g(a , b,\tilde{z})\;,\ ] ] the previous map is an isometry .
note that we have introduced the notation @xmath277 , @xmath278 , @xmath279 and @xmath280 , @xmath281 for the measures .
_ proof .
_ the proof of the proposition ( [ pkin ] ) tells us that any invariant cylindrical function @xmath103 is completely caracterized by a function @xmath271 ( to be more precise , the vector space @xmath215 corresponds to the set of polynomial functions on @xmath282 and after completion with respect to the kinematical measure one sees @xmath215 as the set of functions on @xmath282 ) . yet , by definition , the partial physical hilbert space @xmath238 is the image of @xmath217 by the operator @xmath8 and then any element of @xmath238 can be written as @xmath283 \psi(a ) \in \text{cyl}_{pinv}(\gamma;g)^*$ ] .
it is then natural to identify @xmath238 with the space @xmath217 .
the last point of the proposition is a direct consequence of the right / left invariances of the @xmath0 haar measure .
@xmath218
let us finish this section with some remarks .
remark 1 .
the isometry given in the previous proposition holds for any choice of minimal graph @xmath36 and any choice of observer . each partial physical hilbert space @xmath238 where @xmath36 is any minimal graph are isomorphic one to the other and therefore are equivalent . for that reason
, we will use the notation @xmath284;g)$ ] to denote the partial physical hilbert space where @xmath285=(m_0,\cdots , m_{n-1})$ ] .
we have emphasized the explicit characteristics of the partial physical hilbert space : the topological structure of @xmath1 , the colors of the boundaries , i.e. the masses of the particles and the gauge group @xmath6 of course .
remark 2 .
the order of the particles matters in the definition of @xmath284;g)$ ] .
the space @xmath284;g)$ ] is completely and uniquely caracterised by a given flat graph of the type ( figure [ minimalgraph ] ) where there is an order between the links .
if one modifies the order between the links , then the definition the structure of the partial physical hilbert space is slightly modified .
however , two different orders define isomorphic partial physical hilbert space . for instance , if we consider a sphere @xmath286 with three particles associated to the masses @xmath287 , then the following map is an isometry : @xmath288 it is indeed easy to verify that @xmath289 .
one can establish a more general property that we will discuss in the next section .
remark 3 .
the physical hilbert space @xmath290;g)$ ] is easily obtained as the coset : @xmath291;g ) \ ; \equiv \ ; { \cal h}_{pphys}^{g , n}([m];g)/s^1\;.\ ] ] the action of @xmath171 on @xmath284;g)$ ] is naturally defined by : @xmath292;g ) & \longrightarrow & { \cal h}_{pphys}^{g , n}([m];g ) \\ ( h , f ) & \longmapsto & f^{h}\ ; : \ ; y,\tilde{z } \ ; \mapsto \ ; f^h(a , b,\tilde{z } ) \ ; = \ ; f(h^{-1}ah , h^{-1}bh , h^{-1}\tilde{z})\;. \end{aligned}\ ] ] we have used the notations of the proposition ( [ philbert ] ) with @xmath293 and so on .
therefore , the physical hilbert space is obviously a sub - hilbert space of @xmath284;g)$ ] and can be obtained as the image of the following projector : @xmath294;g ) \ ; \longrightarrow \ ; { \cal h}_{phys}^{g , n}([m];g ) \;\;\;\;\;\;\ ; f \ ; \longmapsto \tilde{f}(a , b,\tilde{z } ) \ ; = \ ; \int dh \ ; f(h^{-1}ah , h^{-1}bh , h^{-1}\tilde{z})\;.\end{aligned}\ ] ] this is in fact a trivial application of refined algebra techniques in the compact @xmath171 case .
the hilbert structure of @xmath290;g)$ ] is defined from this map as the kernel @xmath244 is invariant by the action ( [ s1action ] ) .
remark 4 . from the very definition of the partial physical hilbert space as a subset of @xmath295 , @xmath238 is in fact a subset of @xmath296^{2 g } \otimes \mathbb c[s^2]^{n-1}$ ] where @xmath296 $ ] is the group algebra and @xmath297=\mathbb c[g]/u(1)$ ] is called the 2-sphere algebra .
the group algebra is defined as the algebra of formal sum of group elements . as @xmath6
is compact there is a trivial isomorphism which identifies the spaces @xmath296^{2 g } \otimes \mathbb c[s^2]^{n-1}$ ] and @xmath174 .
remark 5 .
let us recall that any particle @xmath158 $ ] is classically completely caracterized by its mass @xmath249 and the direction of its momentum given by @xmath186 .
the particle @xmath172 plays the role of the observer and the variable @xmath237 does not appear anymore in the definition of partial physical states .
one can interpret this observation by the fact that we work in the rest - frame of the observer and then @xmath237 is set to the identity : the momenta of the other particles are defined in that frame . besides , one can caracterized the particle @xmath40 by an element @xmath298 such that @xmath299 , i.e. @xmath300 is an element of the conjugacy class of the element @xmath248 , denoted @xmath301 : @xmath302 thus , any element of @xmath284;g)$ ] can be viewed as an element of @xmath303 . using the kirillov formulae for the @xmath0 haar measure , @xmath304 the partial physical scalar product between two states @xmath103 and @xmath305 respectively associated to the functions @xmath141 and @xmath20 is then given by : @xmath306)\ ; g(a , b , k).\nonumber\end{aligned}\ ] ] we recall that the functions @xmath141 and @xmath20 are non - null if @xmath307
. this reformulation of the partial physical hilbert space will be convenient in the next section .
remark 6 .
the last remark concerns the normalization of the partial physical scalar product . whose is defined only up to a constant @xmath308)>0 $ ] .
the constant depends a priori on the genus @xmath20 of @xmath1 , the number of particles @xmath4 and the masses of the particles . by a direct calculation
, one shows that the squared norm of the identity function @xmath309 \in { \cal h}_{pphys}^{g , n}([m];g)$ ] is given by the volume of the space of flat @xmath0 connections on the punctured surface @xmath1 , i.e. : @xmath310 \vert \ !
\vert_{pphys}^2 \ ; = \ ; \lambda(g , n,[m])\ ; \sum_{k=1}^{+\infty}\ ; { k^{2 - 2g - n}}\ ; \prod_{e=0}^{n-1 } \frac{\sin(km_e ) } { \sin(m_e ) } \ ; .\end{aligned}\ ] ] at this point , there is not physical reason to impose the value of @xmath311 .
but , we can show that the value of @xmath308)$ ] is completely determined by the coefficients @xmath312)$ ] and the problem to fix it reduces to the fixation of the normalization factor in the ( minimal ) case of the sphere with 3 punctures .
the method is similar to the one proposed in @xcite and is given in the appendix b : if we assume that @xmath312)=\lambda$ ] is a constant independent of the masses , then we show that @xmath308)=\lambda^{n+2g-2}$ ] . |
quantum dots ( qds ) containing discrete energy levels which can be tuned easily in experiment have been attracting considerable attention continuously , and the study of quantum transport properties through qds has become an interesting field which serves as a testing ground for the physical phenomena , such as kondo @xcite and fano effects@xcite .
recent developments on the study include coulomb on - site repulsion @xcite , spin - flip effect caused by the magnetic leads @xcite or rotating magnetic field @xcite , and phonon - assisted tunneling in the molecular qd @xcitewith time - dependent field or ac gate voltage @xcite .
electron transports coherently through an open qd can be investigated in the standard procedure by means of nonequilibrium green function ( ngf ) @xcite , the rate equation @xcite and some other methods@xcite .
the current characteristic properties change dramatically if the qd is connected to additional devices , so various kinds of hybrid qd systems have been studied in the literature .
two aspects mainly involved in those hybrid systems are the ferromagnetic and superconductor properties which evoke the spin degree of freedom of electrons .
the f - qd - f system , i.e. , qd coupled with two ferromagnetic leads , has been studied as spin - valve device @xcite . in the superconductor case , the andreev reflection , conserving the momentum of the reflected particle but not the charge , plays a dominant role at the interface between a normal metal and a superconductor . using multichannel s - matrix method , beenakker demonstrated the resonant andreev tunneling in the normal metal - normal qd - superconductor ( n - nqd - s ) hybrid system in the zero - bias limit @xcite ; later claughton et al .
extended this theory to the finite - bias case @xcite .
sun et al . explored the resonant andreev reflection in this system with multienergy levels in normal qd and found that some extra peaks originating from the andreev reflection superimposed on the conventional current plateaus @xcite . and they also investigated the photon - assisted andreev and normal tunnels in this system . with an applied time - varying external field , it was observed that a series of photon - assisted andreev tunneling peaks appeared in the curve of the average current @xmath3 vs the gate voltage with the negative peaks on the left - hand side and positive peaks on the right - hand side of the original resonant peak in the absence of the external field @xcite .
another interesting hybrid system consists of ferromagnet and superconductor leads coupled with a normal qd ( f - qd - s ) . in ref .
@xcite , zhu et al . provided an efficient method for introducing a spin degree of freedom in this system , and also derived the general current formula of the andreev reflection in a quantum dot connected with two ferromagnetic and one superconductor leads ( f@xmath4 , f@xmath5-qd - s ) @xcite .
considering both the coulomb interaction and the multienergy levels in qd , feng and xiong studied the transport properties in the f - qd - s system 19 .
very recently , cao et al . considered the spin - flip effect in qd of a f - qd - s system and found that the single - resonance peak of the andreev reflection conductance versus the gate voltage would spill into a double - resonance peak with the increasing of the spin - flip scattering strength @xcite . thus far , most of studies relating to andreev reflection have been focused on a single - qd system . on the other hand ,
the electron transport through a double - quantum - dot ( dqd ) , especially the parallel - coupled double - quantum - dot ( pdqd ) , is an interesting subject which may involve the quantum phase interference .
an open pdqd threaded by a magnetic flux is regarded as an aharonov - bohm ( ab ) interferometer in which the coherence of the electron remains partly even in the presence of coulomb repulsion@xcite .
tunable coupling strength between two parallel dots makes the transport properties more complex and interesting@xcite .
the tunable andreev reflection in an ab interferometer has been investigated recently by peng et al.@xcite . in this paper
we shall study the electron transport through a n - pdqd - s structure , i.e. , a parallel - coupled double - quantum dot that is connected with two leads , one is a normal metal and the other is a superconductor . the device threaded by two magnetic fluxes ( see fig.1 ) with the interdot coupling can be regarded as a double ab interferometer . by using the ngf method , the current @xmath6 and the probability of the andreev reflection are derived .
the paper is organized as follows : in sec .
ii , we present the model hamiltonian and derive the general formula of the current and the probability of the andreev reflection with ngf method . in sec .
iii , the properties of the current due to andreev reflection depending on bias voltage , magnetic flux , gate voltage , and interdot coupling strength are studied in detail , and several dufferent effects are predicted .
finally , a brief summary is given in sec .
the hamiltonian of the system schematically shown in fig .
1 can be decomposed into four parts of the following form : @xmath7where @xmath8@xmath9 describes the left normal metal lead , @xmath10 ( @xmath11 ) is the corresponding annihilation ( creation ) operator , and @xmath12 is the bias voltage .
we assume the chemical potential of right lead @xmath13 @xcite .
@xmath14 describes the right superconductor lead with the energy gap @xmath15 .
@xmath16 models the coupled double quantum dots , @xmath17 is the interdot coupling strength , and @xmath18 is the corresponding phase shift induced by the magnetic flux . here
we consider the single energy level in each quantum dot .
@xmath19 is the tunneling part of hamiltonian , @xmath20 ( @xmath21 ) , independent of momentum @xmath22 ( @xmath23 ) , is the hopping strength between @xmath24th quantum dot [ i. e. the quantum dot with energy level @xmath25 ( @xmath26 ) ] and left ( right ) lead , and @xmath27 is the phase shift of the electron tunneling from the left lead to first quantum dot , where * * @xmath28where @xmath29 * * with * * @xmath30 * ( * @xmath26 * ) * being the magnetic flux in the quantum unit * * @xmath31 * * and * * @xmath32 * . * the current from the left lead to a quantum dot can be written as @xcite @xmath33e^{-i\varepsilon ( t_{1}-t^{\prime } ) } , \eqnum{3}\end{aligned}\ ] ] where we define @xmath34 and @xmath35 .
because of the right superconductor lead , it is convenient to introduce 4@xmath364 matrix representation in which @xmath37 and @xmath38 have the forms , @xmath39 @xmath40 where @xmath41 @xmath42 @xmath43 .
so eq . ( 3 ) can be rewritten as @xmath44\}_{11 + 33 } , \eqnum{6}\]]where @xmath45with @xmath46 , @xmath47 , here we consider only a spin - up current @xmath48 [ @xmath49 is easily obtained from @xmath50 by exchanging the spin index ] . to calculate @xmath51 ,
we start from the dyson equation , @xmath52 in which @xmath53 is the retarded green function for the isolated quantum dots , and can be easily obtained as , @xmath54self - energy @xmath55 consists of three parts , such that , @xmath56 where @xmath57 results from the tunneling coupling between qd and the normal metal lead , @xmath58 between qd and superconductor lead , and @xmath59 is from interdot coupling contribution . under the wide - bandwidth approximation @xcite ,
the self - energy becomes @xmath60and @xmath57 is written as @xmath61where @xmath62the self - energy from the coupling between qd and right superconductor lead is given by @xcite , @xmath63where @xmath64and the retarded green function of the isolated superconductor lead @xmath65 is read as @xmath66with @xmath67 being the density states of the right lead in a normal metal state .
we define @xmath68 @xmath42 @xmath69 . by substituting eqs .
( 9)(13 ) into eq .
( 8) , the green function @xmath51 can be derived .
it is , however , convenient to transform eq .
( 8) into the energy representation by making use of the fourier transformation , @xmath70then solving eq .
( 8) in the energy representation , we get ( see appendix a ) @xmath71 here , @xmath72 .
@xmath73 are the elements of the matrix @xmath74 , and the detailed calculation of @xmath75 is given in appendix a. the green function @xmath51 is obtained after making an inverse fourier transformation as @xmath76 in the following , we derive the smaller green function using the keldysh equation @xmath77where , @xmath78with @xmath79and @xmath80where @xmath81@xmath82 is the ratio of superconducting density of states to the normal density of states @xcite .
according to eq .
( 20 ) , @xmath83 can be obtained by substituting eqs .
( 21)-(23 ) and the green functions @xmath84 * , * @xmath85 * * into eq .
( 20 ) with the results of matrix elements @xmath86 and @xmath87 shown in appendix b. the total current , i.e. , the sum of both spin - up and spin - down parts , is found by substituting the retarded green function @xmath88 and the smaller green function @xmath89 shown in appendix b into eq .
( 6 ) , and the result is @xmath90where @xmath91t_{ar},\]]and @xmath92\widetilde{\rho } _ { r}(\varepsilon ) t_{lr}.\]]@xmath93 is seen to be the current due to the andreev reflection contribution and the probability of the andreev reflection @xmath94 is given by @xmath95 .
\nonumber\end{aligned}\]]as explained in refs .
@xcite , the term @xmath96 [ @xmath97 * d*escribes an electron with spin - up tunnelling from the left lead through quantum dot 1 ( 2 ) to the superconductor lead and reflecting a hole back to the left lead through the quantum dot 1 ( 2 ) .
while the term * * @xmath98*[*@xmath99 * ] * is for the spin - up electron tunneling from left lead through dot 1 ( 2 ) to the superconductor lead and a hole reflecting back to the left lead through the dot 2 ( 1 ) .
there exist other channels in which the tunnel electron and reflecting hole go through both two dots . the remaining part in eq .
( 26 ) describes the interference effect of all possible multichannels of the device .
the transmission probability from left to right leads reads @xmath100 \nonumber \\ & & + \gamma _ { 2}^{l}[\gamma _ { 1}^{r}\left\vert \widetilde{g}% _ { 31}^{r}(\varepsilon ) \right\vert ^{2}+\gamma _ { 1}^{r}\left\vert \widetilde{% g}_{32}^{r}(\varepsilon ) \right\vert ^{2 } \eqnum{27 } \\ & & + \gamma _ { 2}^{r}\left\vert \widetilde{g}_{33}^{r}(\varepsilon ) \right\vert ^{2}+\gamma _
{ 2}^{r}\left\vert \widetilde{g}_{34}^{r}(\varepsilon ) \right\vert ^{2 } ] \nonumber \\ & & + 2% % tcimacro{\func{re}}% % beginexpansion \mathop{\rm re}% % endexpansion [ -\gamma _ { 1}^{l}\gamma _ { 1}^{r}e^{-i\frac{\phi } { 2}}\frac{\delta } { % \left\vert \varepsilon \right\vert } \widetilde{g}_{11}^{r}(\varepsilon ) % \widetilde{g}_{12}^{r\ast } ( \varepsilon ) + \cdots ] .
\nonumber\end{aligned}\ ] ] containing many terms which are understood .
we are mainly interested in the physics of the andreev reflections . as a matter of fact only the andreev reflection process contributes to the current ( i. e. , @xmath101 ) at zero temperature , if @xmath102 in the following section , we analyze time evolution of the probability of the andreev reflection @xmath94 and the andreev reflection current @xmath93 as well depending on dc bias voltage , gate voltage , magnetic flux , and interdot coupling strength .
first of all we transform the coupled - qd system into an effective decoupled one with bonding and antibonding dressed states of the qd molecule using the following transformation @xcite , @xmath103 where , @xmath104 and @xmath105 are the annihilation operators of the bonding and antibonding dressed states for the qd molecule . and @xmath106 is defined as 1/2tan@xmath107[2@xmath108/(@xmath109-@xmath110 ) ] .
thus the hamiltonian for coupled double quantum dot @xmath16 is diagonalized as @xmath111where @xmath112 . \eqnum{30}\]]considering the tunneling hamiltonian between double qd and the left lead , we obtain the linewidth @xmath113 and @xmath114 corresponding to the antibonding and bonding states coupled to the left lead , respectively , as @xmath115 in the numerical analysis we set @xmath116 as the energy unit .
the energy dependence of reflection probability @xmath94 on the interdot coupling strength is shown in fig .
@xmath117 for the symmetry [ with @xmath118 @xmath119 , fig.2(a ) ] and asymmetry [ with @xmath120 , @xmath121 , fig.2(b ) and @xmath122 @xmath123 , @xmath124 , fig.2(c ) ] cases , respectively .
other parameters are chosen as @xmath125 , and @xmath126 it is seen in fig . @xmath127
that if the two qds are completely decoupled ( i.e. , @xmath128 ) only a single peak appears ( solid line ) , and when @xmath129 , two peaks emerges at @xmath130 due to the existence of two energy levels .
the maximum of peaks is getting lower when the coupling constant @xmath108 is increasing . * * for the asymmetry case fig .
@xmath131 shows that there are no significant two peaks .
@xmath132 exhibits the complex spectra due to the possible overlap of the broad levels similar to the interpretation of fano effect in a parallel coupled double qd connected to two normal metal leads system @xcite . and
when @xmath17 is large enough such that the broad two energy levels are completely separated , the two single peaks remain at @xmath133and @xmath134 .
it is interesting to see the effect of the magnetic flux on the andreev reflection and fig .
@xmath135 presents the dependence of the andreev reflection probability on the magnetic flux @xmath136 with @xmath137 and therefore @xmath138 .
the height of peaks at @xmath130 can reach @xmath139 when the magnetic flux is nonzero , and magnetic flux also has an influence on the width of the peaks , i. e. , on the linewidth function @xmath140 , here we choose @xmath141 for a solid line , and @xmath142 for a dotted line , the height of peaks are not changed .
electron phase coherence is shown in fig . @xmath143 . when @xmath128 , i. e. , there is no the interdot coupling , the oscillation period of the andreev reflection probability vs magnetic flux is @xmath144 ( solid line ) , while @xmath145 , the period changes to @xmath146when @xmath137 [ see fig . @xmath147 . here , we choose the different energy eigenvalues such that @xmath148 , since at those values the probability of the andreev reflection reaches maximum .
to see the reason of the period changing , we give a general result of the oscillation period when @xmath149 .
the channel paths contain two parts due to the andreev reflection in the hybrid system .
one is the incident electron from the left lead to the superconductor in fig .
@xmath150 , the other is the reflecting hole from the superconductor to the left lead in fig . @xmath151 . the phase shift of electron in the case ( 3 ) in fig .
@xmath150 is @xmath152 , and the corresponding phase shift of hole in the case ( 1 ) in fig .
@xmath151 is @xmath153 , so the total shift of the channel ( 3 ) for the elcetron and ( 1 ) for the hole is @xmath154 .
the interference effect corresponds to all the channel paths reads @xcite @xmath155where parameters @xmath156 are the amplitudes of channel paths . with @xmath157,@xmath158 @xmath159 , eq .
( 33 ) shows the oscillation periods might have @xmath160 , @xmath2 , @xmath161 , @xmath162 , @xmath163 , @xmath164 , @xmath165 , @xmath166 , @xmath167 , @xmath168 , and @xmath169 , suggesting that the oscillation period of reflect probability should be @xmath170 .
the linear depedence of the ratio @xmath0 is further confirmed in fig .
4(b ) , it shows that the oscillation period is @xmath171 for @xmath172 ( solid line ) , @xmath173 for @xmath174 ( dotted line ) , and @xmath175 for @xmath176 ( dashed line ) .
the linear manner of the osicllation period is the same as the one in ref .
22 , however , the mechanism is different : there the transmitted current is analyzed .
@xmath177 shows current - gate - voltage curve ( @xmath93 vs @xmath178 ) for the case that @xmath179 , @xmath180 , @xmath181 , and the bias voltage @xmath182 .
we assume that the energy levels in both dots are connected to the same voltage through the capacitive coupling . when @xmath183 ( solid line ) , there exist two smaller peaks at @xmath184 , and @xmath185 located symmetrically beside the larger peak at @xmath186 with the height that is two times higher than the height of the smaller peaks . * * the interpretation of the phenomena is similar to that in ref .
@xcite such that when @xmath184 the energy level @xmath109 matches with chemical potential @xmath187 in the right superconductor lead and the andreev reflection takes place at this level .
when @xmath186 two energy levels have a symmetric localization about @xmath188 ( @xmath110 is @xmath189 above @xmath188 while @xmath109 is @xmath189 below @xmath188 ) and both @xmath109 and @xmath110 are below the bias voltage , so the andreev reflection that involves those two levels leads to the height of the peaks that are two times higher than that containing only one energy level*. * the interpretation is the same for the gate voltage @xmath190 where the energy level @xmath110 matches with @xmath188 .
some new properties are observed for the coupled case @xmath129 : ( 1 ) the location of the middle peak does not change , while the positions of the left and right peaks shift towards the left and right sides , respectively , with the increasing coupling strength @xmath108 .
( 2 ) the heights of the left and the middle peaks are suppressed significantly ; meantime , the height of the right peak is enhanced .
( 3 ) the widths of the left and middle peaks decrease while the width of the right peak widened .
these interesting phenomena can be explained with eq .
( 30)-(32 ) .
the energy level of dressed state @xmath191 gets larger and @xmath134 becomes smaller with increasing coupling strength @xmath108 , and at the end @xmath191 is higher than energy level of bared state @xmath110 , when @xmath134 is smaller than @xmath192 .
so when the gate voltage takes action on the dressed states , the left peak runs towards the left while the right peak goes opposite , however , the locations of the two peaks remain symmetric about the middle peak .
we assume @xmath193 @xmath194 , @xmath195 , where @xmath196 is introduced due to the coupling strength @xmath108 .
so when the gate voltage becomes @xmath197 , the left peak emerges and with increasing @xmath178 ( up to @xmath198 ) , both two dressed states levels locate symmetrically about @xmath188 .
thus , the location of middle peak does not change at @xmath186 , no matter how the parameter @xmath199 varies .
the linewidth of dressed states , @xmath113 , increases with the increasing of @xmath108 , while @xmath114 decreases under the parameters considered here , and hence , the right peak gets wider and left peak becomes narrower with the increasing of @xmath108 .
the linewidths @xmath140 are directly connected with the coupling strengths between qd and the leads .
the widening of the right - side peak means the increasing of coupling strength between dressed state @xmath191 and the leads .
therefore , the height of the right - side peak is enhanced , and opposite to this , the height of the left - side peak is suppressed .
the suppression of the middle - peak height suggests that the andreev resonance involving two energy levels gets maximum when two energy levels are symmetric .
@xmath200 show a tunable andreev reflection current depending on the magnetic flux when @xmath129 . in fig .
@xmath201 , we fix @xmath202 , and tune the value of @xmath18 for different curves .
one can see that the height of the right peak is enhanced and the height of the left one is suppressed while the height of the middle peak does not change with the increase of @xmath18 .
it is interesting to notice that the left peaks has essentially no changes when @xmath138 ( solid line ) and @xmath203 ( dotted line ) , while the right peak changes conspicuously with respect to the case without the magnetic flux .
figure @xmath204 has the opposite effect compared with fig .
@xmath201 , namely , the height of the left peak is suppressed while the height of the right one is enhanced and the middle peak does not vary with @xmath18 .
we fix @xmath205 , and vary values of magnetic flux @xmath136 in fig .
@xmath206 , it is shown that the height of three peaks can be tuned simultaneously .
we have studied the tunable andreev reflection in a double ab interferometer in terms of the nonequilibrium green function method and observed several features of the reflection current .
it is found that the oscillation period of the andreev reflection probability with the magnetic flux is @xmath2 when interdot coupling vanishes ( @xmath128 , in this case the system reduces to a single interferometer ) , while it is @xmath170 for our double interferometer ( @xmath129 ) where @xmath0 is the ratio of two parts magnetic fluxes , i.e. @xmath207 .
* * the andreev reflection current peaks can not only be tuned by coupling strength @xmath108 , but also by magnetic flux .
this work was supported by the national natural science foundation of china under grant no .
in this appendix , we present the derivation of the green function of eq . ( 18 ) in detail . from eqs .
( 8)-(13 ) and ( 17 ) we have the following formulas for the matrix elements of green function @xmath208 in the energy representation : @xmath209 @xmath210 @xmath211 @xmath212 where @xmath213 and @xmath214 solving those coupled equations , the results are obtained as @xmath215 where @xmath216 @xmath217 @xmath218 @xmath219 here , @xmath220 with the same procedure we have @xmath221 @xmath222 @xmath223 @xmath224 so the final results are @xmath225 where @xmath226 @xmath227 @xmath228 @xmath229
the complete expression of matrix elements of the green function @xmath230 is @xmath231\ } \nonumber \\ & & + i\int \frac{d\varepsilon } { 2\pi } f_r(\varepsilon ) \widetilde{\rho } % _ r(\varepsilon ) \{\gamma _ 1^r\left| \widetilde{g}_{11}^r(\varepsilon ) \right| ^2+\gamma _ 1^r\left| \widetilde{g}_{12}^r(\varepsilon ) \right| ^2 \nonumber \\ & & + \gamma _ 2^r\left| \widetilde{g}_{13}^r(\varepsilon ) \right| ^2+\gamma _ 2^r\left| \widetilde{g}_{14}^r(\varepsilon ) \right| ^2 \eqnum{b1 } \\ & & + 2% % tcimacro{\func{re } } % % beginexpansion \mathop{\rm re } % endexpansion [ -\gamma _ 1^re^{-i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{11}^r(\varepsilon ) \widetilde{g}_{12}^{r*}(\varepsilon ) \nonumber \\ & & + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{-i\frac \phi 2}\widetilde{g}% _ { 11}^r(\varepsilon ) \widetilde{g}_{13}^{r*}(\varepsilon ) + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{i\frac \phi 2}\widetilde{g}_{12}^r(\varepsilon ) % \widetilde{g}_{14}^{r*}(\varepsilon ) \nonumber \\ & & -\gamma _
2^re^{i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{13}^r(\varepsilon ) \widetilde{g}_{14}^{r*}(\varepsilon ) ) -% \sqrt{\gamma _ 1^r\gamma
_ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{11}^r(\varepsilon ) \widetilde{g}_{14}^{r*}(\varepsilon ) \nonumber \\ & & -\sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{12}^r(\varepsilon ) \widetilde{g}_{13}^{r*}(\varepsilon ) ] \ } , \nonumber\end{aligned}\ ] ] @xmath232 \nonumber \\ & & + i\int \frac{d\varepsilon } { 2\pi } f_r(\varepsilon ) \widetilde{\rho } % _ r(\varepsilon ) \{[\gamma _ 1^r\widetilde{g}_{11}^r(\varepsilon ) -\gamma _
1^re^{i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } \widetilde{g}% _ { 12}^r(\varepsilon ) + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{i\frac \phi 2}% \widetilde{g}_{13}^r(\varepsilon ) \nonumber \\ & & -\sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{14}^r(\varepsilon ) ] \widetilde{g}_{13}^a(\varepsilon ) + [ -\gamma _ 1^re^{-i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{11}^r(\varepsilon ) + \gamma _ 1^r\widetilde{g}% _ { 12}^r(\varepsilon ) \eqnum{b2 } \\ & & -\sqrt{\gamma _ 1^r\gamma _
2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{13}^r(\varepsilon ) + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{-i\frac % \phi 2}\widetilde{g}_{14}^r(\varepsilon ) ] \widetilde{g}_{23}^a(\varepsilon ) \nonumber \\ & & + [ \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{-i\frac \phi 2}\widetilde{g}% _ { 11}^r(\varepsilon ) -\sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } \widetilde{g}_{12}^r(\varepsilon ) + \gamma _ 2^r% \widetilde{g}_{13}^r(\varepsilon ) \nonumber \\ & & -\gamma _
2^re^{-i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{14}^r(\varepsilon ) ] \widetilde{g}_{33}^a(\varepsilon ) + [ -% \sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{11}^r(\varepsilon ) \nonumber \\ & & + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{i\frac \phi 2}\widetilde{g}% _ { 12}^r(\varepsilon ) -\gamma _
2^re^{i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } \widetilde{g}_{13}^r(\varepsilon ) + \gamma _ 2^r% \widetilde{g}_{14}^r(\varepsilon ) ] \widetilde{g}_{43}^a(\varepsilon ) \ } , \nonumber\end{aligned}\ ] ] @xmath233 \nonumber \\ & & + i\int \frac{d\varepsilon } { 2\pi }
f_r(\varepsilon ) \widetilde{\rho } % _ r(\varepsilon ) \{[\gamma _ 1^r\widetilde{g}_{31}^r(\varepsilon ) -\gamma _
1^re^{i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } \widetilde{g}% _ { 32}^r(\varepsilon ) + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{i\frac \phi 2}% \widetilde{g}_{33}^r(\varepsilon ) \nonumber \\ & & -\sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{34}^r(\varepsilon ) ) ] \widetilde{g}_{11}^a(\varepsilon ) + [ -\gamma _ 1^re^{-i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{31}^r(\varepsilon ) + \gamma _ 1^r\widetilde{g}% _ { 32}^r(\varepsilon ) \eqnum{b3 } \\ & & -\sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{33}(\varepsilon ) + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{-i\frac \phi 2}\widetilde{g}_{34}^r(\varepsilon ) ] \widetilde{g}_{21}^a(\varepsilon ) \nonumber \\ & & + [ \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{-i\frac \phi 2}\widetilde{g}% _ { 31}^r(\varepsilon ) -\sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } \widetilde{g}_{32}^r(\varepsilon ) + \gamma _ 2^r% \widetilde{g}_{33}^r(\varepsilon ) \nonumber \\ & & -\gamma _
2^re^{-i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{34}^r(\varepsilon ) ] \widetilde{g}_{31}^a(\varepsilon ) + [ -% \sqrt{\gamma _ 1^r\gamma _
2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{31}^r(\varepsilon ) \nonumber \\ & & + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{i\frac \phi 2}\widetilde{g}% _ { 32}^r(\varepsilon ) -\gamma _ 2^re^{i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } \widetilde{g}_{33}^r(\varepsilon ) + \gamma _ 2^r% \widetilde{g}_{34}^r(\varepsilon ) ] \widetilde{g}_{41}^a(\varepsilon ) \ } , \nonumber\end{aligned}\ ] ] @xmath234\ } \nonumber \\ & & + i\int \frac{d\varepsilon } { 2\pi } f_r(\varepsilon ) \widetilde{\rho } % _ r(\varepsilon ) \{\gamma
_ 1^r\left| \widetilde{g}_{31}^r(\varepsilon ) \right| ^2+\gamma _ 1^r\left| \widetilde{g}_{32}^r(\varepsilon ) \right| ^2 \eqnum{b4 } \\ & & + \gamma _ 2^r\left| \widetilde{g}_{33}^r(\varepsilon ) \right| ^2+\gamma _ 2^r\left| \widetilde{g}_{34}^r(\varepsilon ) \right| ^2 \nonumber \\ & & + 2% % tcimacro{\func{re } } % % beginexpansion \mathop{\rm re } % endexpansion [ -\gamma _
1^re^{i\frac \phi 2}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{32}^r(\varepsilon ) \widetilde{g}_{31}^{r*}(\varepsilon ) + % \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{i\frac \phi 2}\widetilde{g}% _ { 33}^r(\varepsilon ) \widetilde{g}_{31}^{r*}(\varepsilon ) \nonumber \\ & & + \sqrt{\gamma _ 1^r\gamma _ 2^r}e^{i\frac \phi 2}\widetilde{g}% _ { 32}^r(\varepsilon ) \widetilde{g}_{34}^{r*}(\varepsilon ) -\gamma _
2^re^{i% \frac \phi 2}\frac \delta { \left| \varepsilon \right| } \widetilde{g}% _ { 33}^r(\varepsilon )
\widetilde{g}_{34}^{r*}(\varepsilon ) \nonumber \\ & & -\sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{32}^r(\varepsilon ) \widetilde{g}_{33}^{r*}(\varepsilon ) -% \sqrt{\gamma _ 1^r\gamma _ 2^r}\frac \delta { \left| \varepsilon \right| } % \widetilde{g}_{31}^r(\varepsilon ) \widetilde{g}_{34}^{r*}(\varepsilon ) ] \}. \nonumber\end{aligned}\ ] ] the results here show that the above matrix elements of @xmath235 do not depend on time @xmath236 , because for the n - qd - s hybrid system , the current should be independent of time for the dc bias @xcite . c. a. stafford , and n. s. wingreen , phys .
76 , 1916 ( 1996 ) ; chi - keung lui , baigeng wang , and jian wang , phys . rev .
b 70 , 205316 ( 2004 ) ; mou yang , and shu - shen li , phys . rev .
b 70 , 045318 ( 2004 ) .
s. a. gurvitz , and ya .
s. prager , phys .
b 53 , 15932 ( 1996 ) ; bing dong , h. l. cui , and x. l. lei , phys . rev .
b 69 , 035324 ( 2004 ) ; xin - qi li , wen - kai zhang , ping cui , jinshu shao , zhongshui ma , and y. j. yan , phys . rev .
b 69 , 085315 ( 2004 ) ; t. m. stace , and s. d. barrett , phys .
92 , 136802 ( 2004 ) .
2 : the probability of the andreev reflection @xmath94 vs @xmath239 in the unit of the energy gap @xmath15 for ( a ) the symmetry case ( @xmath240 @xmath119 ) and asymmetry case ( b ) @xmath122 @xmath241,@xmath242 , ( c ) @xmath243 , @xmath124 . the solid , dotted , dot - dashed , and dashed lines correspond to @xmath108=0 , 0.02 , 0.04 , and 0.08 , respectively .
4 : the periodic oscillation of the andreev reflection probability @xmath249 with the magnetic flux @xmath136 ( @xmath250 , @xmath251 , @xmath252 ) , ( a ) @xmath253 .
solid line : @xmath128 , @xmath254 , dotted line : @xmath255 , @xmath148 , and dashed line : @xmath256 , @xmath257 .
( b ) @xmath255 , @xmath148 .
solid line : @xmath258 , dotted line : @xmath259 , and @xmath143 for dashed line .
fig . 6 : andreev reflection current @xmath93 vs gate voltage @xmath178 ( @xmath260 , @xmath180 , @xmath118 @xmath119 , and @xmath181 ) for bias voltage @xmath182 , @xmath128 ( solid line ) , @xmath255 ( dotted line ) , and @xmath256 ( dashed line ) . fig . 7
: andreev reflection current vs gate voltage with the fixed total magnetic flux that @xmath202 for @xmath255 .
@xmath138 ( solid line ) , @xmath203 ( dotted line ) , @xmath261 ( dot - dashed line ) , @xmath160 ( dashed line ) . | quantum transport through a double aharonov - bohm - interferometer in the presence of andreev reflection is investigated in terms of the nonequilibrium green function method with which the reflection current is obtained .
tunable andreev reflection probabilities depending on the interdot coupling strength and magnetic flux as well are analysised in detail .
it is found that the oscillation period of the reflection probability with respect to the magnetic flux for the double interferometer depends linearly on the ratio of two parts * * magnetic * * fluxes * * @xmath0 * , * i.e. @xmath1 * , * while that of a single interferometer is * * @xmath2*. * the coupling strength not only affects the height and the linewidth of andreev reflection current peaks vs gate votage but also shifts the peak positions .
it is furthermore demonstrated that the andreev reflection current peaks can be tuned by the magnetic fluxes .
pacs : 73.23.-b ; 74.45.+c ; 73.63.kv |
it was suggested a long time ago ( shvartsman 1970 ; michel 1972 ; anile , & treves 1972 ; maraschi , reina , & treves 1974 ) that accretion of a hydrogen plasma onto a neutron star must necessarily imply the presence of an electrostatic field @xmath2 under stationary conditions .
the field @xmath2 arises because the radiation drag , due to the outflowing flux , acts efficiently only on electrons while practically all the mass is in protons .
the electrostatic field compensates the different acceleration of the two species and allows electrons and protons to move at the same speed .
obviously @xmath2 is stronger the larger is the deviation from free fall , a situation which is expected when the luminosity approaches the eddington limit .
the typical values of the charge @xmath3 associated with the electric field are not large enough to influence the space
time around the compact object ( michel 1972 ) , but its presence may nevertheless have interesting astrophysical consequences .
shvartsman ( 1970 ) proposed that such a system could act as an electrostatic accelerator for positrons , for which the electrostatic and radiative forces have the same direction . balancing energy gains and losses ,
due mainly to inverse compton with soft ambient photons , he estimated that positrons may be accelerated to energies of a few tens of mev .
although the `` shvartsman accelerator '' is definitely ingenious , it received little attention in the past , possibly because no obvious positron source seemed available near neutron stars .
however , positrons may be generated by the radioactive decay of the neutron star crust elements subject to the bombardment of accreting particles ( reina , tarenghi , & treves 1974 ; bildsten , shapiro , & wasserman 1992 ) .
pair production may be also expected in accretion models with shocks ( see e.g. shapiro , & salpeter 1975 ) or in hydrostatic atmospheres kept at mildly relativistic temperatures by comptonization ( turolla , _ et al .
_ 1994 ; zane , turolla , & treves 1996 ) . here
we reconsider shvartsman s idea , examining the role of cooling processes different from inverse compton ( bremsstrahlung , non radiative scattering ) and the possibility that accelerated positrons survive annihilation with atmospheric electrons , escaping to infinity .
let us consider spherical accretion onto a neutron star ( @xmath4 , @xmath5 cm ) and for simplicity suppose that the accreting material is ionized hydrogen . neglecting general relativistic corrections ,
the emerging luminosity is related to the accretion rate by @xmath6 when @xmath7 approaches the eddington limit @xmath8 , the flow dynamics is strongly affected by the radiative force ( see e.g maraschi , reina , & treves 1974 , 1978 ; miller 1990 ; park & miller 1991 ; zampieri , turolla , & treves 1993 , hereafter ztt )
. owing to the reduced effective gravity , the flow deviates from free fall and , for @xmath9 , a settling behaviour appears close to the star surface , where the velocity decreases with radius .
even a small deviation from free
fall , which is always present for a non vanishing luminosity , implies that electrons and protons are subject to different accelerations so that an electrostatic field @xmath2 is built up around the star ( shvartsman 1970 ; zeldovich , & novikov 1971 ; michel 1972 ) .
although , in principle , the same argument applies to spherical accretion onto a black hole , the efficiency is so low ( @xmath10 in absence of magnetic fields , see e.g. nobili , turolla , & zampieri 1991 ) that departures from free fall are really negligible for any reasonable value of the accretion rate .
the expression for @xmath2 can be easily derived considering the forces acting on the two species ( @xmath11 ) @xmath12 and @xmath13 where @xmath14 denotes the luminosity in the frame comoving with the accretion flow . equating the two accelerations we get @xmath15 at large radii , where @xmath16 , the previous expression coincides with the result obtained by maraschi , reina , & treves ( 1974 , 1978 ) analyzing the flow equations for the two species @xmath17 an estimate of the largest potential difference
can be easily obtained integrating equation ( 4 ) assuming @xmath18 @xmath19 which shows that , in principle , extremely relativistic positrons may be expected to reach infinity .
here we consider a scenario in which the number density of @xmath20 is not zero at some radius @xmath21 .
positrons are immediately accelerated by the electric field and gain energy at a rate @xmath22 where @xmath23 is the particle velocity . at the same time , positrons are accelerated by the radiative force and suffer radiative losses through inverse compton ( ic ) scatterings with the radiation bath ; @xmath24 and @xmath25 bremsstrahlung ; non radiative scatterings with the flow electrons . in this section
we concentrate on ic losses , which is the only cooling process taken into account by shvartsman , and we show that , even in this simple case , ultra
relativistic energies are not expected to be reached . in the next section
we will verify that bremsstrahlung losses are negligible , but that elastic scatterings may be very efficient in transfering energy to ambient electrons and will ultimately stop the positron flow .
the energy lost by a positron for ic scattering with a radiation field of energy density @xmath26 and radiation pressure @xmath27 is @xmath28 while the gain due to the radiative force is @xmath29 where @xmath30 is the total flux divided by @xmath31 ( gurevich , & rumyantsev 1965 , all the radiation moments are here evaluated in the lab frame ) . for an isotropic radiation field ,
@xmath32 , @xmath33 and the expression for the net power reduces to @xmath34 ( see e.g. rybicki , & lightman 1979 ) . accordingly , the positron energy changes with radius as @xmath35 introducing the dimensionless quantities @xmath36 , @xmath37 , @xmath38 , and @xmath39 , where @xmath40 is the schwarzschild radius , equation ( 10 ) can be written as @xmath41\ , .\end{aligned}\ ] ] whenever @xmath42 the solution of equation ( 11 ) relaxes immediately to the reduced solution @xmath43\ ] ] irrespective of the initial particle energy . if no other processes prevent positrons from escaping ( see the discussion in section 4 ) , the terminal lorentz factor can be obtained integrating numerically equation ( 11 ) once the radiation field in the accreting medium is known . in section 5 numerical results based on the accretion model described in ztt will be presented .
equation ( 12 ) also allows a close comparison between our result and shvartsman s one .
the contribution of electrostatic acceleration can be easily obtained dropping @xmath44 in equation ( 10 ) which is , formally , equivalent to put @xmath45 in equation ( 12 ) .
an upper limit for the terminal lorentz factor in this case is given by @xmath46 , which corresponds to a kinetic energy @xmath47 kev .
this is much less than the value of tens of mev proposed by shvartsman .
the point in shvartsman s original analysis that should be reconsidered is the calculation of the ic losses , which were estimated in a very crude way by multiplying the energy exchange in each collision times an `` optical depth '' parameter , @xmath48 . even assuming that the total number of collisions suffered by the positron in traveling a radial distance @xmath49 , @xmath50 , is indeed related to shvartsman s equation ( 2 ) ( and this is questionable )
, @xmath50 should be @xmath51 , since @xmath52 .
this implies that shvartsman s compton cooling is underestimated by a factor @xmath53 .
moreover the final particle energy is evaluated assuming that no losses are present for @xmath54 which results in a even larger underestimate of ic cooling .
in the previous section we have derived the energy equation for positrons , including only ic cooling . in section 5
we show that , even under this favourable assumption , the positrons terminal lorentz factor is @xmath55 ( see table 1 ) . in order to establish
if indeed these mildly relativistic particles may reach infinity , we need a more thorough understanding of the physical conditions in the flow close to the star surface . in our scenario four different kinds of particles are present : `` primary '' electrons and protons of the accreting material , with the same number density @xmath56 and temperature @xmath57 , together with positrons and electrons , with number density @xmath58 and temperature @xmath59 , that are created in the atmosphere by photon photon or particle collisions .
we will limit our discussion to the case in which @xmath60 , so the number density of electrons is @xmath61 .
the first issue that should be addressed concerns the probability that positrons annihilate with ambient electrons .
this question arises quite naturally since , at variance with shvartsman s scenario in which highly relativistic energies are expected , now the particle energy is not high enough to ensure that the cross section for pair annihilation becomes negligible . in the non relativistic limit ,
the annihilation rate for positrons and electrons with arbitrary distribution functions is ( svensson 1982 ) @xmath62 we introduce a characteristic time @xmath63 for annihilation defined as @xmath64 since @xmath63 depends on @xmath56 , it gives the probability of annihilation on a scale of order the density lengthscale in the accretion flow , @xmath65 . on the other hand , a positron of energy @xmath66 covers a distance @xmath67 in a time @xmath68 @xmath69 and the particle survives annihilation if @xmath70 . a numerical comparison between these two timescales
will be given in the next section for different values of both the emergent luminosity and @xmath21 .
here we only note that , in the more favourable situations in which particles are immediately accelerated up to @xmath71 , this condition is met if @xmath72 this limit is fairly easy to satisfy for near eddington accretion since @xmath73@xmath74@xmath75 close to the star surface .
other cooling mechanisms that could influence the maximum positron energy are radiative and non radiative scattering with ions and electrons in the accretion flow .
the positron
ion bremsstrahlung energy loss is given by @xmath76 where @xmath77 is the charge of the ions .
this process acts on a characteristic timescale @xmath78 which , compared with @xmath63 , yields @xmath79 so , for pure hydrogen , bremsstrahlung losses are negligible for particles surviving annihilation and free free could represent a competitive cooling process only for large @xmath77 . on the other hand , the energy exchange due to collisions is indeed important and can modify substantially the picture outlined up to now .
let us introduce the energy loss rate @xmath80 for a test particle in a plasma of protons and electrons as @xmath81 and focus on the non relativistic regime .
the more efficient process is radiationless , elastic scattering between positrons and electrons ; positron proton equipartition needs a larger timescale by a factor @xmath82 .
the energy loss rate can be obtained from the expression of diffusion coefficients , which , in the limit @xmath23 much larger than the proton velocity , yields ( see e.g. rosenbluth , & sagdeev 1983 ) @xmath83 \ , , \end{aligned}\ ] ] where @xmath84 is the coulomb logarithm and @xmath85 the electron velocity .
the rate @xmath80 is not necessarily positive since a fast particle tends to lose energy to the plasma and a slow one to gain energy .
initial differences in the mean velocities of positrons and electrons are eliminated by collisions in a time @xmath86 .
the importance of collisions can be quantified comparing @xmath87 with the dynamical timescale given by expression ( 15 ) .
in order to investigate the effects of electrostatic plus radiative acceleration on particles injected in different atmospheric layers , we have solved equation ( 11 ) varying both the accretion luminosity and the injection radius @xmath21 .
the run of the flow variables has been taken from the one fluid models computed by ztt for near eddington accretion of a `` cold '' , pure scattering medium onto neutron stars .
the solutions are plotted , together with the reduced solution ( equation 12 ) , in figures 1a , b , c for @xmath880.3 , 0.6 and 0.9 , respectively . for
any given luminosity , we have considered three different values of @xmath21 , which correspond to a scattering optical depth @xmath89 ( see table 1 ) , with the exception of the case @xmath90 , for which the maximum optical depth is @xmath91 at the star surface .
this simple choice allows us to mimic the acceleration of positrons initially created in different atmospheric layers , and , increasing the luminosity , to visualize the process in the presence of more and more expanded atmospheres . in order to account for the anisotropy of the radiation field
, we introduced the closure @xmath92 , adopting for @xmath93 the same expression used by ztt , @xmath94 + 1 /3 $ ] .
the final energies in all the cases we considered are listed in table 1 .
figures 2a , b , c show the comparison between @xmath68 and @xmath95 for the same choice of the parameters .
as can be seen from the figures , the reduced solution is always attained at small radii , but at the same time annihilation acts efficiently in the inner dense atmospheric layers preventing positrons created in optically thick regions from escaping .
the chance to surviving annihilation increases for particles created in photospheric regions and above , although positrons starting at @xmath96 reach infinity with a lower lorentz factor .
the larger terminal velocities are expected for low values of @xmath7 when the photosphere is closer to the star surface .
figures 3a , b , c show @xmath97 together with @xmath87 for two different values of the gas temperature , since the efficiency of coulomb collisions strongly depends on the difference between the test particle and the electron velocities .
we note that only for @xmath98 kev is @xmath99 negative in an extended region near @xmath21 , so the possibility that positrons gain additional energy owing to collisions appears remote . on the other hand , under the more realistic assumption that @xmath100 kev , as expected in typical x ray sources , we find that mildly relativistic positrons are `` de facto '' prevented from reaching infinity , all differences between the mean energies of the two species being washed out by the collective effect of soft collisions .
in this paper we have reconsidered acceleration of positrons near a neutron star accreting close to the eddington limit . the radiative force and the electrostatic thrust due to the non vanishing electric field around the neutron star
are opposed by inverse compton losses and elastic scattering with electrons of the accreting material .
ultimately , these are the main agents responsible for both the acceleration and deceleration of positrons .
the maximum energy which can be achieved by positrons owing to the electrostatic acceleration alone is limited to @xmath47 kev which differs substantially from shvartsman s estimate of few tens of mev .
the main contribution to the particle s acceleration is therefore provided by the radiative force , which acts equally on electrons and positrons .
we have also shown that the largest terminal lorentz factors are still @xmath55 , even if this additional gain is accounted for .
a first consequence is that the maximum energies attained by positrons are in a range where pair annihilation is still important , at least for particles created at @xmath101 .
moreover , elastic scatterings with cold electrons in the accretion flow oppose the particle acceleration and provide an additional source of trapping .
this suggests that only a fraction of positrons will actually leave the star , while the rest should annihilate and eventually contribute to a narrow 511 kev line .
furthermore , no multiplication through electromagnetic showers either on the accreting material or on the photons can be expected , so the positron production depends mainly on the injection process .
moreover , we note that the same conclusions discussed above can be applied to particles injected below @xmath102 ( i.e. below @xmath103 ) , quite independently of the initial conditions . if the source is close to the eddington limit , even if positrons are injected with energies @xmath104 mev , like in radioactive decay , because of the efficiency of ic losses , their final energy will be limited to @xmath105 kev and the possibility of outstream or annihilation will not change .
ccccc 0.3 & 1.5 & 1 & 19 & 358 0.3 & 1 . & 2 & 5 & 358 0.3 & 0.1 & 141 & 0.006 & 296 0.6 & 5 . & 1.5 & 61 & 293 0.6 & 1 . & 13 & 0.5 & 293 0.6 & 0.1 & 855 & 0.002 & 151 0.9 & 5 . & 11 & 5 & 198 0.9 & 1 . & 123 & 0.06 & 198 0.9 & 0.1 & 5000 & 0.0002 & 50 anile , a.m. , & treves , a. 1972 , , 19 , 411 bildsten , l. , salpeter , e.e . , & wasserman , i. 1992 , , 384 , 143 gurevich , l. , & rumyantsev , a.a .
1965 , sov .
jetp , 20 , 1233 maraschi , l. , reina , c. , & treves , a. 1974 , , 35 , 389 maraschi , l. , reina , c. , & treves , a. 1978 , , 66 , 99 michel , f.c .
1972 , , 15 , 153 miller , g.s .
1990 , , 356 , 572 nobili , l. , turolla , r. , & zampieri , l. 1991 , , 383 , 250 park , m.g . , & miller , g.s . , 1991 , , 371 , 708 reina , c. , treves , a. , & tarenghi , m. 1974 , , 32 , 317 rosenbluth , m.n .
, & sagdeev , r.z .
, 1983 , handbook of plasma physics , vol . 1 , ( amsterdam : north holland ) rybicki , g.b . , &
lightman , a.p . ,
1979 , radiative processes in astrophysics , ( new york : wiley ) shapiro , s.l . , & salpeter , e.e .
1975 , , 198 , 671 shvartsman , v.f . 1970 ,
astrofizika , 6 , 309 svensson , r. 1982 , , 258 , 335 turolla , r. , zampieri , l. , colpi , m. , & treves , a. 1994 , , 426 , l35 zampieri , l. , turolla , r. , & treves , a. 1993 , , 419 , 311 ( ztt ) zane , s. , turolla , r. , & treves , a. 1996 , , submitted zeldovich , ya.b . , & novikov , i.d . 1971 , relativistic astrophysics , vol.1 , ( chicago : the university of chicago press ) | as first shown by shvartsman ( 1970 ) , a neutron star accreting close to the eddington limit must acquire a positive charge in order for electrons and protons to move at the same speed .
the resulting electrostatic field may contribute to accelerating positrons produced near the star surface in conjunction with the radiative force .
we reconsider the balance between energy gains and losses , including inverse compton ( ic ) , bremsstrahlung and non radiative scatterings .
it is found that , even accounting for ic losses only , the maximum positron energy never exceeds @xmath0 kev .
the electrostatic field alone may produce energies @xmath1 kev at most .
we also show that coulomb collisions and annihilation with accreting electrons severely limit the number of positrons that escape to infinity . |
study of temporal dynamics has greatly advanced our understanding of correlated electron systems for such dynamics may provide information about essential energy scales associated with the elementary excitations and collective modes in these systems.@xcite the newly developed time - resolved angle - resolved photo - emission spectroscopy ( trarpes ) is a powerful technique to directly measure momentum resolved electronic dynamics,@xcite and has been applied to study ultrafast dynamics in charge density wave materials,@xcite high - temperature superconductors(htscs),@xcite and topological insulators.@xcite smallwood _ et al . _
@xcite have used the technique to study gap and quasiparticle ( qp ) population dynamics in optimally doped d - wave htsc bi@xmath0sr@xmath0cacu@xmath0o@xmath1 .
their results clearly show that qp s relaxation rate is dependent on the momentum and fluence .
these intriguing observations have challenged the understanding of non - equilibrium dynamics in the superconducting ( sc ) state , and called for theoretical explanations for further explorations on htscs with trarpes technique . in this paper
, we apply a two - temperature model @xcite to theoretically study the transient energy distribution and qp relaxation dynamics of d - wave superconductor measured in trarpes .
our theory explains the momentum - resolved dynamics of the photoexcited qps , and is in good agreement with experiments .
in a typical trarpes experiment , a pump laser pulse is first applied to excite the investigated sample into a non - equilibrium state , and arpes technique is then used to measure the temporal electronic dynamics of the sample . presently , trarpes technique has a time - resolution @xmath2 fs.@xcite in the relaxation process of the laser - excited d - wave superconductors , there are two essential time scales determined by the intrinsic interactions , namely , @xmath3 due to the electron - electron ( e - e ) scattering and @xmath4 due to the electron - phonon ( e - p ) scattering .
we consider a typical case where @xmath5 , which is suitable for the present experiment resolution of @xmath6 . in this case
, the e - e scattering thermalizes the excited electronic subsystem into a quasi - equilibrium state in the time scale of @xmath7 .
then the detailed non - equilibrium processes of the electronic subsystem will be washed out during the time @xmath6 , and the quasi - equilibrium state can be characterized by the effective temperature @xmath8 . in the longer time scale @xmath9 ,
the extra energy in the electronic subsystem will be dissipated to the lattice subsystem through electron - phonon interaction , and @xmath8 will decrease to the lattice temperature @xmath10 .
thus the observed gap and qp dynamics of the superconductors will be associated with the time - dependent electronic temperature @xmath8 .
this simple two - temperature scenario has been applied to describe the relaxation dynamics in htsc , @xcite and will be applied to study the trarpes here . the dependence of @xmath8 on the time @xmath11 in the two - temperature model is not known _ a priori _ , but may be determined from the experimental measurements or derived from the microscopic theory . for simplicity , we assume that the dominant process for the electronic system to dissipate energy is the pairwise recombination of qps into cooper pairs , and the rothwarf - taylor equation@xcite for the total qp number still holds , which is supported by the previous thz conductivity measurement.@xcite further exploiting the fact that the total qp density is approximately proportional to the square of electronic temperature @xmath8 , then the equation of the electronic temperature @xmath8 is estimated as@xcite @xmath12 here , the pump laser is assumed to be applied at time @xmath13 , which promotes the electronic temperature from @xmath10 to @xmath14 . for given @xmath10 ,
higher laser fluence will drive the sample into higher excited state and lead to larger @xmath14 .
then eq .
( [ decay ] ) describes the decay of the electronic temperature from @xmath14 to @xmath10 , and the parameter @xmath15 is proportional to the recombination rate of qps@xcite and will be determined by fitting to the experimental results later on . denotes the angle to describe momentum cutline in arpes measurements .
( b ) temperature dependence of sc gap , determined from eq .
( [ gap ] ) .
( c ) and ( d ) : calculated change in the line - momentum - integrated arpes intensity , @xmath16 , defined in eq . ( [ deltai ] ) , between @xmath8=90k and @xmath10=20k , along the momentum cutlines @xmath17 and @xmath18 , respectively .
( e ) and ( f ) : experimental result of the intensity change @xmath16 along the momentum cutlines @xmath17 and @xmath18 , respectively , taken from ref . for pump fluence 5 @xmath19j/@xmath20 .
blue color represents intensity gain and red color intensity loss . ]
now we consider the quasi - equilibrium electronic state characterized by temperature @xmath8 and the resulting consequences in trarpes measurements .
since the trarpes has been reported on the optimally - doped bi@xmath21sr@xmath21cacu@xmath21o@xmath1(bi-2212 ) samples,@xcite our calculations below will be specified to this compound although the main conclusions are expected to apply to general d - wave superconductors .
we use a simple tight - binding model suggested in ref . to describe the normal state energy dispersion , with @xmath22 where the coefficients are @xmath23 . the fermi surface ( fs ) determined by @xmath24
is shown in fig .
[ bandgap](a ) .
a momentum cutline denoted by an angle @xmath25 for arpes measurements is also shown .
we consider a simple d - wave sc gap function for bi-2212 , @xmath26 , and further assume a simple bcs form for the temperature dependence of @xmath27 , which is determined by the self - consistent equation@xcite @xmath28}{(\xi^{2}+\delta^{2})^{1/2}}d\xi.\label{gap}\end{aligned}\ ] ] here , @xmath29 is the boltzmann constant , @xmath30 denotes the density of states at the fermi level of one spin orientation , @xmath31 characterizes the strength of pair potential for the electrons within the cutoff energy @xmath32.@xcite with the weak - coupling condition @xmath33 and the relation @xmath34 , where @xmath35 is the sc critical temperature , eq .
( [ gap ] ) gives the universal function of @xmath36 on @xmath37,@xcite as shown in fig .
[ bandgap](b ) .
while this form is for conventional s - wave superconductor , the deviation from d - wave superconductors is not significant to change the qualitative results below . with the bare band dispersion relation @xmath38 and the gap function @xmath39 , the qp excitation energy @xmath40 is determined as @xmath41 .
the key quality measured in the trarpes is the spectral function @xmath42 in terms of the energy distribution curves , which takes the form below with a lorentzian lineshape@xcite @xmath43.\label{specfuc}\end{aligned}\ ] ] here , @xmath44 characterizes the broadening of the spectral line and is assumed to be a constant .
@xmath45 are the coherence factors . in our calculations
, we set @xmath46 k , @xmath47 ev , and @xmath48 ev .
when the electronic temperature @xmath8 changes , the spectrum function @xmath42 also changes due to the temperature - dependent sc gap .
several physical qualities based on the spectral functions are usually given in the trarpes experiments , and will be calculated below in order to compare with the experimental results .
one is the line - momentum - integrated arpes intensity @xmath49 ( arpes intensity hereafter ) along a momentum cutline @xmath50 ( the choice in our calculations is shown in fig .
[ bandgap](a ) ) , @xmath51 where @xmath52 is the fermi - dirac ( fd ) distribution function , which depends on the electronic temperature @xmath8 .
note that @xmath42 is also a function of @xmath8 via the sc gap function .
in the two - temperature scenario , as @xmath8 decreases from @xmath14 to the equilibrium temperature @xmath10 , the sc gap increases and the thermal distribution of qps is suppressed , then the number of excited qps is reduced .
this process is reflected in the time - evolution of the arpes intensity @xmath49.@xcite a more relevant quantity is the change of the arpes intensity between the two temperatures @xmath8 and @xmath10 , @xmath53 it is convenient to introduce the line - momentum - integrated density of states(dos hereafter ) along the momentum cutline @xmath50 , @xmath54 , then @xmath16 is decomposed into two terms , @xmath55 where @xmath56 is the change of the distribution function , and @xmath57 is the change of the dos along @xmath50 .
this decomposition is vital for us to understand the results below . in fig .
[ bandgap](c ) and ( d ) , we show the results of @xmath16 for the system for two sets of fs angles , i.e. a diagonal cutline @xmath17 ( nodal cutline ) and an off - nodal cutline @xmath18 , respectively . the two temperatures are @xmath58k , and @xmath59k .
@xmath60 is strongly dependent on the cutline . for the nodal cutline , @xmath60 is approximately antisymmetric with respect to @xmath61 , similar to the change of the fermi distribution function @xmath62 .
in particular , @xmath63 at @xmath64 . for the non - nodal cutline ,
the shape is far from the antisymmetric , and @xmath65 is finite and positive , and @xmath66 occurs at @xmath67 .
our calculations are in good agreement with the experimental data,@xcite which are reproduced in fig .
[ bandgap](e ) and ( f ) for comparison .
the strong fs angle dependence of the arpes intensity is associated with the symmetry of the sc gap function . for the nodal cutline of ( @xmath17 ) , the second term in eq .
( [ diffint ] ) vanishes .
furthermore , @xmath68 and @xmath62 is of substantial value only within a window of @xmath69 .
since @xmath70 does nt vary sharply near @xmath64 , we have @xmath16 approximately antisymmetric as numerically shown in fig .
[ bandgap](c ) . for the off - nodal case , the second term in eq .
( [ diffint ] ) plays an important role , and @xmath16 is no longer antisymmetric . at @xmath64
, we have @xmath71 according to eq .
( [ specfuc ] ) and @xmath72 , thus @xmath73 .
another important quality given in trarpes experiments is the angle - resolved qp density . following the experiments,@xcite we define @xmath74 as the integral of the arpes intensity @xmath49 over @xmath61 above the fermi energy along the momentum cutline @xmath50 at temperature @xmath10 , and @xmath75 is the change of @xmath74 when the electronic temperature increased from @xmath10 to @xmath8 , i.e. @xmath76 below we analysis the dynamics of @xmath75 to further explain the trarpes experiments .
it has been observed that the photoexcited qp density is strongly dependent on the fs angle @xmath25 similar to the d - wave sc gap function.@xcite according to our discussions above , the electronic temperature has been increased from @xmath10 to @xmath14 after the laser pumping , because the detailed electronic dynamics to reach such a quasi - equilibrium state has been washed out due to the large time resolution of the arpes measurements . then the sc gap is reduced and even closed if @xmath77 and more qps are also excited .
thus we can calculate the photoexcited qp density according to eq .
( [ idi ] ) , and the results of @xmath78 for the temperatures @xmath14=90k , 80k , 70k , 60k , 50k and @xmath10=20k are given in fig .
[ iniint](a ) .
the calculation results reproduce the angle - dependence relation observed in experiments,@xcite as shown in fig .
[ iniint](c ) .
notice that higher pump fluence corresponds to higher electronic temperature @xmath14 . to understand the results , we recall that @xmath75 has both the contributions from thermal broadening and gap reduction according to eq .
( [ diffint ] ) .
we denote @xmath79 , where @xmath80 the angle dependence of @xmath81 is expected to be weak as long as @xmath82 does nt vary sharply near the fermi energy . specially , if @xmath70 is approximated by its value @xmath83 at fermi energy , @xmath81 will be independent on the fs angle @xmath25 .
while for @xmath84 , its angle dependence comes from the change of dos @xmath85 which is strongly dependent on the d - wave gap function . for the nodal case(@xmath17 ) , @xmath86 gives @xmath87 ; while in the off - nodal region , the gap reduction will increase the dos near the fermi energy and thus @xmath88 . as an example , fig .
[ iniint](b ) shows the two contributions @xmath81 and @xmath84 to @xmath78 for @xmath89k and @xmath59k .
it is found that @xmath81 is only weakly dependent on the fs angle @xmath25 , while @xmath84 strongly depends on @xmath25 in the similar way as the d - wave gap function , which supports our analysis . when the pump fluence is lower , the temperature @xmath14 is closer to @xmath10 , then the gap reduction becomes smaller , and the angle dependence of @xmath78 becomes weak . this tendency is shown in fig .
[ iniint](a ) , and is also found in the trarpes experiments in fig .
[ iniint](c ) . on the fermi surface angle @xmath25 for several temperatures
@xmath8 with @xmath59k .
( b ) thermal broadening contribution @xmath81 and gap reduction contribution @xmath84 to the relative arpes intensity change @xmath78 for temperature @xmath58k and @xmath59k .
( c ) experimental result for the photoexcited qp density , taken from ref . .
] .[iniint ] we further analyze the angle - dependent decay rate of qps . it has been observed that the qps in the off - nodal region decay faster than the ones in the nodal region.@xcite in the two - temperature scenario , the arpes measures the quasi - equilibrium electronic state characterized by the temperature @xmath90 .
we first calculate the normalized @xmath75 for different temperatures at given fs angle @xmath25 , and the results are shown in fig .
[ tempscale](a ) . here
, we have fixed @xmath59k , and varied the temperature @xmath8 from @xmath91k to @xmath92k , and then normalize @xmath75 by its value at @xmath91k .
it is found that for given temperature @xmath8 , the normalized @xmath75 decreases when the angle @xmath25 moves from the nodal region to the off - nodal region . then with the help of eq .
( [ decay ] ) , the temperature dependence of the normalized @xmath75 in fig .
[ tempscale](a ) is mapped to the corresponding time - dependence relation , as shown in fig .
[ tempscale](b ) .
here we have set @xmath93 to fit the experimental results .
our results here reproduce the observations that the decay of normalized @xmath78 depends on the fs angle@xcite shown in fig .
[ tempscale](c ) . for several fs angles with initial electronic temperature @xmath14=90k and lattice temperature @xmath10=20k .
( b ) decay curves of normalized @xmath75 for several fs angles by mapping the curves in ( a ) to the time scale in the help of eq .
( [ decay ] ) . the fitting parameter @xmath94 .
( c ) experimental decay curves of normalized @xmath75 at @xmath17 and @xmath18 respectively , taken from ref . .
] the results in fig .
[ tempscale ] can also be understood from the two contributions @xmath95 and @xmath96 in @xmath75 . for the nodal case ( @xmath17 )
, we have @xmath97 , while @xmath98 if the dos @xmath70 is approximated by its value at fermi energy .
this explains the nearly linear dependence of the normalized @xmath75 on the temperature for @xmath17 in fig .
[ tempscale](a ) . in the off - nodal case
, @xmath96 due to the gap variation contributes to @xmath75 in addition to @xmath95 .
when the temperature @xmath8 decreases , the gap becomes larger and @xmath85 decreases .
thus in the off - nodal region the normalized @xmath75 deviates from the linear temperature - dependence relation and decays faster , as shown in fig .
[ tempscale](a ) . after mapping to the time scale ,
the normalized @xmath75 then shows the angle - dependent decay rate of qps . with different initial electronic temperature @xmath14 and lattice temperature @xmath59k for @xmath17 and @xmath18 respectively .
the fitting parameter @xmath94 in eq .
( [ decay ] ) .
( c ) and ( d ) : experimental decay curves of normalized @xmath75 with different pumping fluences for @xmath17 and @xmath18 respectively , taken from ref . .
higher pumping fluence corresponds to higher initial electronic temperature @xmath14 . ]
it is also observed that the qps at fixed angle relax faster if the sample is pumped by higher fluence laser.@xcite the reason however is not the same as the angle - dependence case discussed above , considering that the superconducting gap is not involved in the nodal direction ( @xmath17 ) .
the nearly linear dependence of @xmath75 on the temperature @xmath8 shown in fig .
[ tempscale](a ) implies that the decay of the normalized @xmath75 along the cutline @xmath17 directly reflects the decay of @xmath8 .
this fluence - dependent decay behavior can be explained by the recombination process of qps into cooper pairs,@xcite and thus justifies the assumed decay equation ( [ decay ] ) for @xmath8 . with eq .
( [ decay ] ) and the parameter @xmath93 , we got the decay of the normalized @xmath75 with different initial temperatures @xmath14 for @xmath17 and @xmath18 , as shown in fig .
[ fluscale](a ) and ( b ) .
the calculation results reproduces the experimental observations that higher fluence induces faster qp decay rate , as shown in fig .
[ fluscale](c ) and ( d ) .
thus , unlike the angle - dependent decay rate of qps which is associated with the d - wave gap dynamics , the fluence - dependent decay rate of qps is due to the pairwise recombination of qps into cooper pairs .
in conclusion , we show that the main features of the trarpes observations for d - wave superconductors so far can be explained within a simple two - temperature scenario . in this picture , the effective electronic temperature affects both the thermal distribution of the quasiparticles and the superconducting gap .
the angle dependence of the photoexcited quasiparticle density and the quasiparticle decay rate are associated with the d - wave gap dynamics , while the fluence - dependent quasiparticle decay rate is attributed to the pairwise recombination of qps into cooper pairs .
different from the original explanations of the experimental results,@xcite the two - temperature scenario here does nt refer to the details of the microscopic scattering processes , which have been washed out due to the thermalization of e - e scattering in the time scale @xmath7 .
our results suggest that the phenomenological two - temperature model could be a good starting point to analyze the trarpes experiments before extracting other interesting microscopic dynamics processes .
furthermore , better time - resolution in experiment will be crucial for the development of more powerful trarpes technique .
we thank jianqiao meng for indicating us to the trarpes technique , and thank weiqiang chen , zijian yao , and hongmin jiang for helpful discussions .
this work is supported by the hong kong grants of university grant council aoe / p-04/08 and grc hku707010 , and by nsfc 11274269 .
99 d. n. basov , r. d. averitt , d. van der marel , m. dressel , and k. haule , rev .
phys . * 83 * , 471 ( 2011 ) .
l. perfetti , p. a. loukakos , m. lisowski , u. bovensiepen , h. eisaki , and m. wolf , phys .
lett . * 99 * , 197001 ( 2007 ) .
f. schmitt , p. s. kirchmann , u. bovensiepen , r. g. moore , l. rettig , m. krenz , j .- h .
chu , n. ru , l. perfetti , d. h. lu , m. wolf , i. r. fisher , and z .- x .
shen , science,*321 * , 1649 ( 2008 ) .
f. schmitt , p. s. kirchmann , u. bovensiepen , r. g. moore , j .- h .
chu , d. h. lu , l. rettig , and m. wolf , new .
* 13 * , 063022 ( 2011 ) .
r. corts , l. rettig , y. yoshida , h. eisaki , m. wolf , and u. bovensiepen , phys .
lett . * 107 * , 097002 ( 2011 ) .
j. graf , c. jozwiak , c. l. smallwood , h. eisaki , r. a. kaindl , d .- h .
lee , and a. lanzara , nat . phys .
* 7 * , 805 ( 2011 ) . c. l. smallwood , j. p. hinton , c. jozwiak , w. zhang , j. d. koralek , h. eisaki , d .- h .
lee , j. orenstein , and a. lanzara , science , * 336 * , 1137 ( 2012 ) .
y. h. wang , d. hsieh , e. j. sie , h. steinberg , d. r. gardner , y. s. lee , p. jarillo - herrero , and n. gedik , phys .
lett . * 109 * , 127401 ( 2012 ) .
a. rothwarf and b. n. taylor , phys .
* 19 * , 27 ( 1967 ) .
r. a. kaindl , m. a. carnahan , d. s. chemla , s. oh , and j. n. eckstein , phys
b * 72 * , 060510(r)(2005 ) . the original rothwarf - taylor equation for the total qp number @xmath99 in ref . reads @xmath100 , where @xmath101 is the external source term to generate qps , @xmath102 is the total number of phonon to break cooper pairs with rate @xmath103 , and @xmath104 is the recombination rate of qps into cooper pairs . in the two temperature model , we assume that the lattice temperature @xmath10 is unchanged and so is the phonon number @xmath102 .
then the above equation for @xmath99 can be simplified as @xmath105 if we neglected the source term @xmath101 and replaced the term @xmath106 by @xmath107 due to the equilibrium condition @xmath108 . with the approximate relation @xmath109 for d - wave superconductors ,
we get eq .
( [ decay ] ) and the parameter @xmath110 .
m. r. norman , m. randeria , h. ding , and j. c. campuzano , phys .
b * 52 * , 615 ( 1995 ) .
m. tinkham , _ introduction to superconductivity _
( mcgraw - hill inc , new york , 1996 ) . j. c. campuzano , h. ding , m. r. norman , m. randeira , a. f. bellman , t. yokoya , t. takahashi , h. katayama - yoshida , t. mochiku , and k. kadowaki , phys . rev .
b * 53 * , r14737 ( 1996 ) . | motivated by recent development in time - resolved angle - resolved photoemission spectroscopy ( trarpes ) for d - wave superconductors , we analyze the non - equilibrium relaxation dynamics of the laser pulse excited sample within the scenario of two - temperature model .
it is shown that the main features reported in the trarpes technique may be understood within this phenomenological picture .
the momentum dependence of the excited quasiparticle density and the relaxation rate is associated with the dynamics of the nodal d - wave superconducting gap , and the fluence dependence of the relaxation rate is related to the recombination process of quasiparticles into cooper pairs . |
hypertoric varieties ( cf .
@xcite , @xcite ) are the hyperkhler analogue of khler toric varieties .
the algebraic construction of hypertoric varieties was given by hausel and sturmfels @xcite .
modelling on their construction , in this paper we construct hypertoric dm stacks and study their orbifold chow rings . according to @xcite ,
the topology of hypertoric varieties is determined by hyperplane arrangements . in this paper
we define stacky hyperplane arrangements from which we define hypertoric dm stacks .
let @xmath0 be a finitely generated abelian group of rank @xmath1 and @xmath2 the natural projection modulo torsion .
let @xmath3 be a homomorphism determined by a collection of nontorsion integral vectors @xmath4 .
we require that @xmath5 has finite cokernel .
the gale dual of @xmath5 is denoted by @xmath6 . a @xmath7 element @xmath8 in @xmath9 and the vectors @xmath10 determine a hyperplane arrangement @xmath11 in @xmath12 .
we call @xmath13 a _ stacky hyperplane arrangement_. for @xmath3 in @xmath14 , we consider the lawrence lifting @xmath15 of @xmath5 where @xmath16 is a finitely generated abelian group with rank @xmath17 . the map @xmath18 is given by vectors @xmath19 .
the generic element @xmath8 determines a lawrence simplicial fan @xmath20 in @xmath21 .
we call @xmath22 a lawrence stacky fan and @xmath23 the lawrence toric dm stack .
the hypertoric dm stack @xmath24 associated to @xmath14 is defined as a quotient stack which is a closed substack of the lawrence toric dm stack @xmath23 , generalizing the construction of @xcite .
the stacky hyperplane arrangement @xmath14 also determines an extended stacky fan @xmath25 introduced in @xcite . here
@xmath26 is the normal fan of the bounded polytope @xmath27 of the hyperplane arrangement @xmath28 .
the toric dm stack @xmath29 defined in @xcite is the associated toric dm stack of @xmath24 . to the map @xmath30 we associate a multi - fan @xmath31 in the sense of @xcite , which consists of cones generated by linearly independent subsets @xmath32 in @xmath33 for @xmath34 ,
see section [ substack ] .
we assume that the @xmath35 .
we prove that each top dimensional cone in @xmath31 gives a local chart for the hypertoric dm stack @xmath24 .
we define a set @xmath36 consisting of all pairs @xmath37 , where @xmath38 is a cone in the multi - fan @xmath31 , @xmath39 such that @xmath40 for @xmath41 . for @xmath42
we consider a closed substack of @xmath24 given by the quotient stacky hyperplane arrangement @xmath43 .
the inertia stack of @xmath24 is the disjoint union of all such closed substacks , see section [ substack ] .
we now describe the orbifold chow ring of @xmath24 .
the multi - fan @xmath31 naturally gives a `` matroid '' @xmath44 .
the vertex set is @xmath45 , and the faces are the subsets @xmath46 such that @xmath32 are linearly independent in @xmath33 .
note that the faces of @xmath44 are the cones in @xmath31 . according to @xcite ,
the ordinary cohomology ring of the hypertoric variety corresponding to the hyperplane arrangement @xmath28 is isomorphic to the `` stanley - reisner '' ring of the matroid @xmath44 .
our result shows that the orbifold chow ring of hypertoric dm stacks is a generalization of the stanley - reisner ring of the matroid @xmath44 to the multi - fan @xmath31 .
let @xmath47 denote all the pairs @xmath48 , where @xmath49 , @xmath38 is a cone in @xmath50 such that @xmath51 and @xmath52 are rational numbers .
then @xmath47 gives rise to a group ring @xmath53=\bigoplus_{(c,\sigma)\in n^{\delta_\mathbf{\beta}}}\mathbb{q}\cdot y^{(c,\sigma)},\ ] ] where @xmath54 is a formal variable .
for any @xmath55 , there exists a unique element @xmath56 such that @xmath57 and @xmath58 , where @xmath59 are nonnegative integers .
we call @xmath60 the @xmath61 of @xmath48 . for @xmath48
we define the _ ceiling function _ @xmath62 by @xmath63 .
note that if @xmath64 , @xmath65 . for two pairs @xmath66 , @xmath67 ,
if @xmath68 is a cone in @xmath69 , define @xmath70 .
let @xmath71 be the minimal cone in @xmath69 containing @xmath72 so that @xmath73 .
we define the grading on @xmath74 $ ] as follows . for any @xmath48 , write @xmath58 , then @xmath75 where @xmath76 is the dimension of @xmath77 .
note that we have integer grading due to the fact that @xmath24 is hyperkhler . by abuse of notation
, we write @xmath78 as @xmath79 .
the multiplication is defined by @xmath80 using the property of ceiling functions we check that the multiplication is commutative and associative .
so @xmath74 $ ] is a unital associative commutative ring .
let @xmath81 be the ideal in @xmath74 $ ] generated by the elements : @xmath82 let @xmath83 be the orbifold chow ring of the hypertoric dm stack @xmath24 .
we have the following theorem : [ chowring ] let @xmath24 be the hypertoric dm stack associated to the stacky hyperplane arrangement @xmath14 .
then there is an isomorphism of graded @xmath84-algebras : @xmath85}{cir(\delta_\mathbf{\beta})}.\ ] ] the orbifold chow ring of the hypertoric dm stack @xmath24 is independent of the generic element @xmath8 .
it only depends on the map @xmath5 .
theorem [ chowring ] is proven by a direct approach .
the inertia stack of a hypertoric dm stack @xmath24 is the disjoint union of closed substacks @xmath86 for all @xmath42 . to determine the ring structure , we identify the 3-twisted sectors as closed substacks of @xmath24 indexed by triples @xmath87 in @xmath88 such that @xmath89 is a integral linear combination of @xmath90 s .
we then determine the obstruction bundle over any 3-twisted sector and prove that the orbifold cup product is the same as the product of the ring @xmath74 $ ] described above .
the multi - fan @xmath69 is equal to the simplicial fan @xmath26 in @xmath91 induced from the stacky hyperplane arrangement @xmath14 if and only if @xmath28 has @xmath92 hyperplanes @xmath93 whose normal polytope is a product of simplices .
so in this case @xmath91 is a stacky fan and the simplicial fan @xmath26 is a product of normal fans of simplices , the toric variety @xmath94 is a product of weighted projective spaces .
then by @xcite the associated hypertoric variety is the cotangent bundle of the toric variety @xmath94 .
so @xmath95 , the cotangent bundle of the toric dm stack @xmath29 .
the ring @xmath74 $ ] coincides ( as vector spaces ) with the deformed ring @xmath96^\mathbf{\sigma}$ ] as defined in @xcite .
let @xmath91 be as above .
then there is an isomorphism of @xmath84-vector spaces @xmath97 here is an example which shows that the orbifold chow ring of @xmath24 is not isomorphic as a ring to the orbifold chow ring of the associated toric dm stack @xmath29 .
consider the weighted projective stack @xmath98 which is a toric dm stack with stacky fan @xmath25 , where @xmath99 , @xmath100 is given by the vectors @xmath101 and @xmath26 is the simplicial fan in the lattice @xmath0 consisting cones @xmath102 and @xmath103 generated by @xmath104 and @xmath105 respectively .
the gale dual map @xmath106 is given by the matrix @xmath107 .
choosing generic element @xmath108 , we get a stacky hyperplane arrangement @xmath109 .
the hypertoric dm stack @xmath24 is the cotangent bundle @xmath110 whose core is the toric dm stack @xmath98 .
both @xmath111 $ ] and @xmath96^{\mathbf{\sigma}}$ ] are generated by @xmath112 , @xmath113 , and @xmath114 . according to theorem 1.1 and the main theorem in @xcite ,
their orbifold chow rings are given as follows : @xmath115 } { ( x_{1}-2x_{2},v^{2}-x_{2},vx_{1},x_{1}x_{2 } ) } \cong \frac{\mathbb{q}[v ] } { ( v^{3})},\ ] ] @xmath116 } { ( x_{1}-2x_{2},x_{1}x_{2},vx_{1},v^{2 } ) } \cong \frac{\mathbb{q}[x_{2},v ] } { ( x_{2}^{2},vx_{2},v^{2})}.\ ] ] it is easy to see that these two rings are not isomorphic .
thus the orbifold chow ring of a hypertoric dm stack is not necessarily isomorphic to the orbifold chow ring of its core .
( however , their chow rings are isomorphic , see theorem 1.1 of @xcite . )
this also proves that the orbifold chow ring has no homotopy invariance property .
we remark that the core of a general hypertoric dm stack can be singular , it is not clear how to define orbifold chow ring .
but in the case of a cotangent bundle over weighted projective space , the core is the weighted projective space and the orbifold chow ring is well - defined . on the other hand ,
the orbifold chow ring of a lawrence toric dm stack is isomorphic to its associated hypertoric dm stack , see @xcite .
computations of orbifold cohomology rings of hypertoric orbifolds in symplectic geometry have been pursued in @xcite .
this paper is organized as follows . in section [ hyperplane ]
we discuss the relation between stacky hyperplane arrangements and extended stacky fans .
we define hypertoric dm stack @xmath24 associated to the stacky hyperplane arrangement @xmath14 . in section [ hypertoricdm ] we discuss the properties of hypertoric dm stacks . in section [ substack ]
we determine closed substacks of a hypertoric dm stack .
this yields a description of its inertia stacks .
we prove theorem [ chowring ] in section [ chow ] , and in section [ app ] we give some examples . in this paper
we work entirely algebraically over the field of complex numbers .
chow rings and orbifold chow rings are taken with rational coefficients . by an orbifold we mean a smooth deligne - mumford stack with trivial generic stabilizer .
we refer to @xcite for the construction of gale dual @xmath117 from @xmath118 .
we denote by @xmath2 the natural map modulo torsion . for cones
@xmath119 in @xmath120 , we use @xmath68 to represent the set of union of the generators of @xmath121 and @xmath122 .
we would like to thank the referee for the nice comments of remark [ lawrencermk - s ] and kai behrend , megumi harada , nicholas proudfoot for valuable discussions .
in this section we define hypertoric deligne - mumford stacks , mimicking the construction of hypertoric varieties in @xcite .
we introduce stacky hyperplane arrangements .
we explain how a stacky hyperplane arrangement gives extended stacky fans .
let @xmath0 be a finitely generated abelian group and @xmath123 a map given by nontorsion integral vectors @xmath124 .
we have the following exact sequences : @xmath125 @xmath126 where @xmath127 is the gale dual of @xmath5 ( see @xcite ) .
the map @xmath127 is given by the integral vectors @xmath128 . choose a generic element @xmath129 which lies in the inage of @xmath127 and let @xmath130 be a lifting of @xmath8 in @xmath131 such that @xmath132 .
note that @xmath8 is generic if and only if it is not in any hyperplane of the configuration determined by @xmath127 in @xmath133 .
let @xmath134 be the dual of @xmath0 and @xmath135 , then @xmath136 is a @xmath1-dimensional @xmath137-vector space .
associated to @xmath8 there is a hyperplane arrangement @xmath138 in @xmath136 defined by @xmath139 the hyperplane @xmath140 so ( [ arrangement ] ) determines hyperplane arrangements in @xmath136 , up to translation induced by the choice of the lifting @xmath130 .
we call @xmath13 a _ stacky hyperplane arrangement_. it is well - known that hyperplane arrangements determine the topology of hypertoric varieties @xcite .
let @xmath141 let @xmath26 be the normal fan of @xmath27 in @xmath142 with one dimensional rays generated by @xmath143 . by reordering , we may assume that @xmath144 are the hyperplanes that bound the polytope @xmath27 , and @xmath145 are the other hyperplanes
. then we have an extended stacky fan @xmath25 defined in @xcite , where @xmath146 is given by @xmath147 , and @xmath148 are the extra data . by @xcite , the extended stacky fan @xmath91 determines a toric deligne - mumford stack @xmath29 .
it is the same stack as in @xcite .
its coarse moduli space is the toric variety corresponding to the normal fan of @xmath27 . according to @xcite , a hyperplane arrangement @xmath28 is _ simple _ if the codimension of the nonempty intersection of any @xmath149 hyperplanes is @xmath149 .
a hypertoric variety is the coarse moduli space of an _ orbifold _ if the corresponding hyperplane arrangement is simple .
let @xmath150 , see figure 1 .
the polytope @xmath27 of the hyperplane arrangement is the shaded triangle whose toric variety is the projective plane .
the extended stacky fan is given by the fan of the projective plane @xmath151 and an extra ray @xmath152 .
if for a generic element @xmath129 the hyperplane arrangement @xmath28 bounds a polytope whose normal fan is @xmath26 , then @xmath25 is a stacky fan defined in @xcite . consider the gale dual map @xmath153 in ( [ exact2 ] ) .
we denote the gale dual map of @xmath154 by @xmath155 + where @xmath21 is a lattice of dimension @xmath156 .
the map @xmath18 is given by the integral vectors @xmath157 and @xmath18 is called the lawrence lifting of @xmath5 .
[ lawrencermk - s ] consider the following commutative diagram for which the rows are exact : @xmath158^{t}}@<-.5ex>[r]\dto & ~\mathbb{z}^{2m}\ar_{[0,1]}@<-.5ex>[l]\ar_{[1,-1]}@<-.5ex>[r]\dto^{(\beta^{\vee},-\beta^{\vee } ) } & ~\mathbb{z}^{m}\ar_{[1,0]^{t } } @<-.5ex>[l]\dto^{\beta^{\vee}}\\ 0\ar@<-.5ex>[r ] & dg(\beta)\ar@<-.5ex>[l]\ar@<-.5ex>[r ] & dg(\beta),\ar@<-.5ex>[l]}\ ] ] where @xmath159^{t}$ ] represents the transpose of the matrix . taking gale dual to the above diagram
yields the following commutative diagram : @xmath160^{t}}@<.5ex>[r]\dto^{id } & ~\mathbb{z}^{2m}\ar^{[1,1]}@<.5ex>[l]\ar^{[1,0]}@<.5ex>[r]\dto^{\beta_{l } } & ~\mathbb{z}^{m}\ar^{[1,-1]^{t}}@<.5ex>[l]\dto^{\beta}\\ ~\mathbb{z}^{m}\ar@<.5ex>[r ] & n_{l}\ar@<.5ex>[l]\ar@<.5ex>[r ] & n.\ar@<.5ex>[l]}\ ] ] so from the functoriality of gale dual we have @xmath161 . since @xmath162
, we get that @xmath163 .
so @xmath157 are the vectors @xmath164 where @xmath165 are the standard bases of @xmath131 . given the generic element @xmath8 , let @xmath166 be the natural image of @xmath8 under the projection @xmath167 . then the map @xmath168 is given by @xmath169 .
for any basis of @xmath170 of the form @xmath171 , there exist unique @xmath172 such that @xmath173 let @xmath174 $ ] be the coordinate ring of @xmath175 .
let @xmath176 we put @xmath177 and @xmath178 where @xmath179 is the complement of @xmath180 and corresponds to a maximal cone in @xmath20 . from @xcite
, @xmath20 is the fan of a lawrence toric variety @xmath181 corresponding to @xmath8 in the lattice @xmath21 , and @xmath182 is the irrelevant ideal .
the construction above establishes the following a stacky hyperplane arrangement @xmath109 also gives a stacky fan @xmath22 which is called a lawrence stacky fan . from proposition 4.3 in @xcite
, @xmath20 is a simplicial fan in @xmath21 .
the rays @xmath183 , @xmath184 are generated by @xmath185 , @xmath186 .
the map @xmath18 is the map ( [ betal ] ) given by @xmath187 .
so by @xcite , @xmath22 is a stacky fan .
[ lawrencetoricdmstack ] the toric dm stack @xmath23 is called the lawrence toric dm stack . for the map
@xmath188 given by @xmath189 , there is an exact sequence @xmath190 applying @xmath191 to ( [ exact3 ] ) yields @xmath192 where @xmath193 and @xmath194 is the torus of dimension @xmath17 . from @xcite and proposition [ lawrencestacky ] ,
the toric dm stack @xmath23 is the quotient stack @xmath195 $ ] , where @xmath196 acts on @xmath197 through the map @xmath198 in ( [ exact5 ] ) .
define an ideal in @xmath199 $ ] by : @xmath200 where @xmath201 is the map in ( [ exact1 ] ) and @xmath202 is the @xmath203-th component of the vector @xmath204 . according to section 6 in @xcite
, @xmath205 is a prime ideal .
let @xmath206 be the closed subvariety of @xmath197 determined by the ideal ( [ ideal1 ] ) .
since @xmath207 acts on @xmath206 naturally and the group @xmath196 acts on @xmath206 through the map @xmath198 , we have the quotient stack @xmath208 $ ] . since @xmath209 is a closed subvariety , the quotient stack @xmath208 $ ] is a closed substack of @xmath23 , and is deligne - mumford .
the hypertoric deligne - mumford stack @xmath24 associated to the stacky hyperplane arrangement @xmath14 is defined to be the quotient stack @xmath208 $ ] .
let @xmath210 , @xmath26 the fan of projective line @xmath211 , and @xmath212 given by @xmath213 .
then the gale dual @xmath214 is given by the matrix @xmath215 $ ] .
choose a generic element @xmath216 in @xmath217 which determines the fan @xmath26 .
the stacky hyperplane arrangement is @xmath109 , @xmath218 and @xmath206 is the subvariety of @xmath219)$ ] determined by the ideal @xmath220 .
then by @xcite , the coarse moduli space is the _ crepant resolution _ of the gorenstein orbifold @xmath221 $ ] , see figure 3 .
the corresponding hyperplane arrangement @xmath28 consists of three distinct points on the real line @xmath222 , and the bounded polyhedron is two segments intersecting at one point .
so the core of the hypertoric variety is two @xmath211 intersecting at one point .
the hypertoric dm stack @xmath24 is a nontrivial @xmath223-gerbe over the _ crepant resolution _ according to the action given by the inverse of the above matrix .
if we change @xmath224 to @xmath225 , we will see an example in section [ substack ] that the hypertoric dm stack is a trivial @xmath223-gerbe over the crepant resolution .
each deligne - mumford stack has an underlying coarse moduli space . in this section
we prove that the coarse moduli space of @xmath24 is the underlying hypertoric variety .
consider again the map @xmath226 in ( [ exact2 ] ) , which is given by the vectors @xmath227 .
as in section 2 , let @xmath166 be the natural image of @xmath8 under the projection @xmath167 . then the map @xmath168 is given by @xmath169 . from the map @xmath228 we get the simplicial fan @xmath20 in ( [ fan ] ) . by @xcite ,
the toric variety @xmath181 , which is the geometric quotient @xmath229 , is the coarse moduli space of the lawrence toric dm stack @xmath23 .
the toric variety @xmath181 is semi - projective , but not projective . in @xcite , from @xmath127 and @xmath8 ,
the authors define the hypertoric variety @xmath230 as the complete intersection of the toric variety @xmath181 by the ideal ( [ ideal1 ] ) , which is the geometric quotient @xmath231 .
we have the following proposition .
the coarse moduli space of @xmath24 is @xmath230 .
let @xmath232 . by the universal property of geometric quotients ( @xcite ) , we have that @xmath233 . from lemma 3.3 in @xcite ,
the stabilizers of points in @xmath234 are the same as the stabilizers of the points in @xmath206 , which are determined by the box elements in the lawrence simplicial fan and extended stacky fan .
so we have the following diagram @xmath235\ar@{}[dr]|{\box}&~\mathcal{x}(\mathbf{\sigma_{\theta}})\dto \\
y(\beta^{\vee},\theta ) \ar @{^{(}->}[r ] & x(\sigma_{\theta}),}}\ ] ] which is cartesian .
the lawrence toric variety @xmath181 is the coarse moduli space of the lawrence toric deligne - mumford stack @xmath23 .
so @xmath24 has coarse moduli space @xmath230 . in @xcite
, the authors began with the map @xmath127 , and assumed that @xmath9 and @xmath0 are free . in our case
@xmath9 is a finitely generated abelian group , the toric variety @xmath181 is again semi - projective since @xmath20 is a semi - projective fan .
the hypertoric variety @xmath230 is the complete intersection determined by the ideal ( [ ideal1 ] ) .
this reduces to the case in @xcite when @xmath9 and @xmath0 are free . from ( [ arrangement ] ) ,
a hyperplane @xmath139 is naturally oriented .
changing the orientation of @xmath236 means changing the map @xmath5 by replacing @xmath237 by @xmath238 . [ orientation ] @xmath24 is independent to the coorientations of the hyperplanes in the hyperplane arrangement @xmath11 corresponding to the stacky hyperplane arrangement @xmath14 . note that changing coorientations _ does _ change the corresponding normal fan of the weighted polytope @xmath27 .
it suffices to prove the proposition when we change the coorientation of one hyperplane , say @xmath239 for some @xmath240
. let @xmath241 .
then we have a new stacky hyperplane arrangement @xmath242 , where @xmath243 is given by @xmath244 . using the technique of gale dual in @xcite , it is easy to check that if the gale dual @xmath127 is given by the integral vectors @xmath245 , then the gale dual @xmath246 is given by the integral vectors @xmath247
. let @xmath248 be the map given by @xmath249 if @xmath250 and @xmath251 , then we have the following commutative diagrams : @xmath252 consider the diagram @xmath253 } \rto^ { } & ~(\mathbb{z}^{2m})^{*}\dto^{[\beta^{\vee},-\beta^{\vee}]}\\ dg(\beta')\rto^ { } & dg(\beta).}\ ] ] applying @xmath254 yields the following diagram of abelian groups @xmath255 recall that @xmath206 is a subvariety of @xmath197 defined by the ideal @xmath205 in ( [ ideal1 ] ) .
when we change the coorientation of @xmath239 , the ideals do not change , so @xmath256 . by ( [ diagram1 ] ) , the following diagram is cartesian : @xmath257 where @xmath258 is determined by the map @xmath259 .
so the groupoid @xmath260 is morita equivalent to the groupoid @xmath261 .
the stack @xmath208 $ ] is isomorphic to the stack @xmath262 $ ] , and @xmath263 .
let @xmath25 be the extended stacky fan induced by @xmath14 .
the toric dm stack @xmath29 is the quotient stack @xmath264 $ ] , where @xmath265 as in @xcite , and @xmath266 is the square - free ideal of the fan @xmath26 .
so every hypertoric dm stack @xmath24 has an associated toric dm stack @xmath29 whose simplicial fan is the normal fan of the bounded polytope @xmath27 in the hyperplane arrangement @xmath28 determined by the stacky hyperplane arrangement @xmath14 .
but by proposition [ orientation ] , @xmath24 does not determine @xmath29 . consider figure 1 again .
the corresponding toric variety is @xmath151 .
if we change the coorientation of the hyperplane @xmath267 , then the corresponding normal fan @xmath26 of @xmath27 changes .
the resulting toric variety is a hirzebruch surface .
so the associated toric dm stacks are different .
but the hypertoric dm stacks are the same .
in this section we consider substacks of hypertoric dm stacks . in particular , we determine the inertia stack of a hypertoric dm stack .
let @xmath109 be a stacky hyperplane arrangement and @xmath25 the extended stacky fan induced from @xmath14 .
let @xmath24 denote the corresponding hypertoric dm stack .
consider the map @xmath268 given by @xmath269 .
let @xmath270 be a partially ordered finite set of cones generated by @xmath271 .
the partial ordering is defined by requiring that @xmath272 if @xmath38 is a face of @xmath77 .
we have the minimum element @xmath273 which is the cone consisting of the origin .
let @xmath274 be the set of all convex polyhedral cones in the lattice @xmath33 .
then we have a map @xmath275 such that for any @xmath276 , @xmath277 is the cone in @xmath33 . then @xmath278 is a simplicial multi - fan in the sense of @xcite .
recall that in section [ hyperplane ] we have the fan @xmath20 for the lawrence toric variety corresponding to @xmath279 .
let @xmath280 be the lawrence lifting of @xmath281 .
we have the following lemma .
[ lawrence ] if @xmath282 forms a cone in @xmath20 , then @xmath283 forms a cone in @xmath31 .
this can be easily proved from the definition of fan @xmath20 in ( [ fan ] ) . for a cone @xmath38 in the multi - fan @xmath31 , let @xmath284 .
then we have a quotient extended stacky fan @xmath285 , where @xmath286 is given by the images of @xmath287 s in @xmath288 .
let @xmath289 , then @xmath290 since @xmath38 is simplicial .
consider the commutative diagrams @xmath291 and @xmath292 applying the gale dual yields @xmath293 and @xmath294 since @xmath295 , the gale dual @xmath296 . and again applying the @xmath297 functor to the above two diagrams ( [ close - hypertoric1 ] ) , ( [ close - hypertoric2 ] ) yields @xmath298 and @xmath299 since @xmath129 , from the map @xmath300 in ( [ close - hypertoric1 ] ) @xmath8 induces @xmath301 in @xmath302 . from the isomorphism @xmath303 in ( [ close - hypertoric2 ] ) , we get @xmath304
. then @xmath109 gives @xmath305 whose induced extended stacky fan is @xmath306 . from ( [ close - hypertoric1]),([close - hypertoric2 ] )
we have the following diagrams @xmath307 taking @xmath254 gives @xmath308 let @xmath309 and @xmath310 the closed subvariety of @xmath311 defined by the ideal @xmath312 where @xmath313 is the dual map of @xmath314 and @xmath315 the @xmath203-th component of the vector @xmath316 . then from the definition of hypertoric dm stacks , we have @xmath317 $ ] .
we have the following result : [ closedsub ] if @xmath38 is a cone in the multi - fan @xmath31 , then @xmath86 is a closed substack of @xmath24 .
let @xmath182 be the irrelevant ideal in ( [ irrelevant ] ) .
the hypertoric stack @xmath24 is the quotient stack @xmath208 $ ] , where @xmath318 is the subvariety determined by the ideal @xmath205 in ( [ ideal1 ] ) .
taking duals to ( [ close - hypertoric1 ] ) , ( [ close - hypertoric2 ] ) we get : @xmath319 and @xmath320 let @xmath321 be the subvariety of @xmath234 defined by the ideal @xmath322 .
then @xmath321 contains the @xmath323-points @xmath324 such that the cone spanned by @xmath325 containing @xmath38 belongs to @xmath31 . from lemma
[ lawrence ] , the @xmath323-point @xmath326 in @xmath321 such that @xmath327 implies that @xmath328 or @xmath329 .
it is clear that @xmath321 is invariant under the @xmath196-action defined by ( [ exact5 ] ) .
let @xmath330 . then from ( [ ideal1]),([idealclose - hypertoric ] ) and ( [ close - hypertoric7]),([close - hypertoric8 ] ) , @xmath331 and the components
@xmath332 are determined by the choice of the generic element @xmath8 .
let @xmath333 be the inclusion given by @xmath334 . from the map @xmath335 in ( [ close - hypertoric6 ] )
, we have a morphism of groupoids @xmath336 which induces a morphism of stacks @xmath337\longrightarrow [ v(\sigma)/g]$ ] . to prove that it is an isomorphism , we first prove that the following diagram is cartesian : @xmath338 this is easy to prove . given an element @xmath339 , under the map @xmath340
, we get @xmath341 .
if there is an element @xmath342 such that @xmath343 , then from the exact sequence in the first row of ( [ close - hypertoric3 ] ) , there is an element @xmath344 such that @xmath345 .
thus we have an element @xmath346 .
so the morphism @xmath337\longrightarrow [ v(\sigma)/g]$ ] is injective .
let @xmath347 be an element in @xmath348 , then there exists an element @xmath349 such that @xmath350 . from ( [ close - hypertoric3 ] ) , @xmath351 determines an element in @xmath196 , so @xmath352 is surjective and @xmath352 is an isomorphism .
clearly the stack @xmath353 $ ] is a closed substack of @xmath24 , so the stack @xmath317 $ ] is also a closed substack of @xmath24
. we now study open substacks of @xmath24 .
let @xmath38 be a top dimensional cone in @xmath31 .
then @xmath354 is a stacky fan , where @xmath355 is given by @xmath237 for @xmath356 .
since @xmath0 has rank @xmath1 , we find that @xmath357 is a finite abelian group .
so in this case the generic element @xmath8 induces zero in @xmath357 .
this is the degenerate case , which means that the corresponding ideal ( [ ideal1 ] ) is zero .
thus @xmath358 note that @xmath359 is a finite abelian group . according to the construction of hypertoric dm stack in section 2 ,
the hypertoric dm stack @xmath360 associated to @xmath361 is the quotient stack @xmath362 $ ] which can be regarded as a local chart of the hypertoric orbifold @xmath208 $ ] .
if @xmath38 is a top - dimensional cone in the multi - fan @xmath31 , then @xmath363 is an open substack of @xmath24 .
since @xmath38 is a top dimensional cone in @xmath31 , from ( [ exact2 ] ) we get a basis @xmath364 of @xmath170 .
let @xmath365 be the open subvariety of @xmath366 defined by the monomials @xmath367 in ( [ irrelevant ] ) .
let @xmath368 , i.e. the points in @xmath365 staisfying ( [ ideal1 ] ) .
then we have the groupoid @xmath369 associated to the action of @xmath196 on @xmath370 .
it is clear that this groupoid defines an open substack of @xmath24 .
next we show that this substack is isomorphic to @xmath363 .
consider the following commutative diagram : @xmath371 applying gale dual and @xmath254 , we obtain @xmath372 we construct a morphism @xmath373 . for @xmath374 , we set @xmath375 if @xmath376 is component of a monomial of @xmath377 in ( [ irrelevant ] ) or @xmath378 if @xmath379 is component of a monomial of @xmath377 in ( [ irrelevant ] ) , then from ( [ ideal1 ] ) , the corresponding @xmath379 or @xmath376 can be represented as linear component of @xmath380 for @xmath381 .
let @xmath382 be the morphism given by @xmath383 for @xmath381 , and @xmath384 to the corresponding @xmath385 or linear combination of @xmath380 for @xmath381 in the above analysis .
then @xmath386 induces a morphism @xmath373 .
hence we have a morphism of groupoids @xmath387\longrightarrow [ v_{\sigma}\times g{\rightrightarrows}v_{\sigma}],\ ] ] where @xmath335 is the morphism in ( [ propsecond ] ) .
this morphism determines a morphism of the associated stacks .
the isomorphism of these two stacks comes from the following cartesian diagram : @xmath388 let @xmath389 be the sublattice generated by @xmath38 , and @xmath390 . note that when @xmath38 is a top dimensional cone , @xmath391 is the local orbifold group in the local chart of the coarse moduli space of the hypertoric toric dm stack .
namely : [ finitegroup ] let @xmath38 be a top - dimensional cone in the multi - fan @xmath31
. then @xmath392 .
the proof is the same as the proof for a top dimensional cone in a simplicial fan in proposition 4.3 in @xcite .
recall that @xmath196 acts on @xmath393 via the map @xmath394 in ( [ exact5 ] ) .
we write @xmath395 [ fixpoint ] let @xmath396 be a point fixed by @xmath342 . if @xmath397 , then @xmath398 .
since @xmath196 acts on @xmath175 through the matrix @xmath399 $ ] in ( [ exact3 ] ) , we have that @xmath400 .
the lemma follows immediately .
given the multi - fan @xmath31 , we consider the pairs @xmath37 , where @xmath38 is a cone in @xmath31 , @xmath39 such that @xmath401 for @xmath41 .
note that @xmath38 is the minimal cone in @xmath31 satisfying the above condition .
let @xmath36 be the set of all such pairs @xmath37 .
[ fixbox ] there is an one - to - one correspondence between @xmath342 with nonempty fixed point set and @xmath402 .
moreover , for such @xmath351 and @xmath37 we have @xmath403\cong \mathcal{m}(\mathcal{a}(\sigma))$ ] .
let @xmath402 .
since @xmath38 is contained in a top dimensional cone @xmath77 in @xmath31 , we have @xmath404 . by lemma [ finitegroup ] , @xmath405 .
hence @xmath406 determines an element in @xmath407 .
using the morphism @xmath335 in ( [ propsecond ] ) , we see that @xmath351 fixes a point in @xmath206
. conversely , suppose @xmath342 fixes a point @xmath326 in @xmath206 , where @xmath324 .
by lemma [ fixpoint ] , the point @xmath326 satisfies the condition that if @xmath397 then @xmath398 . from the definition of @xmath366 , there is a cone in @xmath20 containing the rays for which @xmath398 . by lemma
[ lawrence ] , the rays @xmath408 for which @xmath409 is a cone in @xmath31 which we call @xmath38 .
so @xmath351 stabilizes @xmath410 in @xmath411 through @xmath258 in ( [ prop4.6 ] ) for any top dimensional cone @xmath77 containing @xmath38 , and @xmath351 corresponds to an element @xmath402 . from the definition of @xmath321 and @xmath348 in proposition [ closedsub ] , we have @xmath412 and @xmath353\cong [ y^{g}/g]$ ] which is @xmath86 .
we determine the inertia stack of a hypertoric dm stack .
[ inertia ] the inertia stack of @xmath24 is given by @xmath413 the hypertoric dm stack @xmath414 $ ] is a quotient stack .
its inertia stack is determined as @xmath415.\ ] ] by proposition [ fixbox ] , the stack @xmath403 $ ] is isomorphic to the stack @xmath86 for some @xmath42 .
let @xmath25 be an extended stacky fan , where @xmath416 , the simplicial @xmath417 fan @xmath26 is the fan of weighted projective plane @xmath418 , and @xmath419 is given by the vectors @xmath420 , where @xmath421 are the generators of the rays in @xmath26 .
choose generic element @xmath422 .
then @xmath109 is the stacky hyperplane arrangement whose induced extended stacky fan is @xmath91 . a lifting of @xmath8 in @xmath423 through the gale dual map @xmath127 is @xmath424 .
the corresponding hyperplane arrangement @xmath425 consists of 4 lines , see figure 2 .
take @xmath426 , then @xmath42 , where @xmath38 is the ray generated by @xmath427 .
consider the following diagram @xmath428 we have the quotient extended stacky fan @xmath285 , where @xmath429 is given by the vectors @xmath430 , and @xmath431 is the extra data in the quotient extended stacky fan .
taking gale dual , we get @xmath432 where @xmath127 is given by the matrix @xmath433 $ ] and @xmath314 is given by @xmath434 $ ] . the associated generic element @xmath435 and the lifting of @xmath436 in @xmath437 is @xmath438 .
so the quotient hyperplane arrangement @xmath305 is a line with three distinct points @xmath439 .
the bounded polyhedron of this hyperplane arrangement is two segments intersecting at one point , see figure 3 .
the core of @xmath86 corresponds to these two segments , hence is two @xmath211 s meeting at one point .
adding the stacky structure the twisted sector @xmath86 corresponding to the element @xmath406 is the trivial @xmath223-gerbe over the _ crepant resolution _ of the stack @xmath221 $ ] .
in this section we discuss the orbifold chow ring of hypertoric dm stacks .
we determine its module structure , then compute the orbifold cup product .
we first consider the ordinary chow ring for hypertoric dm stacks . according to @xcite ,
the cohomology ring of @xmath24 is generated by the chern classes of some line bundles defined as follows .
applying @xmath191 to ( [ exact2 ] ) , we have @xmath440 [ linebdle ] for every @xmath237 in the stacky hyperplane arrangement , define the line bundle @xmath441 over @xmath24 to be the trivial line bundle @xmath442 with the @xmath196-action on @xmath323 defined via the @xmath203-th component of the morphism @xmath443 in the above exact sequence . for any @xmath49
, there is a cone @xmath444 such that @xmath445 where @xmath446 are rational numbers .
let @xmath47 denote all the pairs @xmath48 .
then @xmath47 gives rise a group ring @xmath53=\bigoplus_{(c,\sigma)\in n^{\delta_\mathbf{\beta}}}\mathbb{q}\cdot y^{(c,\sigma)},\ ] ] where @xmath54 is a formal variable . by abuse of notation
, we write @xmath78 as @xmath79 .
the multiplication is given in terms of the ceiling function for fans which we define below .
since the multi - fan @xmath69 is simplicial , we have the following lemma . for any @xmath49
, there exists a unique cone @xmath447 and @xmath448 such that @xmath449 and @xmath450 where @xmath451 .
@xmath452 @xmath60 is called the fractional part of @xmath48 . now for @xmath55 , from lemma [ lemma4.5 ] , we write @xmath58 , where @xmath59 s are nonnegative integers .
we define the _ ceiling function _ @xmath62 by @xmath453 note that if @xmath64 , @xmath65 . for two pairs @xmath66 ,
@xmath67 , if @xmath68 is a cone in @xmath69 , define @xmath70 .
let @xmath71 be the minimal cone in @xmath69 containing @xmath72 so that @xmath73 .
the ceiling function @xmath62 is an integral linear combination of @xmath237 s for @xmath381 .
we define the grading on @xmath74 $ ] as follows . for any @xmath48 , write @xmath58 , then @xmath75 where @xmath76 is the dimension of @xmath77 .
let @xmath81 be the ideal in @xmath74 $ ] generated by the elements in ( [ ideal2 ] ) .
the multiplication @xmath454 is defined by ( [ product ] ) .
the multiplication ( [ product ] ) is associative .
for any three pairs @xmath455 , if @xmath456 is a cone in @xmath457 , let @xmath458 be the minimal cone in @xmath457 containing @xmath459 such that @xmath460 .
then we check from the properties of ceiling function that @xmath461 and @xmath462 are both equal to @xmath463 it is easy to check that the product preserves the grading , and the proof is left to the readers .
consider the map @xmath268 which is given by @xmath269 .
we take @xmath45 as the vertex set .
the _ matroid complex _
@xmath44 is defined using @xmath5 by requiring that @xmath464 iff the normal vectors @xmath465 are linearly independent in @xmath33 .
the _ stanley - reisner ring _ of the matroid @xmath44 is @xmath466=\frac{\mathbb{q}[y^{b_{1}},\cdots , y^{b_{m}}]}{i_{m_{\beta}}},\ ] ] where @xmath467 is the matroid ideal generated by the set of square - free monomials @xmath468 it is clear that @xmath469 $ ] is a subring of @xmath74 $ ] under the injection @xmath470 .
[ coarsecoh ] let @xmath109 be a stacky hyperplane arrangement and @xmath24 the corresponding hypertoric dm stack , then we have an isomorphism of graded rings @xmath471}{cir(\delta_\mathbf{\beta})},\ ] ] given by @xmath472 , where @xmath81 is the ideal generated by elements in ( [ ideal2 ] ) .
let @xmath230 be the coarse moduli space of the hypertoric dm stack @xmath24 . by @xcite , we have @xmath473}{cir(\delta_\mathbf{\beta})},\ ] ] given by @xmath474 , where @xmath475 is the @xmath476-equivariant weil divisor on @xmath230 .
let @xmath477 be the first lattice vector in the ray generated by @xmath237 , then @xmath478 for some positive integer @xmath479 .
by @xcite , the chow ring of the stack @xmath24 is isomorphic to the chow ring of its coarse moduli space @xmath480 via @xmath481 , and @xmath482 for @xmath483 .
note that we have an isomorphism : @xmath484}{cir(\delta_\mathbf{\beta})}\cong \frac{\mathbb{q}[y^{b_{1}},\cdots , y^{b_{m}},y^{b_{1}^{'}},\cdots , y^{b_{m}^ { ' } } ] } { ( cir(\delta_\mathbf{\beta}),y^{b_{1}}+y^{b_{1}^{'}},\cdots , y^{b_{m}}+y^{b_{m}^{'}})}\ ] ] for which the right side is the chow ring of the lawrence toric dm stack @xmath23 defined in definition [ lawrencetoricdmstack ] , see @xcite .
so in the dm stack situation @xmath485 .
this point of view will be covered and generalized in @xcite .
let @xmath486 denote the orbifold chow ring of @xmath24 , which by definition is @xmath487 as a group . by proposition [ inertia ]
, we have @xmath488 for @xmath402 , there is an exact sequence of vector bundles , @xmath489 where @xmath490 denotes the normal bundle of @xmath86 in @xmath24 .
on the other hand , there is a surjective morphism @xmath491 restricting this to @xmath86 yields a surjective map @xmath492 moreover , the element in the local group represented by @xmath406 acts trivially on the kernel .
let @xmath406 act on @xmath493 with eigenvalue @xmath494 , where @xmath495 .
it follows that the age function on the component @xmath86 assumes the value @xmath496 hence @xmath486 as a graded module coincides with @xmath497.\ ] ] note that @xmath486 is @xmath498-graded , due to the fact that @xmath24 is hyperkhler .
again since the multi - fan @xmath31 is simplicial , we have the following result , similar to lemma 4.6 in @xcite .
let @xmath77 be a cone in the multi - fan @xmath31 .
if @xmath499 , and suppose @xmath500 are contained in a cone @xmath447 .
then @xmath501 is contained in a cone of @xmath31 .
let @xmath24 be the hypertoric dm stack associated to the stacky hyperplane arrangement @xmath14 , then we have an isomorphism of graded @xmath502-modules : @xmath503}{cir(\delta_\mathbf{\beta})}\cong \bigoplus_{(v,\sigma)\in box(\delta_{\mathbf{\beta}})}a^{*}(\mathcal{m}(\mathcal{a}(\sigma)))[deg(y^{(v,\sigma)})].\ ] ] we use arguments similar to those in proposition 4.7 of @xcite . from lemma [ lemma4.5 ]
it is easy to see that @xmath53\cong \bigoplus_{(v,\sigma)\in box(\delta_{\mathbf{\beta}})}y^{(v,\sigma)}\cdot \mathbb{q}[m_{\beta}].\ ] ] consider @xmath504 .
it is an ideal of the ring @xmath505 $ ] , so @xmath503}{cir(\delta_\mathbf{\beta})}\cong \bigoplus_{(v,\sigma)\in box(\delta_{\mathbf{\beta}})}\frac{y^{(v,\sigma)}\cdot \mathbb{q}[m_{\beta}]}{y^{(v,\sigma)}\cdot cir(\delta_\mathbf{\beta})}.\ ] ] for an element @xmath402 , let @xmath506 .
then we have an induced _ matroid complex _
@xmath507 , where @xmath508 is the map in the quotient stacky hyperplane arrangement @xmath43 and the quotient extended stacky fan @xmath509
. similarly from @xmath508 , we have multi - fan @xmath510 in @xmath511 . by lemma [ coarsecoh ] , @xmath512/cir(\delta_{\beta(\sigma)})$ ] . for any element @xmath42
, we construct an isomorphism @xmath513}{cir(\delta_\mathbf{\beta(\sigma)})}[deg(y^{(v,\sigma ) } ) ] \longrightarrow \frac{y^{(v,\sigma)}\cdot \mathbb{q}[m_{\beta}]}{y^{(v,\sigma)}\cdot cir(\delta_\mathbf{\beta})}.\ ] ] as follows .
consider the quotient stacky hyperplane arrangement @xmath305 .
the hypertoric dm stack @xmath86 is a closed substack of @xmath24 .
consider the morphism @xmath514\to \mathbb{q}[y^{b_{1}},\ldots , y^{b_{m}}]$ ] given by @xmath515 . by lemma [ multifan
] , it is easy to check that the ideal @xmath516 is mapped to the ideal @xmath467 , so we have a morphism @xmath517\to \mathbb{q}[m_{\beta}]$ ] . since @xmath469 $ ] is a subring of @xmath74 $ ] .
let @xmath518[deg(y^{(v,\sigma)})]\to y^{(v,\sigma)}\cdot \mathbb{q}[m_{\beta}]$ ] be the morphism given by @xmath519 .
using similar arguments as in proposition 4.7 of @xcite , we find that the ideal @xmath520 goes to the ideal @xmath521 , so we have the morphism @xmath522 such that @xmath523)=[y^{(v,\sigma)}\cdot y^{b_{i}}]$ ] .
conversely , for @xmath402 , since @xmath38 is simplicial , for @xmath524 we can choose @xmath525 such that @xmath526 and @xmath527 for @xmath528 .
we consider the following morphism @xmath529\to \mathbb{q}[y^{\widetilde{b}_{1}},\ldots , y^{\widetilde{b}_{l}}]$ ] given by : @xmath530 again by lemma [ multifan ] , the ideal @xmath467 is mapped by @xmath531 to the ideal @xmath516 .
then @xmath531 induces a surjective map @xmath469\to \mathbb{q}[m_{\beta(\sigma)}]$ ] and a surjective map @xmath532\to \mathbb{q}[m_{\beta(\sigma)}][deg(y^{(v,\sigma)})]$ ] . using the same computation as in proposition 4.7 in @xcite ,
the relations @xmath521 is seen to go to @xmath520 .
this yields another morphism @xmath533}{y^{(v,\sigma)}\cdot cir(\delta_\mathbf{\beta})}\longrightarrow \frac{\mathbb{q}[m_{\beta(\sigma)}]}{cir(\delta_\mathbf{\beta(\sigma)})}[deg(y^{(v,\sigma)})]\ ] ] so that @xmath534 .
so @xmath522 is an isomorphism .
we conclude by lemma [ coarsecoh ] . in this section
we compute the orbifold cup product .
first for any @xmath535 , we have the following lemma : if @xmath536 is a cone in the multi - fan @xmath457 , there exists a unique @xmath537 such that @xmath538 is a cone in the multi - fan @xmath457 and @xmath539 has no fractional part .
let @xmath540 and @xmath541 the minimal cone in @xmath536 containing @xmath542 .
then @xmath543 satisfies the conditions of the lemma .
the notation @xmath544 means that the triple @xmath545 satisfies the conditions in lemma [ triples ] . by @xcite ,
the 3-twisted sector @xmath546 is the moduli space of 3-pointed genus 0 degree 0 orbifold stable maps to @xmath24 .
let @xmath547 be a genus 0 twisted curve with stacky structures possibly at @xmath548 .
consider a constant map @xmath549 with image @xmath550 .
this induces a morphism @xmath551 where @xmath552 is the local group of the point @xmath553 .
let @xmath554 be generators of @xmath555 and @xmath556 .
the @xmath557 fixes the point @xmath553 . by proposition [ fixbox ]
, @xmath557 corresponds to an element @xmath558 .
an argument similar to that in proposition 6.1 in @xcite shows that 3-twisted sectors of the hypertoric dm stack @xmath24 are given by @xmath559 where @xmath560 is the cone in @xmath31 satisfying @xmath561 , @xmath562 .
let @xmath563 be the evaluation map .
we have the obstruction bundle ( see @xcite ) @xmath564 over the 3-twisted sector @xmath565 , @xmath566 where @xmath567 is the embedding , @xmath568 is the @xmath569-covering branching over three marked points @xmath570 , and @xmath569 is the group generated by @xmath571 . let @xmath402 , say @xmath404 for some top dimensional cone @xmath77 .
let @xmath572 be the inverse of @xmath406 as an element in the group @xmath573 .
equivalently , if @xmath574 for @xmath41 , then @xmath575 .
[ equalto3 ] let @xmath576 such that @xmath577 . then if @xmath578 , @xmath579 , where @xmath580 , then @xmath581 or @xmath582 .
let @xmath583 , with @xmath584 and @xmath585 . then @xmath586 . from the condition we have @xmath587 and @xmath588 .
so if @xmath589 belongs to @xmath590 and @xmath541 , then @xmath591 exist and if @xmath592 , then @xmath593 . if @xmath589 belongs to @xmath590 , but not @xmath541 , then @xmath594 does nt exist and @xmath595 , @xmath596 .
the cases that @xmath589 belongs to @xmath597 but not @xmath598 , to @xmath599 but not @xmath600 are similar .
we omit them .
the stack @xmath24 is an abelian dm stack , i.e. the local groups are all abelian groups . for any 3-twisted sector @xmath565 , the normal bundle @xmath601 can be split into the direct sum of some line bundles under the group action .
it follows from the definition that if @xmath602 , then the action of @xmath406 on the normal bundle @xmath601 is given by the diagonal matrix @xmath603 .
a general result in @xcite and @xcite about the obstruction bundle and lemma [ equalto3 ] imply the following proposition .
[ obstructionbdle ] let @xmath604 be a 3-twisted sector of the stack @xmath24 such that @xmath605 .
then the euler class of the obstruction bundle @xmath564 on @xmath565 is @xmath606 where @xmath441 is the line bundle over @xmath24 defined in definition [ linebdle ] .
@xmath452 to prove the main theorem , we introduce two lemmas .
for any two pairs @xmath607 , there exist two unique elements @xmath608 such that @xmath609 and @xmath610 , @xmath611 , where @xmath612 are nonnegative integers .
let @xmath613 be the unique element in @xmath36 such that @xmath577 in the local group given by @xmath68 .
let @xmath583 , with @xmath584 and @xmath585 .
the existence of @xmath591 means that @xmath589 belongs to @xmath614 .
let @xmath560 be the cone in @xmath31 such that @xmath615 , with @xmath616 or @xmath267 .
let @xmath617 be the set of @xmath203 such that @xmath616 and @xmath618 exist , @xmath619 the set of @xmath240 such that @xmath589 belongs to @xmath560 but not to @xmath541 .
we have the following lemma for the ceiling functions : @xmath620 . by the definition of ceiling functions
, we have @xmath621 and @xmath622 .
the lemma follows .
if @xmath68 is a cone in @xmath69 for the two pairs @xmath623 , then the product @xmath624 in ( [ product ] ) can be given by @xmath625 first for a fixed ray @xmath626 and @xmath627 , by the definition of ceiling functions , we find that @xmath628 since @xmath629 , by lemma [ gauss ] , we need to check that @xmath630 this can be easily proven using ( [ check ] ) and lemma [ equalto3 ] . by proposition [ moduleisom ]
, it remains to prove that the orbifold cup product is the same as the product in the ring @xmath74 $ ] .
by lemma [ keylem ] , we need to prove that the orbifold cup product is the same as the product in ( [ productmain ] ) .
it suffices to consider the canonical generators @xmath79 , @xmath631 for @xmath42 .
consider @xmath632 with @xmath42 .
the element @xmath37 determines a twisted sector @xmath86 . the corresponding twisted sector to @xmath237 is the whole hypertoric stack @xmath24 .
it is easy to see that the 3-twisted sector relevant to this product is @xmath633 , where @xmath634 denotes the inverse of @xmath406 in the local group .
it follows from the dimension formula in @xcite that the obstruction bundle over @xmath635 has rank zero .
it is immediate from definition that @xmath636 if there is a cone in @xmath457 containing @xmath637 .
this is the third case in ( [ productmain ] ) .
now consider @xmath638 , where @xmath639 . by ( [ 3-sector ] )
, we see that if @xmath536 is not a cone in @xmath31 , then there is no 3-twisted sector corresponding to the elements @xmath640 .
thus the product is zero by definition . on the other hand , by definition of the ring @xmath74 $ ] , we have @xmath641 .
so @xmath642 . if @xmath536 is a cone in @xmath31 , let @xmath643 such that @xmath644 and @xmath645 in the local group .
then we have the 3-twisted sector @xmath565 .
let @xmath563 be the evaluation maps .
the element @xmath631 is the class @xmath385 in the cohomology of the twisted sector @xmath86 . from the definition of orbifold cup product @xcite , @xcite
, we have : @xmath646 where @xmath647 is the composite of @xmath648 and the natural involution @xmath649 .
let @xmath583 , with @xmath584 and @xmath585 .
let @xmath617 denote the set of @xmath203 such that @xmath650 and @xmath591 exist , @xmath619 the set of @xmath240 such that @xmath408 belongs to @xmath560 , but not belong to @xmath651 .
if some @xmath652 , for example , @xmath653 , then @xmath654 is a torsion element in @xmath0 which means that the action of @xmath654 is trivial on the hypertoric dm stack . then the 3-twisted sector corresponding to @xmath640 is isomorphic to the twisted sector @xmath655 and the obstruction bundle over @xmath655 is zero by @xcite . in this case
the set @xmath617 and @xmath619 are all empty .
so @xmath656 .
this is again the third case in ( [ productmain ] ) .
now we assume that @xmath657 .
if @xmath658 , then @xmath659 , @xmath660 and @xmath661 .
so the 3-twisted sector corresponding to @xmath640 is isomorphic to the twisted sector @xmath662 and the obstruction bundle over @xmath662 is zero by @xcite .
the set @xmath619 is the set @xmath240 such that @xmath663 .
so we have @xmath664 which is the second case in ( [ productmain ] ) . if @xmath665 , then @xmath666 and the obstruction bundle over the 3-twisted sector @xmath565 is given by proposition [ obstructionbdle ] .
so we have : @xmath667 since @xmath668 , we have @xmath669 which is the first case in ( [ productmain ] ) . @xmath452
in this section we compute some examples of the orbifold chow rings of hypertoric dm stacks .
in particular , we relate the hypertoric stack to crepant resolutions .
let @xmath99 and @xmath26 the fan of projective line @xmath211 generated by @xmath670 .
let @xmath671 be the map given by @xmath672 and @xmath673 for @xmath674 .
consider the following exact sequences @xmath675 @xmath676 where the gale dual @xmath127 is given by the column vectors of the matrix @xmath677.\ ] ] note that @xmath678 is unimodular in the sense of @xcite .
taking @xmath254 yields @xmath679 so @xmath680 .
choose @xmath681 in @xmath682 , then it is a generic element .
the extended stacky fan @xmath25 is induced from the stacky hyperplane arrangement @xmath109 , where @xmath28 is the hyperplane arrangement whose normal fan is @xmath26 .
it is easy to see that the toric dm stack is the projective line @xmath211 .
the hypertoric dm stack is the crepant resolution of the gorenstein orbifold @xmath683 $ ] . to see this , from the construction of hypertoric dm stack
, we have : @xmath684 where @xmath198 is given by the matrix @xmath685 $ ] .
let @xmath686 $ ] be the coordinate ring of @xmath687 .
so the ideal @xmath205 in ( [ ideal1 ] ) is generated by the following equations : @xmath688 hence @xmath206 is the subvariety of @xmath689 determined by the above ideal .
the action of @xmath196 on @xmath206 is through the map @xmath198 in ( [ resolution ] ) .
the hypertoric dm stack associated to @xmath14 is @xmath414 $ ] . from proposition
[ orientation ] , the hypertoric dm stack is independent to the coorientations of the hyperplanes .
this means that we can give the stacky hyperplane arrangement @xmath14 as follows .
let @xmath690 for @xmath691 .
then the gale dual map @xmath692 is given by the matrix @xmath693,\ ] ] which is exactly the matrix in lemma 10.2 in @xcite , from which it follows that the coarse moduli space @xmath230 of @xmath414 $ ] is the crepant resolution of the gorenstein orbifold @xmath683 $ ] .
the core of the hypertoric dm stack @xmath24 is a chain of @xmath694 copies of @xmath211 with normal crossing divisors corresponding to the multi - fan @xmath31 .
the @xmath695-action defining the gorenstein orbifold @xmath683 $ ] is given by @xmath696 for @xmath697 .
there are @xmath694 twisted sectors each of which is isomorphic to @xmath698 with age @xmath385 .
there are only dimension zero cohomology on the untwisted sector and twisted sectors .
so we prove the following proposition : [ crepant ] the chow ring of @xmath24 is isomorphic to the ring @xmath701}{\{y_{i}y_{j } : 1\leq i , j\leq n-1\}},\ ] ] which is isomorphic to the orbifold cohomology ring of the gorenstein orbifold @xmath683 $ ] . by theorem 1.1 , the chow ring of @xmath24 is isomorphic to the ring : @xmath702}{\{y_{1}-y_{n}+y_{3}+\cdots+y_{n-1},y_{i}y_{j } : 1\leq i , j\leq n-1\}}\ ] ] which we can easily check that this ring is isomorphic to the orbifold cohomology ring of @xmath683 $ ] in proposition [ orbifold ] .
y. ruan @xcite conjectured that , among other things , the orbifold cohomology ring of a hyperkhler orbifold is isomorphic to the ordinary cohomology ring of a hyperkhler resolution ( which is crepant ) . for the orbifold @xmath683 $ ] , the crepant resolution @xmath230 is smooth , we have that @xmath703 . from proposition [ crepant ] ,
the conjecture is true .
a conjecture equating gromov - witten theories of an orbifold and its crepant resolutions , as proposed in @xcite , is recently proven in genus 0 for @xmath704 $ ] , see @xcite .
the comparison of two gromov - witten theories requires certain change of variables . for @xmath704 $ ] case ,
see @xcite . for @xmath705 $ ] case the following change of variables is found in @xcite : @xmath706 under this change of variables , the genus zero gromov - witten potential of the crepant resolution is seen to coincide with the genus zero orbifold gromov - witten potential of the orbifold @xmath707 $ ] , see @xcite . for a toric orbifold
, it is known that adding rays in the simplicial fan can give a crepant resolution . in the end of the paper
we compute an example and explain that adding rays in the stacky hyperplane arrangement does nt give a smooth hypertoric variety which means that it is not easy in general to give a crepant resolution in hyperkhler geometry .
let @xmath25 be an extended stacky fan , where @xmath416 , the simplicial fan @xmath26 is the fan of weighted projective plane @xmath708 , and @xmath709 is given by the vectors @xmath710 , where @xmath421 are the generators of the rays in @xmath26 .
the generic element @xmath711 determines the fan @xmath26 .
the stacky hyperplane arrangement @xmath109 induces @xmath91 .
the hypertoric dm stack is @xmath712 . from theorem 1.1 , @xmath713}{(x_{1}-x_{3},x_{2}-2x_{3},x_{4}^{2 } ,
x_{1}x_{2}x_{3},x_{4}x_{2},x_{4}x_{1}x_{3 } ) } \cong \frac{\mathbb{q}[x_{3},x_{4}]}{(x_{4}^{2},x_{3}^{3},x_{3}x_{4})}.\ ] ] let @xmath714 and consider the new map @xmath715 which is given by the vectors @xmath716 . choose generic element @xmath717 and we get a new stacky hyperplane arrangement @xmath718 which induces the extended stacky fan @xmath719 .
the hypertoric dm stack @xmath720 is the stack corresponding to @xmath721 . from the definition of box
, @xmath722 is again a box element which determines a twisted sector .
we compute that @xmath723 is isomorphic to @xmath724}{(x_{1}-x_{3},x_{2}-2x_{3}-x_{4},x_{2}x_{4},x_{1}x_{2}x_{3},x_{1}x_{3}x_{4},v^{2},vx_2,vx_4 ) } \cong \frac{\mathbb{q}[x_{3},x_{4},v]}{(x_{3}x_{4}+x_{4}^{2},x_{3}^{3},x_{3}^{2}x_{4},v^{2},vx_3,vx_4)}.\ ] ] we check that @xmath725 is not isomorphic to the ring @xmath723 .
we give two comments here .
first , if the crepant resolution conjecture is true , then @xmath720 is not a hyperkhler resolution . on the other hand , the map @xmath5 is given by the matrix @xmath726 $ ] and the map @xmath727 given by @xmath728 $ ] it is easy to see that @xmath729 is not unimodular which means that @xmath720 is not smooth , see @xcite . so adding columns in @xmath730
ca nt make it unimodular which means that adding rays in the stacky hyperplane arrangement can not give a hyperkhler resolution .
second , since @xmath720 is still a hypertoric orbifold , we wish that @xmath720 is a partial resolution and there is a morphism @xmath731 .
but this is impossible because the two rings have different dimensions and it would violate the mckay correspondence statement if there exists a morphism @xmath731 , see @xcite,@xcite . holy12 d. abramovich , t. graber and a. vistoli , algebraic orbifold quantum product , in _ orbifolds in mathematics and physics ( madison , wi,2001 ) _ , 1 - 24 , _ contem .
_ 310 , amer .
soc . , 2002 .
a. acampo - neuen and j. hausen , toric prevarieties and subtorus actions , _ geometriae dedicata _ 87 , 35 - 64 ( 2001 ) .
r. bielawski and a. dancer , the geometry and topology of toric hyperkhler manifolds , _ comm .
geom_. 8 ( 2000 ) , 727760 .
l. borisov , l. chen and g. smith , the orbifold chow ring of toric deligne - mumford stacks , _
soc_. 18 ( 2005 ) , no.1 , 193 - 215 , math.ag/0309229 .
j. bryan and t. graber , the crepant resolution conjecture , math.ag/0610129 .
j. bryan , t. graber and r. pandharipande , the orbifold quantum cohomology of @xmath221 $ ] and hurwitz - hodge integrals , math.ag/0510335 .
j. bryan and y. jiang , the genus zero crepant resolution conjecture for the orbifold @xmath707 $ ] , preprint . b. chen and s. hu , a derham model for chen - ruan cohomology ring of abelian orbifolds , math.sg/0408265 .
w. chen and y. ruan , a new cohomology theory for orbifolds , _ comm .
_ 248 ( 2004 ) , no .
1 , 131 , math.ag/0004129 . w. chen and y. ruan , orbifold gromov - witten theory , in _ orbifolds in mathematics and physics ( madison , wi , 2001 ) _ , 2585 , _ contem . math .
_ 310 , amer
soc . , 2002 .
math.ag/0103156 .
d. cox , the homogeneous coordinate ring of a toric variety , _ j. of algebraic geometry _ , 4 ( 1995 ) , 17 - 50 .
d. cox , recent developments in toric geometry , in _ algebraic geometry ( santa cruz , 1995 ) _ , 389436 ,
pure math .
62.2 , amer .
soc . , 1997 .
w. fulton , _ introduction to toric varieties _ ,
annals of mathematics studies 131 , princeton university press , princeton , nj , 1993 . r. goldin and m. harada , orbifold cohomology of hypertoric varieties , math.dg/0607421 .
a. hattori and m. masuda , theory of multi - fans , _
osaka j. math_. 40 ( 2003 ) , 1 - 68 .
t. hausel and b. sturmfels , toric hyperkhler varieties , _ documenta mathematica _ , 7 ( 2002 ) , 495 - 534 . t. jarvis , r. kaufmann and t. kimura , stringy k - theory and the chern character , math.ag/0502280 .
y. jiang , the orbifold cohomology ring of simplicial toric stack bundles , _ illinois journal of mathematics _ , to appear , math.ag/0504563 .
y. jiang and h.- h. tseng , note on orbifold chow ring of semi - projective toric dm stacks , math.ag/0606322 .
h. konno , cohomology rings of toric hyperkahler manifolds , _ intern .
j. of math_.(11 ) ( 1997 ) , no.8 , 1001 - 1026 .
kronheimer , the construction of ale spaces as hyper - khler quotients , _ j. differential geom .
_ 29 ( 1989 ) , no .
3 , 665683 .
s. keel and s. mori , quotients by groupoids , _ ann . of math_.
( 2 ) 145 ( 1997 ) , no.1 , 193 - 213 .
g. laumon and l. moret - bailly , _ champs algbriques .
( french ) [ algebraic stacks ] _ , ergebnisse der mathematik und ihrer grenzgebiete .
3 . folge .
springer - verlag , berlin , 2000 .
y. ruan , cohomology ring of crepant resolutions of orbifolds , in _ gromov - witten theory of spin curves and orbifolds _ , 117126 , _ contem .
_ 403 , amer .
soc . , 2006 ,
r. p. stanley , _ combinatorics and commutative algebra _ , 2nd ed . , birkhuser boston , 1996 .
a. vistoli , intersection theory on algebraic stacks and on their moduli spaces , _ invent . math .
_ , 97 ( 1989 ) 613 - 670 .
t. yasuda , motivic integration over deligne - mumford stacks , math.ag/0312115 .
t. yasuda , twisted jets , motivic measures and orbifold cohomology , _ compos . math .
_ 140 ( 2004 ) , no .
2 , 396422 . | hypertoric varieties are determined by hyperplane arrangements . in this paper
, we use stacky hyperplane arrangements to define the notion of hypertoric deligne - mumford stacks .
their orbifold chow rings are computed . as an application ,
some examples related to crepant resolutions are discussed . |
integrable spin chains often provide a fertile ground for studying key theoretical concepts in a simple framework that captures the essential features of the problems under consideration .
an important example of this assertion is the analysis of the entanglement of a quantum system , which can be considered as one of the fundamental characteristics of the quantum realm @xcite .
one of the most common ways of measuring the degree of entanglement of a state of a quantum system @xmath2 is via the bipartite entropy of a subsystem @xmath3 @xcite .
this entropy is defined by @xmath4 , where @xmath5 is the reduced density matrix of the subsystem @xmath3 , @xmath6 is the density matrix representing the state of the whole system , and @xmath7 is an appropriate entropy functional ( von neumann , rnyi , etc . ) .
the small class of models for which the entanglement entropy can be evaluated in closed form ( at least in the thermodynamic limit ) includes certain integrable spin chains , like the lipkin meshkov
glick model @xcite , its @xmath8 generalization @xcite and the nearest - neighbors heisenberg @xmath9 and @xmath10 models @xcite .
as is well known , the latter two models are critical ( gapless ) for a certain range of values of the applied magnetic field , their corresponding virasoro algebras having central charge respectively equal to @xmath11 and @xmath12 . in both cases , the bipartite rnyi entropy of a block of @xmath13 consecutive spins when the whole chain is in its ground state scales as @xmath14 in the critical phase , where @xmath15 is the rnyi parameter ( @xmath16 for the von neumann entropy ) and @xmath17 is the central charge .
this behavior is consistent with the scaling of the rnyi entanglement entropy of a ( @xmath18)-dimensional conformal field theory ( cft ) @xcite .
in fact , the logarithmic scaling of the ground state entanglement entropy is a characteristic feature of critical ( fermionic ) one - dimensional lattice models with short - range interactions ( see , e.g. , ref .
@xcite ) . in a previous paper @xcite
, we showed that the above results also apply to a large class of supersymmetric spin chains with general ( not necessarily short - range ) interactions , which turn out to be equivalent to a suitable free fermion model .
the critical character of these chains ( for appropriate values of the chemical potential @xmath19 ) was ascertained via the analysis of the low - temperature behavior of the free energy per spin .
indeed , we proved that when the dispersion relation @xmath20 of the corresponding free fermion model is monotonic in the interval @xmath21 $ ] , for @xmath22 the free energy per spin is approximately given ( in natural units @xmath23 ) by @xmath24 where @xmath25 is the fermi velocity ( or effective speed of `` sound '' ) and @xmath26 .
this is precisely the expected behavior of the free energy for any critical model ( @xmath17 being the central charge of its virasoro algebra ) , since at low temperatures the free energy of a quantum system is determined by its lowest energy levels , and the free energy per spin of a ( @xmath18)-dimensional cft with central charge @xmath17 satisfies for sufficiently small @xmath27 @xcite .
we also studied the ground - state rnyi entanglement entropy of the above mentioned supersymmetric spin chains , showing that in the thermodynamic limit @xmath28 it again behaves as that of a @xmath1-dimensional cft with central charge @xmath26 .
the aim of this paper is to extend the results of ref .
@xcite by suppressing the requirement that the dispersion relation be monotonic in @xmath21 $ ] . as shown in section [ sec.disp ] ,
this makes it possible to treat a host of naturally arising models , like supersymmetric spin chains with near and next - to - near interactions , or with long - range rational interactions , whose dispersion relation is not always monotonic .
in fact , the entanglement entropy of fairly arbitrary energy eigenstates of one - dimensional free fermionic systems ( in particular , of the ground state of such systems with a non - monotonic dispersion relation ) has been previously studied in the literature ; see , e.g. , refs .
@xcite . in general , the entanglement entropy of the ground state of these models grows logarithmically with the size @xmath13 of the subsystem , with a constant prefactor determined by the number of boundary points of the fermi `` surface '' in @xmath29 .
this logarithmic scaling is a manifestation of the so - called `` area law '' , which is believed to hold for critical fermionic systems in an arbitrary number of dimensions @xcite .
we shall show that the @xmath0 supersymmetric chains studied in this paper do indeed satisfy the area law .
more precisely , by analyzing the low - temperature behavior of the free energy we shall first show that the models under consideration are critical for @xmath30 , where @xmath31 and @xmath32 respectively denote the minimum and maximum values of the dispersion relation .
( as explained in section [ sec.crit ] , strictly speaking this is only true if the roots of the equation @xmath33 are all simple . ) from the latter analysis it also follows that the central charge of these models is equal to the number @xmath34 of disjoint intervals that make up the fermi sea .
we shall next study the ground state rnyi entanglement entropy , showing that in the thermodynamic limit @xmath28 it behaves as @xmath35 .
we shall explicitly compute the ( non - universal ) constant @xmath36 , and prove that the prefactor @xmath37 is equal to @xmath38 .
this is in agreement with the value of the central charge deduced from the low - temperature analysis of the free energy , and once again confirms the conjecture that the entanglement properties of critical fermion models are entirely determined by the topology of their fermi surface @xcite .
we shall end this section with a few words on the paper s organization .
in section [ sec.model ] we recall the definition of the supersymmetric chains under consideration and review their main properties .
section [ sec.disp ] is devoted to the analysis of the models dispersion relation and the construction of simple examples of supersymmetric chains , featuring both short- and long - range interactions , with a non - monotonic dispersion relation . in section [ sec.crit ]
we derive the asymptotic behavior of the models free energy per spin at low temperature , showing that they are critical in an appropriate range of the chemical potential , and determine the central charge of the corresponding virasoro algebra .
the asymptotic behavior of the entanglement entropy of the models ground state is determined in section [ sec.gsee ] using a particular case of the fisher hartwig conjecture for toeplitz matrices @xcite rigorously proved by bttcher and silbermann @xcite .
we briefly state our conclusions and outline several future developments suggested by the present work in section [ sec.conc ] .
the paper ends with three appendices in which we present a review of the application of the fisher hartwig conjecture in the present context , as well as the proofs of several technical results used throughout section [ sec.gsee ] .
the type of models we shall study in this work is the class of @xmath0 supersymmetric spin chains with translationally invariant interactions introduced in ref . @xcite . in the latter models
each site is occupied either by a scalar boson or a spinless fermion , whose creation operators we shall respectively denote by @xmath39 and @xmath40 , the subindex @xmath41 indicating the site on which these operators act .
thus the hilbert space is the @xmath42-dimensional subspace @xmath43 of the infinite - dimensional fock space defined by the constraints @xmath44 the hamiltonian of the models under consideration is given by to @xmath45 unless otherwise stated . ]
@xmath46 where the operator @xmath47 is the total fermion number @xmath48 so that the real parameter @xmath19 has the natural interpretation of the fermions chemical potential .
the real - valued function @xmath49 giving the strength of the interaction between two particles @xmath50 sites apart is assumed to satisfy the constraint @xmath51 but is otherwise arbitrary was implicitly assumed to be nonnegative for all @xmath52 in ref .
@xcite . here
we shall drop the latter requirement , which is not essential in what follows . ] . in other words ,
the chain is _ closed _ , i.e , translationally invariant .
finally , @xmath53 is the @xmath0 spin permutation operator , defined by @xcite @xmath54 equivalently , let @xmath55 ( with @xmath56 ) be a state of the canonical spin basis , where @xmath57 and @xmath58 respectively denote the state with one boson or one fermion .
the action of @xmath53 on the latter state is then given by @xmath59 where @xmath60 if @xmath61 while for @xmath62 @xmath63 equals the number of fermions in the state @xmath64 occupying the sites @xmath65 .
the operator @xmath53 is clearly invariant under the supersymmetry transformation @xmath66 ( @xmath67 ) , and on @xmath43 we have @xmath68 , where @xmath69 is the total boson number .
hence the hamiltonian is indeed supersymmetric - invariant , up to a constant term and the usual relabeling @xmath70 .
the fundamental feature of the @xmath0 supersymmetric chain , explained in detail in refs .
@xcite , is that it can be mapped into a free - fermion model by interpreting the boson state @xmath57 as the fermion vacuum .
more precisely , consider the operators @xmath71 which can be regarded as a new set of fermion creation operators as they obviously satisfy the canonical anticommutation relations ( car ) on @xmath43 .
it was shown by haldane @xcite that on @xmath43 the @xmath0 permutation operator @xmath53 can be simply expressed as @xmath72 substituting into eq .
we readily obtain @xmath73 where @xmath74 we thus see that the spin chain is indeed equivalent to a free - fermion model with hopping amplitude @xmath75 and chemical potential @xmath76 . since the hamiltonian is translationally invariant on account of eq .
, it can be diagonalized by the discrete fourier transform @xmath77 indeed , the operators @xmath78 obviously satisfy the car , and can therefore be considered as a new set of fermionic operators .
moreover , a straightforward calculation shows that @xcite @xmath79 where @xmath80h_n(j)\,.\ ] ] likewise , the system s total momentum operator @xmath81 is given by @xcite @xmath82 with @xmath83 thus the operator @xmath84 creates a ( non - localized ) fermion with well - defined energy @xmath85 and momentum @xmath86 .
it follows from eq . that the spectrum of @xmath87 is the set of numbers of the form @xmath88 with @xmath89 , whose corresponding eigenstates are given by @xmath90
an essential requirement making it possible to study the chain or , equivalently , its fermionic counterpart in the thermodynamic limit is the existence of a smooth function @xmath20 _ independent of @xmath45 _ such that when @xmath91 we have @xmath92 when this is the case , we shall refer to @xmath20 as the model s _
dispersion relation_. from the latter equation and the identity @xmath93 it follows that the dispersion relation is always symmetric about @xmath94 , namely @xmath95 likewise , @xmath96 implies that @xmath97 .
it is also customary to extend @xmath20 to the whole real line as a @xmath98-periodic function , in which case eq .
entails that @xmath99 .
for instance , for the @xmath0 haldane shastry chain @xcite , whose interaction strength is given by @xmath100 it was shown in ref .
@xcite that @xmath101 hence in this case holds with @xmath102 and no error term .
in fact , it can be shown that eq
. also holds ( again with no error term ) for a suitable dispersion relation @xmath103 in the more general chain with elliptic interactions studied in ref .
@xcite .
we shall next present a few relevant examples of models of the form for which the dispersion relation is guaranteed to exist . to this end , it is convenient to rewrite eq . to take into account conditions , namely @xmath104h_n(j)\\ + 2[1-\pi(n)]\,\pi(l)\,h_n(n/2)\,,\end{gathered}\ ] ] where @xmath105 denotes
the parity of the integer @xmath50 and @xmath106 is the integer part of @xmath107 . clearly , the values of @xmath108 with @xmath109 appearing in the latter equation are no longer restricted by eq . .
for this reason , from now on we shall implicitly restrict the domain of @xmath110 to the range @xmath109 , since for @xmath111 we simply have @xmath112 . in this vein , we shall say ( with a slight abuse of language ) that the interaction is _ independent of _
@xmath45 if there is a fixed function @xmath113 such that @xmath114 for @xmath109 . if this is the case we shall simply write @xmath115 , again implicitly assuming that we are restricting ourselves to the range @xmath109 . an important class of models of the form for which the dispersion relation @xmath20 is guaranteed to exist are those whose interaction strength @xmath110 is short - ranged and independent of @xmath45 . by this
we mean that there is a positive integer @xmath116 ( the range of the interaction ) such that @xmath117 for @xmath118 , and @xmath119 with @xmath120 independent of @xmath45 and @xmath121 .
obviously , in this case we have @xmath122\,.\ ] ] in fact , the same is true if we drop but assume instead that the limit @xmath123 exists for all @xmath124 . on the other hand ,
the short range of the interaction @xmath110 is by no means a necessary condition for the existence of the dispersion relation @xmath20 . indeed , suppose for simplicity that @xmath115 is independent of @xmath45 , and that the series @xmath125 is absolutely convergent
. then clearly holds with @xmath126\,.\ ] ] for instance , for the power law interaction @xmath127 with @xmath128 the previous series can be summed in closed form in terms of the polylogarithm function @xcite @xmath129 namely ( taking , for simplicity , @xmath130 ) @xmath131 , \label{cepli}\end{gathered}\ ] ] where @xmath132 denotes riemann s zeta function ( cf .
[ fig.epp ] ) . chain with power - law interaction @xmath133 for several values of the exponent @xmath134 between @xmath135 and @xmath136 . ] from the integral representation @xmath137 where @xmath138 is euler s gamma function , we obtain the equivalent expression @xmath139 using the latter formula in the identity @xmath140 and reversing the order of integration we arrive at the somewhat simpler expression @xmath141 remarkably , for @xmath142 eq .
reduces to eq .
( see , e.g. , @xcite ) .
thus the @xmath0 chain with rational interaction @xmath143 has the same dispersion relation as the haldane shastry chain .
this is of course not entirely unexpected , since for _ fixed _
@xmath144 we have @xmath145 .
note , however , that for @xmath146 both interactions , although negligibly small as @xmath91 , differ by a factor@xmath147 ( cf . fig .
[ fig.hs-rat ] ) .
hs chain ( red ) with the simple inverse - square law @xmath148 ( blue ) for @xmath149 .
inset : same plot for the range @xmath150 . ] of course , although holds for a wide range of interesting interactions , it is not universally true .
for instance , it is not satisfied by the @xmath45-independent interaction @xmath151 , since @xmath152 converges for @xmath153 while the series @xmath154 is divergent . in a previous paper
@xcite we analyzed the critical behavior of supersymmetric spin chains of the type whose dispersion relation is monotonic in the interval @xmath21 $ ] .
these models include the @xmath0 haldane shastry chain ( cf . ) and , more generally , its elliptic generalization introduced in ref . @xcite . as is apparent from fig .
[ fig.epp ] ( and can be analytically checked differentiating eq . ) , the chain with power - law interactions also exhibits this property .
however , this behavior is not universal , and there are in fact simple examples of supersymmetric chains of the form with a non - monotonic dispersion relation . indeed , consider to begin with the chain with nearest and next - to - nearest interactions , whose hamiltonian ( up to an irrelevant multiplicative constant ) is given by @xmath155 with @xmath156 , @xmath157 and @xmath158 .
note that when @xmath159 the fermionic version of the latter model can be mapped to the ( closed ) heisenberg @xmath9 chain by a wigner
jordan transformation @xcite . from eq . with @xmath160 and @xmath161
we easily obtain @xmath162 since @xmath163 the dispersion relation will have a critical point in @xmath164 if and only if @xmath165 ( more precisely , a maximum for @xmath166 and a minimum for @xmath167 ) .
thus in this case the dispersion relation is not monotonic in @xmath21 $ ] provided that @xmath165 .
the same is clearly true for chains of the form with interactions of finite range @xmath168 , for suitable values of the interaction strengths .
it is also easy to construct simple examples of chains of the form with long - range interactions with a non - monotonic dispersion relation .
take , for instance , @xmath169 whose dispersion relation is given by @xmath170\,.\ ] ] if @xmath171 , differentiating the latter equation we obtain @xmath172\,,\ ] ] with @xmath173 it can be shown ( cf .
[ fig.ep23 ] ) that the function @xmath174 increases monotonically over the interval @xmath164 , with @xmath175 and @xmath176 . ] , so that @xmath177 changes sign once ( from positive to negative ) in @xmath164 if and only if @xmath178 .
we conclude that the dispersion relation of the chain with interaction is not monotonic on @xmath21 $ ] provided that @xmath179 . in particular ,
for @xmath180 the dispersion relation is not monotonic in @xmath21 $ ] even if the interaction strength is positive for @xmath181 ( see fig .
[ fig.ep23 ] for a plot of @xmath20 when @xmath182 ) . .
inset : plot of the function @xmath174 in eq . .
in this section we shall study the critical properties of the spin chain when its dispersion relation @xmath20 is not necessarily monotonic over the interval @xmath21 $ ] . to this end
, we shall examine the low temperature behavior of the helmholtz free energy per spin @xmath183 which for this model is given by ( cf .
@xcite ) @xmath184\,{\mathrm{d}}p\,.\ ] ] in the previous expressions @xmath185 denotes the partition function of the chain with @xmath45 spins , and @xmath186 ( in natural units @xmath23 ) .
as remarked in the introduction , at low temperatures the free energy of a critical model should satisfy eq . .
moreover , it was shown in ref .
@xcite that when @xmath20 is monotonic and nonnegative in the interval @xmath21 $ ] the model is critical when the chemical potential @xmath19 lies in the interval @xmath187 , with central charge @xmath26 , and noncritical for @xmath19 outside the closed interval @xmath188 $ ] .
we shall next extend this result to the more general case in which @xmath20 is not necessarily monotonic ( nor nonnegative ) in @xmath21 $ ] . to begin with
, it is immediate to show that the model is not critical when @xmath19 lies outside the interval @xmath189 $ ] . indeed , suppose first that @xmath190 , so that @xmath191 , @xmath192 and @xmath193 in contradiction with the asymptotic behavior characteristic of a critical model .
similarly , when @xmath194 we have @xmath195{\mathrm{d}}p\ ] ] and @xmath196\,{\mathrm{d}}p\\&<t{\mathrm e}^{-{\beta}(\mu-{{{\mathcal e}}_{\text{max}}})}\,,\end{aligned}\ ] ] again in disagreement with eq . .
this conclusion is also borne out by the fact that when @xmath194 or @xmath190 the spectrum is clearly gapped , with energy gap respectively equal to @xmath197 or @xmath198 .
let us now consider the more interesting case in which @xmath199 , in which the spectrum is clearly gapless .
we shall suppose that the equation @xmath33 has @xmath200 roots @xmath201 in the interval @xmath164 , which we will assume to be _
we start by expressing the free energy as @xmath202 where @xmath203{\mathrm{d}}p\ ] ] and the last integral is extended to the subset of the interval @xmath21 $ ] defined by the inequality @xmath204 . clearly , as @xmath205 the main contribution to the integral in eq
. comes from an increasingly small neighborhood of the `` turning points '' @xmath206 , near which @xmath207 is small . to exploit this fact , we choose @xmath208 small enough that @xmath209\cap[p_j-{\delta}p , p_j+{\delta}p]=\emptyset$ ] for @xmath210 and @xmath211 on @xmath212 $ ]
this is certainly possible , since by hypothesis @xmath213 for all @xmath214 .
obviously , @xmath215 depends only on the dispersion relation @xmath20 , and is therefore independent of @xmath27 .
calling @xmath216-\cup_{i=0}^m[p_i-{\delta}p , p_i+{\delta}p]$ ] we have @xmath217{\mathrm{d}}p=\int_a\log\bigl[1+{\mathrm e}^{-{\beta}|{{\mathcal e}}(p)-\mu|}\bigr]{\mathrm{d}}p\\ + \sum_{i=0}^m\int_{p_i-{\delta}p}^{p_i+{\delta}p}\log\bigl[1+{\mathrm e}^{-{\beta}|{{\mathcal e}}(p)-\mu|}\bigr]{\mathrm{d}}p\,.\end{gathered}\ ] ] the first integral can be easily estimated .
indeed , let @xmath218 be the minimum value of @xmath207 on the compact set @xmath219 , which is clearly positive since @xmath220 on @xmath219 , and denote by @xmath221 the length of @xmath3 .
we then have @xmath222{\mathrm{d}}p{\leqslant}{\mathrm e}^{-a{\beta}}|a|{\leqslant}\pi { \mathrm e}^{-a{\beta}}\,,\ ] ] with @xmath218 ( and @xmath221 ) obviously independent of @xmath27 .
consider next the integral @xmath223{\mathrm{d}}p\,.\ ] ] to analyze its low temperature behavior , we perform the change of variable @xmath224 separately in each of the intervals @xmath225 $ ] and @xmath226 $ ] . since @xmath213 ,
this change of variable is one to one and @xmath227 in both of the latter intervals , and we have @xmath228 the asymptotic behavior of these integrals as @xmath205 can be easily determined taking into account that by construction @xmath177 does not vanish on both intervals @xmath225 $ ] and @xmath226 $ ] , and therefore @xmath229 as @xmath230 implies that @xmath231 . since the integral @xmath232 is convergent we have @xmath233 where we have set @xmath234 moreover , if @xmath235 is independent of @xmath236 we have @xmath237 using this inequality and the integral @xmath238 in eq .
we thus have @xmath239 from eqs .
we finally obtain the asymptotic estimate @xmath240 this is the low temperature behavior of the free energy of a @xmath1-dimensional cft with @xmath34 free bosons with fermi velocities @xmath241 .
thus in this case the model is critical , with central charge @xmath242 .
the situation is markedly different if any of the roots of the equation @xmath33 is not simple .
indeed , assume that @xmath243 is a root of order @xmath244 of the latter equation , so that we can write @xmath245\ ] ] with @xmath246 we now choose @xmath208 such that @xmath209\cap[p_j-{\delta}p , p_j+{\delta}p]=\emptyset$ ] for @xmath210 and @xmath211 on @xmath247 $ ] for all @xmath214 . proceeding as before we again arrive at eqs . - and obtain the estimate for the first integral in eq . .
in order to analyze the low temperature behavior of the integral @xmath248 , we again perform the change of variable in each of the intervals @xmath249 $ ] and @xmath250 $ ] , thus obtaining eq . with @xmath251 . in each of the latter intervals we now have @xmath252\ ] ] and @xmath253\,,\ ] ] so that @xmath254\,.\ ] ] substituting into eq . and
proceeding as before we easily obtain @xmath255 ( see ref .
@xcite for more details on the evaluation of the last integral ) .
thus at low temperatures the contribution of @xmath243 to the free energy , given by @xmath256 dominates over the @xmath257 contribution coming from the simple roots @xmath206 . moreover , since the coefficient of @xmath258 in eq .
is always negative , this term can not be compensated by similar terms in eq . coming from other multiple roots .
we thus conclude that when @xmath199 , but the equation @xmath33 has at least one multiple root , the model can not be critical .
a similar analysis shows that this is also the case when @xmath259 or @xmath260 is smooth on @xmath21 $ ] , so that @xmath261 vanishes at the points where @xmath103 attains its maximum and minimum values over the latter interval . ] .
this shows that the model is critical if and only if @xmath30 and all the roots of the equation @xmath33 are simple . when that is the case , the central charge of the model is equal to the number of connected components of its fermi sea ( or , equivalently , half the number of connected components of its fermi `` surface '' ) .
thus , the universality class of the model depends exclusively on the topology of its fermi sea , which confirms the general assertion in ref .
as mentioned in the introduction , one of the hallmarks of a critical fermionic lattice model in one dimension with short - range interactions is the logarithmic growth of its ground state bipartite entanglement entropy with the length @xmath13 of the block of spins considered .
more precisely , let @xmath262 denote the rnyi entropy of the block when the whole chain is in its ground state @xmath263 , where @xmath264 .
the expected behavior of @xmath265 in this type of models is then @xmath266 where @xmath17 is the central charge of the corresponding virasoro algebra and @xmath36 is a non - universal constant ( independent of @xmath13 ) .
we showed in a previous paper @xcite that the latter formula is also valid for the supersymmetric chains when their dispersion relation is monotonic ( and nonnegative ) in the interval @xmath21 $ ] , even in the case of long - range interactions . in this section
we shall extend this result to a general model of the type , whose dispersion relation need not be monotonic ( or nonnegative ) in @xmath21 $ ] . to this end ,
recall first of all that the ground state entanglement entropy @xmath265 can be expressed in terms of the eigenvalues of the ground state correlation matrix @xmath267 , with matrix elements @xmath268 indeed , it was shown in ref .
@xcite that @xmath269 where @xmath270\ ] ] and @xmath271 $ ] are the eigenvalues of the matrix @xmath272
. the asymptotic behavior of @xmath265 can be determined following the method developed by jin and korepin @xcite for the @xmath9 model . to this end , for @xmath273 we define the complex - valued function @xmath274\,,\ ] ] where @xmath275}z$ ] and @xmath276 .
this function has a logarithmic branch cut on the set @xmath277 and no other singularities on a sufficiently small open subset ( independent of @xmath278 ) containing the interval @xmath279 $ ] then for @xmath280 $ ] @xmath281 is real and satisfies @xmath282 . on the other hand ,
if @xmath283 then for @xmath284 with @xmath285 and @xmath280 $ ] integrating the derivative of @xmath286 along the segment from @xmath52 to @xmath287 we obtain @xmath288 thus @xmath289 for @xmath280 $ ] provided that @xmath290 . ] . by cauchy s theorem and eq
. , if @xmath291 is the path sketched in fig .
[ fig.path ] we therefore have @xmath292 where @xmath293 as explained in appendix [ app.fh ] , the latter integral can then be approximated using a proved case of the fisher hartwig conjecture to estimate the logarithmic derivative of @xmath294 . in eq . .
] in order to derive the asymptotic behavior of @xmath294 , we first need to determine the symbol of the toeplitz matrix @xmath295 ( see again appendix [ app.fh ] for the definition of the symbol and its calculation in two simple cases ) .
we shall compute this symbol for a general model of the type , whose dispersion relation is not assumed to be monotonic over @xmath21 $ ] .
more precisely , we shall only suppose that the equation @xmath33 has @xmath200 _ simple _ roots @xmath201 in the interval @xmath164 .
from the symmetry of @xmath103 around @xmath94 ( cf .
it then follows that the remaining roots of the equation @xmath33 in the interval @xmath296 are @xmath297 . in general ,
the system s ground state @xmath263 is determined by the conditions for all @xmath298 , so that the ground state is unique . ]
@xmath299 \hat a_k{|\psi\rangle}&=0\,,\qquad { \varepsilon}_n(k)>\mu\ , , \end{aligned } \right.\ ] ] so that @xmath300 it immediately follows from eq . that the matrix elements of the correlation matrix @xmath267 are given by @xmath301 where the sum ranges over the set @xmath302 of integers in the range @xmath303 $ ] satisfying the condition @xmath304 .
in the thermodynamic limit @xmath91 the latter formula becomes @xmath305 where the integral is extended to the subset of the interval @xmath296 defined by the inequality @xmath204 .
in fact , by the @xmath98-periodicity of the integrand we can replace the interval @xmath296 by any interval of length @xmath98 , which we shall take as @xmath306 $ ] .
let us suppose , for definiteness , that @xmath307 ( the case @xmath308 is dealt with similarly ) . from the simple nature of the roots @xmath309 , @xmath310 , it then follows that the subintervals of the interval @xmath311 on which @xmath312 is negative are @xmath313 with @xmath314 , and @xmath315 with @xmath316 . by eq . , the symbol of the toeplitz matrix @xmath317 is given by @xmath318 thus @xmath319 is piecewise constant and alternates between the two values @xmath320 and @xmath321 .
the discontinuities of this symbol at the points @xmath322 ( with @xmath323 ) suggest the ansatz @xmath324 for suitable @xmath325 and @xmath326 . to verify this ansatz
, we note that for @xmath327 ( with @xmath323 ) we have @xmath328 whereas for @xmath329 ( with @xmath330 ) @xmath331 and thus in either case @xmath332 comparing the latter formula with eq .
we arrive at the system @xmath333 these equations easily imply that @xmath334 is an integer multiple of @xmath98 for @xmath335 .
we shall take @xmath336 , so that calling @xmath337 we have @xmath338 from the equations with @xmath339 and @xmath340 we then obtain @xmath341 so that we can take @xmath342 finally , from the equation with @xmath340 we have @xmath343{\mathrm e}^{-2{\mathrm i}{\beta}\sum_{k=0}^m(-1)^kp_k}\\ = \big[{\lambda}+(-1)^{m}\big]\bigg(\frac{{\lambda}+1}{{\lambda}-1}\bigg)^{-\sum_{k=0}^m(-1)^k\frac{p_k}\pi}\,,\end{gathered}\ ] ] which can also be written as @xmath344 where @xmath345 and @xmath346 denotes the parity of @xmath347 .
it is easy to check that with this choice of @xmath325 and @xmath326 eqs .
are all satisfied . since eq .
coincides with the first eq . ,
as explained in appendix [ app.fh ] , the condition @xmath348 is satisfied , so that we can apply the fisher hartwig conjecture to estimate @xmath349 .
to this end ( using the notation in appendix [ app.fh ] ) , note first of all that @xmath350 and , by eq . , @xmath351 moreover , from eq .
it easily follows that @xmath352^{2(m+1)}\ ] ] and @xmath353^{4(-1)^{i+j}{\beta}^2}\\ & \times \prod_{0{\leqslant}j < i{\leqslant}m}\big[2\sin\bigl(\tfrac{p_i+p_j}2\bigr)\big]^{-4(-1)^{i+j}{\beta}^2}\,.\end{aligned}\ ] ] equation and the fisher hartwig conjecture thus yield the asymptotic formula @xmath354^{-2{\beta}^2 } ( { \lambda}+1)^l\bigg(\frac{{\lambda}+1}{{\lambda}-1}\bigg)^{\!\!-lp}\\\label{dlgen } \times \big[g(1+{\beta})g(1-{\beta})\big]^{2(m+1)}\big(1+o(1)\big)\,,\end{gathered}\ ] ] with @xmath355^{(-1)^{i+j}}\end{gathered}\ ] ] independent of @xmath13 and @xmath356 .
we shall next use the approximate formula and eq . to derive an asymptotic formula for the rnyi entanglement entropy of the ground state of a general model of the form in the limit @xmath28 .
first of all , from eq .
we easily obtain @xmath357\ , , \label{dlapprox}\end{gathered}\ ] ] with @xmath358\\= 1+{\gamma}_e+\sum_{n=1}^\infty\frac{{\beta}^2/n}{n^2-{\beta}^2}\end{gathered}\ ] ] ( cf .
. in fact , the dominant term ( proportional to @xmath13 ) in the previous expression does not contribute to eq .
, since by cauchy s residue theorem we have @xmath359 thus eqs . - yield @xmath360{\mathrm{d}}{\lambda}\,.\end{gathered}\ ] ] moreover , it is straightforward to verify that the integral along the circular arcs of @xmath291 vanishes identically , since each of these arcs is mapped to the opposite of the other by the transformation @xmath361 , and the integrand changes sign under the latter mapping ( cf .
eqs . , and ) .
we thus obtain @xmath362{\mathrm{d}}{\lambda}\,.\end{gathered}\ ] ] in order to evaluate these integrals , we note that along the segments @xmath363 with @xmath364 we have @xmath365 so that @xmath366 on the other hand , @xmath367 is negative for sufficiently small @xmath368 , while @xmath369 tends to @xmath370 as @xmath371 and has the same sign as @xmath372 , so that @xmath373}w=\mp\pi\,.\ ] ] we thus have @xmath374 \equiv -{\mathrm i}b(x)\mp\frac12\,,\ ] ] with @xmath375 from eq
. it immediately follows that @xmath376i_1({\alpha})\\+(m+1)i_2({\alpha})\,,\end{gathered}\ ] ] with @xmath377 { \mathrm{d}}x\ , .
\label{i2}\end{aligned}\ ] ] the value of the integral @xmath378 can be deduced from ref .
@xcite ( cf .
also @xcite ) , namely @xmath379 ( see appendix [ app.int ] for an elementary derivation of the latter formula ) .
we thus obtain @xmath380 where @xmath381 { \mathrm{d}}x\ , .
\label{calfinal}\end{gathered}\ ] ] comparing with eq .
, we see that the ground - state rnyi entanglement entropy of the model behaves as that of a critical system with central charge @xmath242 , as expected .
moreover , the constant term @xmath36 is given in this case by @xmath382 where the first term is model dependent ( it depends on @xmath19 and @xmath20 through the momenta @xmath206 ) , while @xmath383 is a universal constant ( independent of @xmath13 and @xmath206 ) characteristic of the class of models under consideration .
it is shown in appendix [ app.cal ] that @xmath383 in eq . can be expressed as @xmath384{\mathrm{d}}x\,,\ ] ] where @xmath385 is the digamma function .
in particular , eq .
implies that @xmath383 coincides with the function @xmath386 defined in eq .
( 64 ) of ref .
since for @xmath387 we have @xmath388 , eq . yields the formula derived in ref . @xcite for the rnyi entanglement entropy of the model when its dispersion relation is monotonic over the interval @xmath21 $ ] ( which , as explained in the latter reference , includes the @xmath9 model studied in ref .
in fact , using the ideas of ref .
@xcite eq . can be written in the simpler form @xmath389 ( see appendix [ app.cal ] for details ) . from the previous expression it is straightforward to evaluate @xmath383 numerically for any specific value of the rnyi parameter @xmath390
[ fig.cal ] .
it can be numerically verified that @xmath383 vanishes for @xmath391 , and attains its maximum value @xmath392 for @xmath393 ( cf .
[ fig.cal ] ) . in eq .
] it also follows from eq .
that @xmath394 as @xmath395 , and that when @xmath396 @xmath383 tends to a finite ( nonzero ) limit , given by @xmath397 taking the @xmath398 limit in the previous formulas we obtain the following asymptotic expression for the von neumann entropy @xmath399 : @xmath400 where @xmath401 and the universal constant @xmath402 is given by @xmath403 note , in particular , that the latter equation agrees with the formula for the analogous constant @xmath404 in ref .
@xcite .
the formula - ( or its counterpart for the von neumann entropy ) provides an excellent approximation to the ground - state rnyi entanglement entropy of the supersymmetric chain for even moderately large values of @xmath13 . as an example , in fig .
[ fig.error ] we have represented the relative error @xmath405 , where @xmath406 is the approximation to the von neumann entropy @xmath7 , for the finite - range chain in the case @xmath407 and @xmath408 .
the value of @xmath7 has been numerically computed diagonalizing the correlation matrix @xmath267 and using the exact formula ( with @xmath16 ) .
as explained in section [ sec.disp ] , for the value of @xmath409 considered the dispersion relation has exactly one maximum in the interval @xmath164 , and hence is not monotonic . in particular , for @xmath410 the fermi sea consists of two disjoint intervals , as is also apparent from the inset in fig .
[ fig.error ] . as can be seen from the latter figure , the relative error decreases ( though not monotonically ) from @xmath411 to @xmath412 when @xmath13 ranges from @xmath413 to @xmath414 . of the approximation @xmath406 in the rhs of eq . to the von neumann ground - state entanglement entropy of the chain with @xmath407 and @xmath408 as a function of the block length @xmath13 .
inset : dispersion relation of the chain with the latter values of @xmath409 and @xmath19 .
( the interval @xmath415 is in this case is @xmath416 , @xmath417 , @xmath418 and @xmath419 . ) ]
in this paper we have analyzed the critical behavior of a large class of supersymmetric spin chains whose dispersion relation @xmath20 is not assumed to be monotonic in the interval @xmath21 $ ] . we have examined the conditions under which the dispersion relation is well - defined ( i.e. , is a continuous function ) in the thermodynamic limit , providing several simple examples of models of this type , with both short- and long - range interactions , whose dispersion relation is not monotonic .
the main conclusion of our work is that the criticality properties of the supersymmetric chains are determined exclusively by the topology ( the number of points ) of their fermi `` surface '' .
more precisely , through the analysis of the free energy per spin in the critical ( gapless ) phase , we have shown that these models are equivalent to a system of @xmath34 free bosons with fermi velocities @xmath420 , where @xmath421 are the points of the fermi surface in the interval @xmath21 $ ] .
in particular , the central charge is equal to the number @xmath34 of connected components ( intervals ) of the fermi sea .
this result is corroborated by the asymptotic behavior of the ground - state rnyi entanglement entropy @xmath265 as the block size @xmath13 tends to infinity , which has been derived applying a proved case of the fisher hartwig conjecture .
indeed , we have shown that @xmath422 , where @xmath36 is a nonuniversal constant ( independent of @xmath13 ) which we have computed in closed form in terms of the momenta @xmath421 . in particular , for large @xmath13 the entanglement entropy exhibits the logarithmic growth characteristic of @xmath1-dimensional conformal field theories with central charge @xmath242 .
this behavior , which is typical of critical ( fermionic ) one - dimensional lattice models with short - range interactions ( see , e.g. , @xcite ) , was recently established by the authors for supersymmetric spin chains of the type considered here under the assumption that the dispersion relation is monotonic in @xmath21 $ ] .
the present work opens up several possible lines for future research . in the first place
, one could consider a generalization of our results on the ground - state entanglement entropy to more general situations ( for instance , considering excited states , as in ref .
@xcite ) , in which the fermi sea is not necessarily a finite union of disjoint intervals but exhibits a more complicated topological structure
. another interesting generalization of the present work is the analysis of the entanglement of a subset consisting of the union of two or more disjoint blocks .
in fact , the entanglement entropy of this type of subsystems has already been discussed in ref .
@xcite , giving rise to an unproved conjecture on the asymptotic behavior of the determinant of a block toeplitz matrix .
the authors would like to thank p. tempesta for several enlightening discussions .
this work was partially supported by spain s mineco under research grant no .
fis2015 - 63966-p .
jac would also like to acknowledge the financial support of the universidad complutense de madrid through a 2015 predoctoral scholarship .
in this appendix we shall briefly review the fisher hartwig conjecture on the asymptotic behavior of the determinant of a toeplitz matrix when its order tends to infinity .
( recall that a matrix @xmath27 is toeplitz if its matrix elements @xmath423 depend only on @xmath424 . )
if @xmath425 is a ( complex - valued ) function defined on the unit circle @xmath426 , we define its fourier coefficients @xmath427 ( @xmath428 ) by @xmath429 note that the last integral can in fact be extended to any interval of length @xmath98 , by the @xmath98-periodicity of the integrand . for any @xmath430
, the function @xmath431 defines a toeplitz matrix @xmath432 of order @xmath13 through the relation @xmath433 we shall say that the function @xmath17 is the _ symbol _ of the toeplitz matrix @xmath432 .
the fisher hartwig conjecture applies to matrices @xmath432 whose symbol satisfies certain requirements that we shall next describe .
more precisely @xcite , @xmath17 should be of the form @xmath434^{{\alpha}_r}\ , , \label{cz}\ ] ] where @xmath435 , @xmath436 is a nonvanishing smooth function with zero winding number and @xmath437 note that @xmath438 has in general ( i.e. , unless @xmath236 is an integer ) a single jump discontinuity at @xmath439 . if @xmath17 satisfies eq .
, we denote by @xmath440 ( @xmath428 ) the @xmath63-th fourier coefficient of @xmath441 ( which is well defined and smooth , from the smoothness of @xmath325 and the assumption on its winding number ) , and define @xmath442 it is immediate to show that @xmath443 ( resp .
@xmath444 ) can be analytically prolonged to the interior ( resp .
exterior ) of the unit circle .
it also follows from the definition of @xmath445 that on the unit circle we have the _ wiener hopf decomposition _ @xmath446
let us further set @xmath447=\exp\bigg(\sum_{n=1}^{\infty}n\,l_nl_{-n}\bigg)\ ] ] and @xmath448\prod_{r=1}^rb_+\bigl({\mathrm e}^{{\mathrm i}{\theta}_r}\bigr)^{{\beta}_r-{\alpha}_r}b_-\bigl({\mathrm e}^{{\mathrm i}{\theta}_r}\bigr)^{-{\alpha}_r-{\beta}_r}\\ & \hphantom{={}}\times\prod_{1{\leqslant}s\ne r{\leqslant}r}\big(1-{\mathrm e}^{{\mathrm i}({\theta}_s-{\theta}_r)}\big)^{({\alpha}_r+{\beta}_r)({\beta}_s-{\alpha}_s)}\\ & \hphantom{={}}\times\prod_{r=1}^r\frac{g(1+{\alpha}_r+{\beta}_r)g(1+{\alpha}_r-{\beta}_r)}{g(1 + 2{\alpha}_r)}\,,\end{aligned}\ ] ] where the barnes @xmath449-function is the entire function defined by @xmath450\ ] ] and @xmath451 is the euler mascheroni constant .
the fisher hartwig conjecture states @xcite that if @xmath432 is the toeplitz matrix with symbol then when @xmath28 we have @xmath452 with @xmath453 the above conjecture was actually proved by bttcher and silbermann @xcite in the case @xmath454 which , as we shall see below , is the relevant one for our purposes .
in fact , as explained in section [ sec.gsee ] , we shall only need to consider the case in which @xmath455 for all @xmath116 and @xmath325 is a constant ( i.e. , independent of @xmath456 ) .
the fisher hartwig conjecture simplifies considerably in this case , since @xmath457=1\,,\quad{\mathrm e}^{l_0}=b\,,\ ] ] and therefore @xmath458 with @xmath459 and @xmath460 note also that the product @xmath461 reduces to @xmath462\ , .
\label{gsimp}\ ] ] we shall be mainly interested in the case in which @xmath463 , where @xmath356 is a spectral parameter and @xmath267 is the correlation matrix of a block of @xmath13 spins of the @xmath0 supersymmetric spin chain . as a first example , let us express in the form - the symbol of the matrix @xmath432 when the chain s dispersion relation is monotonic in the interval @xmath21 $ ] . to begin with , in this case we have @xmath464 where @xmath465 $ ] is the fermi momentum @xcite .
thus the symbol of the toeplitz matrix @xmath267 is @xmath466 and that of @xmath432 is therefore given by @xmath467 note that @xmath17 has two jump discontinuities on the unit circle at the points @xmath468 .
we shall next show that @xmath469 for suitable @xmath236 and @xmath470 .
indeed , first of all we have @xmath471 on the other hand , if @xmath472 then @xmath473 while for @xmath474 we have @xmath475 hence @xmath476 in order for eq .
to hold we must therefore have @xmath477 from which we easily get @xmath478 although these equations admit an infinite number of solutions @xmath479 provided that @xmath480 , it will prove convenient for our purposes to take @xmath481 where @xmath275}z$ ] and @xmath276 .
note , in particular , that @xmath325 is a nonvanishing constant .
it is also important to observe that @xmath482\implies|{\operatorname{re}}{\beta}|=\frac1{2\pi}\arg_{(-\pi,\pi]}\bigg(\frac{{\lambda}+1}{{\lambda}-1}\bigg)<\frac12\,,\ ] ] since by definition @xmath483}z{\leqslant}\pi$ ] and @xmath484}\bigg(\frac{{\lambda}+1}{{\lambda}-1}\bigg)=\pi&\iff \frac{{\lambda}+1}{{\lambda}-1}\in(-\infty,0)\\&\iff{\lambda}\in(-1,1)\,.\end{aligned}\ ] ] thus the fisher hartwig conjecture can be applied provided that @xmath356 lies outside the closed interval @xmath279 $ ] , with @xmath485 , @xmath486 and @xmath487 by eqs . and , when @xmath28 the characteristic polynomial @xmath488 is given by @xmath489 with @xmath236 defined by eq .
this is precisely the formula used by jin and korepin @xcite for the determination of the asymptotic behavior of the ground state entanglement entropy of the @xmath9 model . as a second example
, we shall consider a simple case in which the dispersion relation @xmath103 is not monotonic in @xmath21 $ ] . more precisely ,
suppose that @xmath103 is nonnegative and has a single maximum in the open interval @xmath164 ( cf .
[ fig.disp ] ) . with a single maximum in @xmath164 .
] for @xmath490 , the equation @xmath33 has one root @xmath491 in the interval @xmath164 , and the determinant @xmath294 is approximately given by eq . .
we shall next determine the asymptotic behavior of @xmath294 when @xmath492 , where @xmath32 is the maximum value of @xmath20 , and thus the equation @xmath33 has two roots @xmath493 in the interval @xmath164 . to begin with , from eq .
it follows that in this case @xmath494 thus the symbol of @xmath317 is given by @xmath495 in other words , @xmath496 alternatively takes on the two values @xmath320 and @xmath321 on each of the four intervals @xmath497 on which @xmath312 has constant sign , starting with @xmath320 . since the symbol has four discontinuities at the points @xmath468 and @xmath498 , we shall try to express it as @xmath499 for suitably chosen @xmath325 , @xmath500 .
in fact , we only need compute @xmath501 , which is straightforward : @xmath502 combining the previous equations with eq . and comparing with eq .
we immediately arrive at the system @xmath503 from the first equation it follows that @xmath504 must be an integer . choosing the simplest solution @xmath505 and dividing the last equation by the first one we again obtain eq . for @xmath236 .
finally , from the last equation it follows that @xmath506 note that @xmath236 is still given by eq .
, so the condition @xmath348 , necessary for the validity of the fisher hartwig conjecture , also applies in this case if @xmath507 $ ] . since now @xmath508 , @xmath509 and @xmath510 by the fisher hartwig conjecture , the determinant @xmath294 is given in this case by @xmath511^{\!\!-2{\beta}^2}\notag\\ & \hphantom{={}\bigg(}\times({\lambda}+1)^l\bigg(\frac{{\lambda}+1}{{\lambda}-1}\bigg)^{\frac l\pi(p_1-p_0-\pi)}\notag\\ & \hphantom{={}\bigg(}\times g(1+{\beta})^4g(1-{\beta})^4\big(1+o(1)\big)\ , .
\label{dlp0p1}\end{aligned}\ ] ]
in this appendix we shall provide an elementary derivation of the integral @xmath378 in eq . , which appears in the asymptotic expression of the rnyi entanglement entropy of the model .
to begin with , we have @xmath512 with @xmath513\,\frac{{\mathrm{d}}x}{1-x^2}\,,\ ] ] or equivalently ( performing the change of variables @xmath514 ) @xmath515\frac{{\mathrm{d}}t}t\,.\ ] ] integrating by parts we obtain the equivalent expression @xmath516 on the other hand , differentiating eq . with respect to @xmath15
we easily get @xmath517 and hence @xcite @xmath518 solving this linear differential equation with the initial condition @xmath519 we finally obtain @xmath520 which immediately yields eq . .
in this appendix we derive eq . for the constant @xmath383 in eq . , and show that it can be more simply expressed by means of the integral . to this end
, we use the elementary identity @xmath521 which can be immediately derived from the well - known infinite product for the gamma function .
our starting point is the definition of @xmath522 , which can be written as @xmath523 { \mathrm{d}}x\,,\ ] ] with @xmath524 and @xmath525 from the relation @xmath526 and the functional identity @xmath527 satisfied by the digamma function it immediately follows that @xmath528-z{\gamma}_e-\frac12\,,\ ] ] and therefore @xmath529\,.\ ] ] substituting into eq . and using eqs . and we easily obtain eq . .
in order to prove eq . , we first make the change of variable @xmath530 in eq . , which yields @xmath531\\&\kern9em\times\big[\psi\bigl(\tfrac12+{\mathrm i}\,w\bigr )
+ \psi\bigl(\tfrac12-{\mathrm i}\,w\bigr)\big]\,{\mathrm{d}}w\\ & = \frac2{{\mathrm i}\pi}\int_0^\infty\log\biggl[\frac{2\cosh(\pi{\alpha}w)}{\big(2\cosh(\pi w)\big)^{\alpha}}\biggr]\frac{{\mathrm{d}}}{{\mathrm{d}}w}\log\bigg [ \frac{{\gamma}\bigl(\tfrac12-{\mathrm i}\,w\bigr)}{{\gamma}\bigl(\tfrac12+{\mathrm i}\,w\bigr)}\bigg]\,{\mathrm{d}}w\,.\end{aligned}\ ] ] we next integrate by parts , taking into account that by stirling s formula we have @xmath532 = o(w\log w)\ ] ] while @xmath533=o\bigl({\mathrm e}^{-2\pi\min(1,{\alpha})w}\bigr)\,,\ ] ] so that the boundary term vanishes .
we thus obtain @xmath534\log\bigg [ \frac{{\gamma}\bigl(\tfrac12-{\mathrm i}\,w\bigr)}{{\gamma}\bigl(\tfrac12+{\mathrm i}\,w\bigr)}\bigg]{\mathrm{d}}w\,.\end{gathered}\ ] ] on the other hand , from gauss s integral representation of the digamma function @xcite @xmath535 it easily follows that @xmath536\frac{{\mathrm e}^{-t}}t\,{\mathrm{d}}t\ ] ] and hence @xmath532 = { \mathrm i}\int_0^\infty\bigg[{\operatorname{csch}}(t/2)\sin(wt)-2w{\mathrm e}^{-t}\bigg]\frac{{\mathrm{d}}t}t\,.\ ] ] substituting into the last formula for @xmath383 and using the elementary identity @xmath537 we obtain @xmath538 with @xmath539{\mathrm{d}}w.\ ] ] the latter integral can be evaluated in closed form by elementary means .
indeed , @xmath540 while @xmath541 and therefore @xmath542 we thus obtain @xmath543 from which eq . easily follows .
32ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop _
_ , ed .
( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , , vol . , ( ) pp . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) , , , and , eds . , @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) | we study the critical behavior and the ground - state entanglement of a large class of @xmath0 supersymmetric spin chains with a general ( not necessarily monotonic ) dispersion relation .
we show that this class includes several relevant models , with both short- and long - range interactions of a simple form .
we determine the low temperature behavior of the free energy per spin , and deduce that the models considered have a critical phase in the same universality class as a @xmath1-dimensional conformal field theory ( cft ) with central charge equal to the number of connected components of the fermi sea .
we also study the rnyi entanglement entropy of the ground state , deriving its asymptotic behavior as the block size tends to infinity .
in particular , we show that this entropy exhibits the logarithmic growth characteristic of @xmath1-dimensional cfts and one - dimensional ( fermionic ) critical lattice models , with a central charge consistent with the low - temperature behavior of the free energy .
our results confirm the widely believed conjecture that the critical behavior of fermionic lattice models is completely determined by the topology of their fermi surface . |
the two - dimensional time - dependent ginzburg - landau equation ( tgle ) describing the behavior of the film is written , in dimensionless variables as : @xmath54 where @xmath55 is the complex , dimensionless superconducting order parameter normalized with respect to @xmath0 .
we consider the static case where the external fields are absent , thus the electro - magnetic potentials can be absorbed into the order parameter .
we include a small sixth order term in the gl free energy to ensure a numerical stability .
the coefficient @xmath56 is a functional of the order parameter and includes the effects of long - range elastic potentials , temperature @xmath57 , and disorder : @xmath58+\int d\mathbf{r } ^{\prime } u(\mathbf{r}-\mathbf{r}^{\prime } ) |\psi ( \mathbf{r}^{\prime } , t)|^{2}\,,\ ] ] where @xmath59 describes quenched fluctuations ( @xmath60 $ ] ) of @xmath0 and @xmath61 is the non - local kernel generated by elasticity interactions .
the temperature @xmath57 is measured in units of superconducting transition temperature @xmath0 , and the unit of the length is the superconducting , zero temperature coherence length @xmath62 .
to couple elasticity and superconductivity we utilize the pressure dependence of @xmath63 which is taken in a linearized form as @xmath64 where @xmath65 is the change in the internal pressure induced by elastic forces .
the latter is derived from the condition of the continuity of the deformation fields @xmath66 and the stress balance at the film - substrate interface ( the @xmath67 plane ) , @xmath68and @xmath69 and at the film - vacuum ( @xmath70 , @xmath71 is the thickness of the film ) interface : @xmath72 the superscripts @xmath73 and @xmath74 refer to the film and to the substrate respectively .
these nine coupled equations are solved using a plain - wave ansatz , yielding the exact expressions for the deformation fields which , in their turn , are used for calculating the internal pressure of the film .
the fourier transform of the pressure in the film then reads : @xmath75 using this expression together with the relation @xmath76 in eq .
( [ eq.tcp ] ) one finds the parameter @xmath77 of the tdgl in the form given in ( [ eq.alpha ] ) .
we define the interaction potential @xmath78 through its fourier transform @xmath79 , where @xmath80 $ ] and @xmath38 is the scale free part of @xmath81 [ note , that @xmath82 . here @xmath8 is the bulk modulus , the linear thermal expansivity of the film is @xmath83 , and the relative volume change due to the change in pressure @xmath65 is @xmath84 .
the tgle ( [ eq.tgle ] ) is solved on a fine discrete grid by the numerical integration in real- and fourier spaces ( _ quasi - spectral split - step method _ ) , taking into account the full non - local elastic potential .
we use periodic boundary conditions and @xmath85 grid points for a system of the size @xmath86 .
the temperature is chosen as @xmath48 and the coupling constant of @xmath87 ( in figs . 2 and 3 ) for a film of thickness @xmath88 .
the used gl parameters are @xmath89 , @xmath90 , @xmath91 , and @xmath92 .
the parameters of the elastic kernel are @xmath93 , @xmath94 ( shear moduli ) and @xmath95 ( modified poisson numbers ) see supplementary materials for details .
here we derive @xmath97 for the problem of a thin superconducting elastic film coupled to a rigid substrate with different elastic properties and lattice mismatch . in the following ,
we use the superscripts @xmath73 for all quantities related to the film , and @xmath74 for the substrate .
both materials are described by the three dimensional displacement fields @xmath98/@xmath99 .
we start our consideration using a plain wave ansatz for these fields : @xmath100 where the misfit strain on the film is captured by the real exponential terms of the film in @xmath101-direction . @xmath102 and @xmath103 are special solutions of the two - dimensional problem due to superconductivity , which we will discuss later . in order to determine the nine free parameters @xmath104 , @xmath105 , and @xmath106 we need nine equations .
first , we consider the interface film - vacuum at @xmath70 , where @xmath71 is the thickness of the film .
the stress boundary condition for a flat surface demands : @xmath107 .
so , for the film we get from hooke s law @xmath108\,,\end{aligned}\ ] ] where the effect of a lattice mismatch of both materials is captured by the @xmath109-term . in general , the lattice constants of both materials are different , i.e. , the misfit parameter @xmath110 is non - zero , where @xmath111 and @xmath112 are the lattice constants of the film and substrate , respectively .
this lattice mismatch leads to a compression of the film - lattice at the interface , such that the lattice spacings match there .
we note , that this description not only applies to crystalline structures but also for amorphous materials , where the lattice constants are averaged quantities .
the coefficients @xmath113 and @xmath114 are the shear modulus and the modified poisson number of the film [ @xmath115 , where @xmath116 is the poisson number ] . in the following we use the fact , that all interface equations are independent of the @xmath117 and @xmath118 coordinate and set @xmath119 .
therefore , the first three of the nine equations are @xmath120 next , we consider the interface between film - substrate at @xmath67 , for which we also need the expressions for the stress tensor elements @xmath121\,.\end{aligned}\ ] ] at this interface we need to fulfill six more conditions , where three of them are resulting from the stress balance @xmath122and the last three equations are the continuity equations for the displacement fields @xmath123 these nine equations can now be solved in order obtain the free parameters of our ansatz .
the two special solutions for @xmath124 in x- and y - direction , @xmath125 , already used in the ansatz for the deformation fields of the film are obtained through solution of the extended stress - strain balance @xmath126 which takes into account that the systems energy is the sum of the superconducting energy and the mechanical deformation energy .
this relation introduces two phenomenological parameters : @xmath127 and @xmath128 which together define the coupling constant of the elastic interaction and superconductivity which is determined explicitly below .
( [ eq.couple ] ) can be solved and gives @xmath129 for @xmath130 . using the solutions for all coefficients we find the internal pressure of the film @xmath131 this expression simplifies to @xmath132 if we use the homogeneous parametrization @xmath133 , @xmath134 . due to the special solution for @xmath124
the pressure @xmath135 is proportional to @xmath136 , where @xmath137 are the fourier components of the order parameter .
now , we need to consider the connection of elasticity and superconductivity . for that
, we first expand @xmath138 in changes of ( internal ) pressure : @xmath139 the pressure change @xmath65 is related to a volume change @xmath140 as @xmath141 , where @xmath8 is the bulk modulus .
therefore we can write @xmath142 where @xmath143 is the relative length change of the system and @xmath144 the linear thermal expansivity of the film . using the above result in the pressure - expanded critical temperature , the parameter @xmath77 in the tgle obtains thus the non - local form shown in the * methods summary * section : @xmath145+\int d\mathbf{r}^{\prime } u(\mathbf{r}-\mathbf{r}^{\prime } ) |\psi ( \mathbf{r}^{\prime } , t)|^{2}\,,\ ] ] where @xmath146 includes also spatial variations of @xmath138 due to weak disorder .
if we define the scale free elastic kernel @xmath147 where the explicit form of the elastic kernel is quite involved and given in section [ app.epot ] . finally , we write the fourier transform of @xmath36 as @xmath79 , where @xmath148 $ ] the tgle equation @xmath54 is solved by a quasi - spectral split step method for periodic boundary conditions on a two - dimensional square grid with grid size of @xmath149 for a system size @xmath150 .
_ split step _ means , that for each time step the equation is solved partially in real space and partially in fourier space ( the diffusion part ) @xcite . in the first step we calculate @xmath151 where the non - local part of @xmath77 is calculated by fast - fourier - transformation ( fft ) upfront . in general
the tgle without long - range potential has only two stable solutions : a homogeneous one and a striped phase .
even if a long - range potential is present , the stability of these two solutions is only destroyed under certain conditions which can be found out by linear stability analysis ( see next section ) .
e.g. , a coulomb potential or even screened coulomb potential can not create this kind of instability .
however , the elastic potential does , if the model parameters are chosen appropriately .
another possible non - local potential which destroys the homogeneous / striped solutions is a box - potential @xcite @xmath153 . in order to determine whether the elastic long - range interactions lead to an instability of the homogeneous solution @xmath18 of the gl equation
, we need to check the linear stability of the equation , by expanding the order parameter around @xmath18 . after this expansion and fourier transformation
, we find that the stability depends on the behavior of the function @xmath154 we know that @xmath155 and @xmath156 and therefore @xmath157 ( for @xmath158 ) and @xmath159 for @xmath160 .
if @xmath161 changes its sign at intermediate @xmath162 the homogeneous solution becomes unstable , see fig .
[ fig.stab ] .
@xmath165 \right .
{ \nonumber}\\ & & \left .
+ d~q\left [ e^{4d~q}\mu_{s}\vartheta^{{(s)}}\left ( 2\mu^{{(p)}}+2\mu^{{(p)}}\nu^{{(p)}}\sigma_{0}-\mu^{{(s)}}\vartheta^{{(s)}}\right ) \right . \right . { \nonumber}\\ & & \left .
~~~\left .
-\mu^{{(s)}}\vartheta^{{(s)}}\left ( \mu^{{(s)}}\vartheta^{{(s)}}+ 2\mu^{{(p)}}\left [ 1+\nu^{{(p)}}+\nu^{{(s)}}-\nu^{{(p)}}\sigma_{0}\right ] \right ) \right .
{ \nonumber}\\ & & \left .
~~~\left .
-2e^{2d~q}\left ( { \mu^{{(p)}}}^{2}\left [ 1 + 2\nu^{{(p)}}\right ] \left [ 1 + 2\nu^{{(s)}}\sigma_{0}\right ] -{\mu^{{(s)}}}^{2}{\vartheta^{{(s)}}}^{2}-\mu^{{(p)}}\mu^{{(s)}}\vartheta^{{(s)}}(\nu^{{(p)}}+\nu^{{(s)}}-2\nu^{{(p)}}\sigma_{0})\right ) \right ] \right .
{ \nonumber}\\ & & \left .
+ \mu^{{(s)}}\left ( e^{d~q}-1\right ) ^{2}\left [ 4\mu^{{(s)}}\nu^{{(s)}}\left ( e^{2d~q}-1\right ) \vartheta^{{(s)}}\left [ \sigma_{0}-1\right ] \right . \right . { \nonumber}\\ & & \left .
\left . ~+\mu^{{(p)}}\left
[ 1 + 2\nu^{{(s)}}\left ( 2 - 2\vartheta^{{(s)}}-3\sigma_{0}\right ) + 2\nu^{{(p)}}\left [ 2\vartheta^{{(s)}}+1\right ] \left [ \sigma_{0}-1\right ] -2e^{d~q}\left ( \nu^{{(s)}}\left ( 2\nu^{{(p)}}\left [ \sigma_{0}-1\right ] + 2\sigma_{0}-3\right ) -1\right ) \right .
\right . \right . { \nonumber}\\ & & \left .
~~~-e^{2d~q}\left ( 3\vartheta^{{(s)}}+2 + 2\nu ^{{(p)}}\left [ \sigma_{0}-1\right ] \left ( \nu^{{(s)}}\left [ 4\sigma_{0}-2\right ]
-1\right ) \right ) \right ] \right ] \right .
{ \nonumber}\\ & & \left .
+ 4{\mu^{{(p)}}}^{2}\nu^{{(s)}}e^{2d~q}\left [ 1 + 2\nu^{{(p)}}\right ] \left [ 4\sigma_{0}-3\right ] \sinh\left ( d~q\right ) \right\ } \right/ { \nonumber}\\ & & \left\ { 3d~q\left\ { -{\mu^{{(s)}}}^{2}\left ( e^{2d~q}-1\right ) ^{2}{\vartheta^{{(s)}}}^{2}\right .
{ \nonumber}\\ & & \left .
+ 2\mu^{{(p)}}\mu^{{(s)}}\left ( e^{2d~q}-1\right ) \vartheta^{{(s)}}\left ( 1+\nu^{{(p)}}+\nu^{{(s)}}-\nu^{{(p)}}\sigma_{0}+e^{2d~q}\left [ 1+\nu ^{{(p)}}\sigma_{0}\right ] \right ) \right
. \right .
{ \nonumber}\\ & & \left .
\left . + { \mu^{{(p)}}}^{2}(1 + 2\nu^{{(p)}})\left ( e^{4d~q}\vartheta^{{(s)}}+2\nu^{{(s)}}\left [ \sigma_{0}-1\right ] -1 - 2e^{2d~q}\left [ 1 + 2\nu^{{(s)}}\sigma_{0}\right ] \right ) \right\ } \right\}\end{aligned}\ ] ] for the simulations ( cf . evolution of amplitude and phase in figs . 2 and 3 ) we used the typical values : @xmath167 , @xmath168 , @xmath91 , @xmath169 , @xmath170 , @xmath171 , @xmath172 ( the substrate is more rigid ) , @xmath173 , @xmath174 ( the lattice mismatch in the normal state is not relevant ) , @xmath175 ( linear dimension of the system ) , and @xmath176 ( number of discrete grid points per dimension ) .
lasarev , b. & kan , l. measurements at low temperatures and high pressures : superconductivity of sn and in under uniform compression pressure of 1750 kg/@xmath177 . _
_ 14 ( 12 ) , 463 - 473 ( 1944 ) .
sacp , b. , chapelier , c. , baturina , t. i. , vinokur , v. m. , baklanov , m. r. & sanquer , m. disorder - induced inhomogeneities of the superconducting state close to the superconductor - insulator transition .
_ * 101 * , 157006 ( 2008 ) .
sambandamurthy , g. , engel , l.w . , johansson , a. , peled , e. & shahar , d. experimental evidence for a collective insulating state in two - dimensional superconductors .
lett . _ * 94 * , 017003 ( 2005 ) .
baturina , t. i. , mironov , a. yu . , vinokur , v. m. , baklanov , m. r. & strunk , c. localized superconductivity in the quantum - critical region of the disorder - driven superconductor - insulator transition in tin thin films .
_ * 99 * , 257003 ( 2007 ) . | self - organized regular patterns are ubiquitous in nature , and one of their most celebrated manifestations is the abrikosov vortex lattice@xcite : under an applied magnetic field , the homogeneous superconductivity becomes unstable and cast itself into a regular texture of the `` normal '' filaments , called abrikosov vortices , immersed into a superconducting matrix .
its prediction and the experimental discovery became a breakthrough in our understanding of superconductivity and founded a new direction in physics .
here we show that the interplay between the superconducting order parameter and elastic fields@xcite , which are intimately connected to the very existence of the superconductivity itself@xcite , can result in a novel superconducting state dual to the abrikosov state : a regular texture of superconducting islands .
the fact that both patterns emerge within the framework of the ginzburg - landau description of superconductivity@xcite indicates that the formation of regular structures may be a generic feature of any phase transition .
emergence of superconducting island arrays is not specific to the effect of the elastic forces , but can be caused by any inherent mechanism generating long - range non - local interactions in the ginzburg - landau functional , for example , by the coulomb forces .
in particular , our findings suggest the formation of a superconducting island textures as a scenario for a superconductor - to - insulator transition in thin films .
sixty years ago the volume change accompanying the transition of a superconductor from the normal to the superconducting state was first observed@xcite .
this discovery followed by finding the dependence of the superconducting critical temperature , @xmath0 , on the isotopic mass@xcite and
the change in elastic constants at the transition into the superconducting state@xcite geared the line of research crowned eventually by the triumph of the bardeen - cooper - schrieffer ( bcs ) microscopic theory of superconductivity , which demonstrated an intimate connection between superconductivity and elastic properties of the material@xcite . according to the bcs theory ,
@xmath0 is related to elasticity via@xcite @xmath1 where @xmath2 is the debye frequency , @xmath3 is the density of states at the fermi level , and @xmath4 is the effective interaction strength between electrons mediated by electron - phonon coupling .
all three parameters , @xmath5 , and @xmath4 in ( [ bcseq ] ) depend on pressure and comprise several microscopic effects including both the changes in the electron and phonon spectrum , as well as structural transformations .
the link between elastic properties and superconductivity has been illustrated by numerous experiments revealing the influence of an external pressure on @xmath0 ( see refs 11 and 12 for extensive reviews ) . mechanical stresses , which affect superconductivity , are inherent to thin films attached to rigid substrates@xcite ( fig.1 )
. indeed , enforcing on the film its own lattice spacing , the substrate impels _ internal _ strains in the film . on top of that , the rigid coupling between the film and the substrate can give rise to a peculiar scenario of the superconducting transition itself .
upon cooling the film , the superconductivity nucleates first in the regions where @xmath0 is elevated by fluctuations .
accordingly , mechanical properties , and , in particular , the lattice spacing in these regions should have changed . the rigid substrate , however , obstructs expansions ( contractions ) of the film associated with the local onset of superconductivity . as a result ,
additional local stresses emerge , which , in their turn , promote further growth ( and the appearance of new ) superconducting nuclei .
the role of the substrate is thus twofold : first , it exerts elastic forces in regions where superconducting droplets appear first and , second , the substrate mediates the elastic coupling between remote parts of the film , effectively transforming local distortions into long - range elastic coupling . appearing and evolving initial superconducting nuclei induce additional elastic distortions . as a result of this positive feedback between the superconducting order parameter and mechanical stress , a spatial instability of the superconducting state develops forming eventually a periodic pattern of superconducting islands ( see fig.1 ) .
our starting point is the time - dependent ginzburg - landau ( tdgl ) equation@xcite combined with elastic stress balance equations ( see methods ) .
this model offers a comprehensive description of formation and temporal evolution of the superconducting nuclei towards the final equilibrium configuration of the order parameter .
we choose a random initial configuration for the order parameter which mimics the random distribution of the superconducting nuclei expected to form in a homogeneously disordered film upon cooling it down to @xmath0 .
however , the final equilibrium island pattern does not depend on the particular choice of the initial configuration .
the advantage of this approach is that all the richness and complexity of the microscopic elasticity - superconductivity interrelations is accounted for by the phenomenological coupling constant @xmath6 , where @xmath7 is the pressure , @xmath8 is the bulk elastic modulus , and @xmath9 is the linear expansion parameter ( with @xmath10 being the linear dimension of the film ) .
these quantities can be inferred from experiments and contain all the information about the electronic degrees of freedom and the lattice excitation spectrum@xcite . in the absence of elastic interactions due to connection to a substrate
, one has the standard equilibrium solution of the conventional tdgl equation : a spatially uniform order parameter @xmath11 describing the homogeneous superconducting state at @xmath12 , and @xmath13 at @xmath14 .
the non - local elastic interaction , @xmath15 , coupling the values of the superconducting order parameter at different points @xmath16 and @xmath17 of the film [ see eq.(3 ) of methods ] distorts @xmath18 and can give rise to an instability of the uniform solution resulting in the formation of a regular island texture for certain values of the elastic parameters . to investigate this process we solve the coupled tdgl and elasticity equations by numerical integration using a quasi - spectral technique .
figure 2 displays a sequence of snapshots for the temporal development of the spatial structure of the amplitude of the order parameter .
the time is measured in the units of the ginzburg - landau time @xmath19 , with @xmath20 $ ] .
starting from the random order parameter configuration , the system evolves through three clearly distinguishable major stages : ( i ) initial amplification of small fluctuations and emergence of an amorphous structure of islands ; ( ii ) appearance of a polycrystalline configuration of well separated superconducting islands ; and ( iii ) slow relaxation of polycrystalline structure to a regular island lattice .
the evolution of the modulus of the order parameter is quantified by the correlation function @xmath21 , where @xmath22 denote the fourier components of the order parameter , and the normalization factor @xmath23 is chosen such that @xmath24 when the island texture is fully periodic .
the first stage is relatively fast : establishing of an amorphous island pattern takes only about 10@xmath19 , while achieving a polycrystalline structure requires a 10 times longer period . in this intermediate time scale @xmath25
evolves logarithmically towards the polycrystalline state . in the final stage , for @xmath26 , see fig.2 , where the island polycrystal relaxes towards the regular lattice , the correlation function of the modulus of the order parameter ceases to be an indicative characteristic quantity . at longest time - scales it is rather the temporal development of the spatial distribution of the _ phase _ of the order parameter that characterizes the evolution of the system .
for the chosen material constants the final state is a superconducting state with the uniform phase distribution corresponding to a long - ranged phase coherence .
figure 3 shows that the macroscopic phase - coherent state ( which appears as a uniformly colored frame ) is achieved via the motion and recombination of vortex - antivortex pairs which are initially present _ en mass _ in the system due to coalescence of the independent superconducting nuclei in the first stage of the order parameter evolution .
the vortex - antivortex pairs are clearly displayed by the phase - cut " lines that appear as sharp color jumps from black to white .
in the same time interval one distinguishes a well pronounced hexagon substructure which shows that different islands have different phases . the global phase coherence and therefore the superconducting state get established by the time of order @xmath27 , which is by an order of magnitude longer than the time for establishing a robust distribution of the amplitude of the order parameter
. - coupling constant , @xmath28 , plane .
above @xmath0 the film is in the normal state , whereas below @xmath0 the film can be either a homogeneous superconductor or a textured superconductor consisting of isolated islands .
the phase boundary is determined by the stability condition of the tgle ( see text ) .
the right panel shows the amplitude of the order parameter @xmath29 as function of the coupling constant @xmath28 .
the onset of the island texture ( crossing the phase boundary ) is marked by the bifurcation point where @xmath30 and @xmath31 start to diverge .
the observed hysteretic behavior the island structure forms , upon adiabatically increasing the coupling , at larger value of @xmath28 than the reverse transition , upon decreasing @xmath28 , to the spatially homogeneous order parameter state indicates that the formation of the island texture occurs via a first order transition .
the path of change of @xmath28 is indicated by the vertical arrow in the left panel .
the inset in the right panel shows the hysteretic effect corresponding sweeping temperature at fixed @xmath32 , first , from @xmath0 down to @xmath33 , and then heating up ( the process shown by the horizontal dotted line in the left panel ) .
it is noteworthy that the island structure persists even upon heating the system above @xmath0 of the film material , due to the increased local @xmath0 within the islands .
the dotted line indicates the usual @xmath34 behavior of the order parameter without elastic interactions . ]
it is important that , while being a general phenomenon , the formation of an island texture requires the elastic coupling to be strong enough .
namely , the texture appears under the condition that the coupling constant @xmath28 exceeds some temperature dependent critical value @xmath35 .
the function @xmath35 is determined by means of linear stability analysis of the tdgl .
writing down the fourier transform of the elastic interaction term @xmath36 as @xmath37 , where @xmath38 is the scale free elastic kernel in the momentum space and @xmath39 is the diffusion coefficient in the ginzburg - landau equation ( see method ) , one finds that the island state forms if @xmath40 , where @xmath41 is the maximum value of @xmath38 at @xmath42 .
this gives @xmath43 . the lower bound for the coupling constant at which the homogeneous state becomes unstable and the island texture can appear is @xmath44 .
the left panel of the figure 4 shows the phase diagram in the reduced @xmath45-@xmath46 coordinates for the chosen system parameters which are given in the methods .
the right panel of fig.4 shows the behavior of the amplitude of the order parameter @xmath47 at the isotherm @xmath48 as function of the coupling constant @xmath28 . the protocol for finding this dependence is as follows : we start with a coupling constant @xmath49 and an initial random distribution of the order parameter , and let the system evolve until a stationary distribution of @xmath29 is achieved .
after that @xmath28 is increased and the system evolves again until a new stationary state is achieved ( adiabatic increase ) , and so forth . upon crossing the phase boundary a bifurcation in the @xmath50-dependence corresponding to the islands formation occurs and a _ finite _ difference between the maximal and minimal values of the amplitude of the order parameter appears . in other words
@xmath50 shows finite jumps to the @xmath51 and @xmath52 branches at some @xmath28 appreciably beyond the phase boundary .
on the descending part of the cycle , @xmath51 and @xmath52 merge at smaller value of @xmath28 closer to the phase boundary , i.e. at smaller @xmath28 .
the change of @xmath28 is indicated by the vertical arrow in the left panel of fig.4 . shown in the right panel inset
is the hysteretic behavior of the order parameter corresponding to sweeping the temperature across the phase transition line .
the protocol is nearly the same as for the coupling constant variation scheme ; the temperature is first decreased from @xmath53 down to @xmath48 and then increased back at the fixed coupling constant @xmath32 .
the temperature sweeping is indicated by the horizontal arrow in the left panel of fig .
4 .
note that the island structure persists even as the system is heated above @xmath0 ; this is the manifestation of the positive feedback between the superconductivity and elasticity in this films bonded to the substrate giving rise to a local increase in superconducting transition temperature within the islands .
the observed hysteretic behaviors indicate that the transition between the homogeneous superconducting and the island texture states is of the first order .
the pattern formation and , in particular , the spontaneous emergence of electronic nanometer - scale structures , due to the existence of competing states is ubiquitous in nature and is found in a wealth of complex systems and physical phenomena ranging from magnetism , superconductivity and superfluidity to liquid crystals and biological dynamics@xcite .
the ginzburg - landau equation , one of the most universal models in theoretical physics , offers an adequate quantitative description of the pattern formation .
here we have revealed the instability and emergence of an electronic textures , due to coupling of the superconducting order parameter with elastic stress fields , generating a non - local term in the ginzburg - landau functional .
it is noteworthy that the general form of the nonlocality in the ginzburg - landau equation , induced by the interplay between the local nonlinearity and the nonlocal couplings , is universal and can result from various physical mechanisms . in particular , while the results presented here are stable with respect to weak disorder , we expect that the interplay of strong disorder with coulomb interactions will give an essential contribution to the nonlocal coupling in the ginzburg - landau functional .
note , in this connection the interesting discussion of the effects of disorder leading to a spatially inhomogeneous distribution of the order parameter@xcite .
we thus conclude that a consistent consideration of the superconductivity in thin films has to include the effects of the interplay between strong disorder with coulomb interactions and elasticity .
the observed island texture is an array of granules of a good " superconductor immersed into a sea of a bad " one with much smaller amplitude of the order parameter and is equivalent to an array of small superconducting droplets connected via weak links , i.e. , a josephson junction array ( jja ) . in our simulations , where thermal and quantum fluctuations are absent , the josephson coupling
, however weak , eventually establishes a global phase coherence and thus global superconductivity .
the fluctuations , dissipation , and disorder can break down this phase uniformity : the intermediate stages of phase distributions of our simulations represent the final states of the``real world '' systems .
thus the possible eventual configuration of the order parameter phase ( see the `` gaudi - mosaic '' phase distribution in the inset of the last phase - frame of the fig .
3 ) can correspond even the insulating state .
this is in accord with the general result that jjas turn insulators if the josephson coupling between the superconducting islands becomes sufficiently weak@xcite .
thus our finding that the transition into a superconducting state in two dimensional superconductors elastically tied to a substrate can occur via the formation of nanoscale superconducting island pattern offers a possible scenario for the superconductor - to - insulator transition observed in various thin superconducting films@xcite .
this supports the conjecture@xcite that the formation of a regular array of superconducting droplets is an inherent property of the critical region of the superconductor - to - insulator transition .
we thank yuri galperin and david hinks for useful discussions .
this work was supported by the u.s .
department of energy office of science under the contract no .
de - ac02 - 06ch11357 , by the program `` quantum macrophysics '' of the russian academy of sciences , and by the russian foundation for basic research ( grant nos .
09 - 02 - 01205 and 09 - 0212206 ) . |
we want to compute the jacobi symbol , since this is easier to typeset and less ambiguous than the more usual @xmath3 .
@xmath4 is the time to multiply @xmath0-bit numbers .
@xmath5 means @xmath6 for some constant @xmath7 . ]
@xmath8 for @xmath0-bit integers @xmath9 and @xmath10 , where @xmath9 is odd positive .
we give three algorithms based on the @xmath11-adic gcd from stehl and zimmermann @xcite .
first we give an algorithm whose worst - case time bound is @xmath12 ; we call this the _ cubic _ algorithm although this is pessimistic since the algorithm is quadratic on average as shown in @xcite , and probably also in the worst case .
we then show how to reduce the worst - case to @xmath13 by combining sequences of `` ugly '' iterations ( defined in section [ sec : cubic ] ) into one `` harmless '' iteration .
finally , we obtain an algorithm with worst - case time @xmath1 .
this is , up to a constant factor , the same as the time bound for the best known algorithm , apparently never published in full , but sketched in bach @xcite and in more detail in bach and shallit @xcite ( with credit to bachmann @xcite ) .
the latter algorithm makes use of the knuth - schnhage fast continued fraction algorithm @xcite and an identity of gauss @xcite .
although this algorithm has been attributed to schnhage , schnhage himself gives a different @xmath1 algorithm @xcite which does not depend on the identity of gauss .
the algorithm is mentioned in schnhage s book @xcite , but no details are given there . with our algorithm
it is not necessary to compute the full continued fraction or to use the identity of gauss for the jacobi symbol .
thus , it provides an alternative that may be easier to implement .
it is possible to modify some of the other fast gcd algorithms considered by mller @xcite to compute the jacobi symbol , but we do not consider such possibilities here . at best
they give a small constant factor speedup over our algorithm .
we recall the main identities satisfied by the jacobi symbol : @xmath14 ; @xmath15 ; @xmath16 for @xmath17 odd ; and @xmath18 if @xmath19 .
note that all our algorithms compute @xmath8 with @xmath10 even positive and @xmath9 odd positive .
for the more general case where @xmath10 is any integer , we can reduce to @xmath10 even and positive using @xmath20 if @xmath10 is negative , and @xmath21 if @xmath10 is odd .
we first describe a cubic algorithm to compute the jacobi symbol .
the quadratic algorithm in section [ sec : harmless - iterations ] is based on this cubic algorithm , and the subquadratic algorithm in section [ sec : sub - quadratic ] uses the same ideas as the quadratic algorithm but with an asymptotically fast recursive implementation . for @xmath22 , the notation @xmath23 denotes the @xmath11-adic valuation @xmath24 of @xmath9 , that is the maximum @xmath25 such that @xmath26 , or @xmath27 if @xmath28 . throughout the paper we use the binary division with positive quotient defined by algorithm [ binarydividepos ] .
compared to the `` centered division '' of @xcite , it returns a quotient in @xmath29 $ ] instead of in @xmath30 $ ] . note that the quotient @xmath31 is always odd .
@xmath32 with @xmath33 @xmath31 and @xmath34 such that @xmath35 , @xmath36 @xmath37 [ bdp : step1 ] * return * @xmath38 . with this binary division ,
we define algorithm cubicbinaryjacobi , where the fact that the quotient @xmath31 is positive ensures that all @xmath39 terms computed remain positive , and @xmath9 remains odd , thus @xmath8 remains well - defined . , the _ _ [ binary gcd ] _ algorithm no longer terminates _ ''
. however , with a modified stopping criterion as in algorithm cubicbinaryjacobi , the algorithm terminates ( we prove this below ) . ] @xmath32 with @xmath40 jacobi symbol @xmath8 @xmath41 , @xmath42 @xmath43 @xmath44 [ cbj : step5 ] @xmath45 [ cbj : step6 ] @xmath46 , @xmath42 @xmath47 * then * return @xmath48 * else * return @xmath49 algorithm cubicbinaryjacobi is correct ( assuming it terminates ) . proof .
we prove that the following invariant holds during the algorithm , if @xmath50 are the initial values of @xmath17 : @xmath51 this is true before we enter the while - loop , since @xmath52 , @xmath53 , and @xmath54 . for each step in the while loop ,
we divide @xmath10 by @xmath55 , swap @xmath9 and @xmath56 , replace @xmath9 by @xmath57 , and divide @xmath58 by @xmath55 .
the jacobi symbol is modified by a factor @xmath59 for the division of @xmath10 by @xmath55 , by a factor @xmath60 for the interchange of @xmath9 and @xmath61 , and by a factor @xmath62 for the division of @xmath58 by @xmath55 . at the end of the loop ,
we have @xmath63 ; if @xmath64 , since @xmath65 , we have @xmath66 , otherwise @xmath67 .
the quantity @xmath68 is non - increasing in algorithm cubicbinaryjacobi . proof . at each iteration of the `` while '' loop
, @xmath9 becomes @xmath69 , and @xmath10 becomes @xmath70 . in matrix notation @xmath71
therefore @xmath72 becomes @xmath73 since @xmath74 , the first term is bounded by @xmath9 . in the second term , @xmath75 , thus the second term is bounded by @xmath76 , which is bounded by @xmath77 for @xmath78 , and equals @xmath79 for @xmath80 .
if @xmath78 , then @xmath72 is multiplied by a factor at most @xmath81 . if @xmath82 then @xmath72 decreases , but by a factor which could be arbitrarily close to @xmath83 .
the only case where @xmath72 does not decrease is when @xmath80 and @xmath84 ; in this case @xmath72 is unchanged .
this motivates us to define three classes of iterations : _ good _ , _ bad _ , and _
ugly_. let us say that we have a _ good _ iteration when @xmath78 , a _
iteration when @xmath82 , and an _ ugly _ iteration when @xmath80 and @xmath84 .
since @xmath31 is odd and @xmath85 , this covers all possibilities . for a bad iteration ,
@xmath86 becomes @xmath87 , and for an ugly iteration , @xmath86 becomes @xmath88 .
we denote the matrices corresponding to good , bad and ugly iterations by @xmath89 , @xmath90 and @xmath91 respectively .
thus @xmath92 the effect of @xmath93 successive ugly iterations is easily seen to be given by the matrix @xmath94 2 - 2(-1/4)^m & \ ; 4 + ( -1/4)^m \end{array } \right ) .
\label{eq : um}\ ] ] assume we start from @xmath95 , and after @xmath96 successive ugly iterations we get values @xmath97 .
then , from equation ( [ eq : um ] ) , @xmath98 we can not have @xmath99 or the algorithm would have terminated .
however , @xmath100 must be an integer .
this gives an upper bound on @xmath93 . for @xmath101 of @xmath0 bits ,
the number of successive ugly iterations is bounded by @xmath102 ( a precise statement is made in lemma [ lemma2 ] ) .
if there were no bad iterations , this would prove that for @xmath0-bit inputs the number of iterations is @xmath103 , since each sequence of ugly iterations would be followed by at least one good iteration .
bad iterations can be handled by a more complicated argument which we omit , since they will be considered in detail in
[ sec : harmless - iterations ] when we discuss the complexity of the quadratic algorithm ( see the proof of theorem [ thm : qbj ] ) .
since the number of iterations is @xmath103 from theorem [ thm : qbj ] , and each iteration costs time @xmath104 , the overall time for algorithm cubicbinaryjacobi is @xmath105 . note that this worst - case bound is almost certainly too pessimistic ( see [ results ] ) .
suppose we have a sequence of @xmath106 ugly iterations .
it is possible to combine the @xmath93 ugly iterations into one _ harmless _ iteration which is not much more expensive than a normal ( good or bad ) iteration .
also , it is possible to predict the maximal such @xmath93 in advance . using this trick , we reduce the number of iterations ( good , bad and harmless ) to @xmath107 and their cost to @xmath13 . without loss of generality ,
suppose that we start from @xmath108 = @xmath86 .
since @xmath9 is odd , we never have @xmath109 .
[ lemma2 ] if @xmath110 , then we have exactly @xmath111 ugly iterations starting from @xmath112 , followed by a good iteration if @xmath113 is even , and by a bad iteration if @xmath113 is odd .
we prove the lemma by induction on @xmath113 . if @xmath114 , @xmath115 is odd , but @xmath9 is odd , so @xmath116 is even , which yields @xmath78 in binarydividepos , thus @xmath39 yield a good iteration . if @xmath117
, @xmath115 is even , which implies that @xmath116 is odd , thus we have @xmath80 .
if we had @xmath84 in binarydividepos , this would mean that @xmath118 , or equivalently @xmath119 , which is incompatible with @xmath117 .
thus we have @xmath120 , and a bad iteration .
now assume @xmath121 .
the first iteration is ugly since @xmath122 divides @xmath115 , which implies that @xmath116 is odd .
thus @xmath80 , and implies that @xmath84 .
after one ugly iteration @xmath86 becomes @xmath123 , thus @xmath115 becomes @xmath124 , and the @xmath11-valuation of @xmath115 decreases by @xmath11 . from the above
, we see that , for a sequence of @xmath93 ugly iterations , @xmath125 satisfy the three - term recurrence @xmath126 and similarly for @xmath127 .
it follows that @xmath128 , and similarly @xmath129 , for @xmath130 .
we can modify algorithm cubicbinaryjacobi to consolidate @xmath93 consecutive ugly iterations into one harmless iteration , using the expressions ( [ eq : am])([eq : bm ] ) for @xmath100 and @xmath131 ( we give an optimised evaluation below ) .
it remains to modify step [ cbj : step6 ] of cubicbinaryjacobi to take account of the @xmath93 updates to @xmath132 .
since @xmath80 for each ugly iteration , we have to increment @xmath132 by an amount @xmath133 where we write @xmath134 for @xmath135 .
however , @xmath136 for @xmath137 , so the terms involving division by @xmath138 `` collapse '' mod @xmath11 , leaving just the first and last terms .
the terms involving two divisions by @xmath11 are all equal to @xmath139 mod @xmath11 , using the observation that @xmath140 mod @xmath122 is constant for @xmath141 .
thus @xmath142 one further simplification is possible .
since @xmath143 , and @xmath144 is odd , we can replace @xmath145 by @xmath144 in the last term , and use the fact that @xmath146 to obtain @xmath147 we can economise the computation of @xmath100 and @xmath131 from ( [ eq : am])([eq : bm ] ) by first computing @xmath148 where the divisions by @xmath149 and by @xmath150 are exact ; then @xmath151 , @xmath152 . from these observations , it is easy to modify algorithm cubicbinaryjacobi to obtain algorithm quadraticbinaryjacobi . in this algorithm , steps [ qbj : step8][qbj : step12 ] implement a harmless iteration equivalent to @xmath106 consecutive ugly iterations ; steps [ qbj : step14][qbj : step15 ] implement bad and good iterations , and the remaining steps are common to both .
step [ cbj : step6 ] of algorithm cubicbinaryjacobi is split into three steps [ qbj : step5 ] , [ qbj : step14 ] and [ qbj : step16 ] . in the case of a harmless iteration ,
the computation of @xmath153 satisfying ( [ eq : delta ] ) is implicit in steps [ qbj : step5 ] , [ qbj : step11 ] and [ qbj : step16 ] .
@xmath32 with @xmath40 jacobi symbol @xmath8 @xmath41 , @xmath42 [ qbj : step3 ] @xmath43 @xmath154 [ qbj : step5 ] @xmath44 @xmath155 [ qbj : step8 ] @xmath156 @xmath157 @xmath158 [ qbj : step11 ] @xmath159 [ qbj : step12 ] @xmath160 [ qbj : step14 ] @xmath46 [ qbj : step15 ] @xmath154 , [ qbj : step16 ] @xmath42 [ qbj : step17 ] @xmath47 * then * return @xmath48 * else * return @xmath49 [ thm : qbj ] algorithm quadraticbinaryjacobi is correct and terminates after @xmath107 iterations of the `` while '' loop ( steps [ qbj : step3][qbj : step17 ] ) if the inputs are positive integers of at most @xmath0 bits , with @xmath161 .
correctness follows from the equivalence to algorithm cubicbinaryjacobi . to prove that convergence takes @xmath107 iterations ,
we show that @xmath72 is multiplied by a factor at most @xmath162 in each block of three iterations .
this is true if the block includes at least one good iteration , so we need only consider harmless and bad iterations .
two harmless iterations do not occur in succession , so the block must include either ( harmless , bad ) or ( bad , bad ) . in the first case , the corresponding matrix is @xmath163 for some @xmath96 .
we saw in
[ sec : cubic ] that the matrix @xmath91 leaves @xmath72 unchanged , so @xmath164 also leaves @xmath72 unchanged , and we need only consider the effect of @xmath165 .
suppose that @xmath112 is transformed into @xmath166 by @xmath165 .
thus @xmath167 we see that @xmath168 the case of two successive bad iterations is similar just replace @xmath165 by @xmath169 in the above , and deduce that @xmath170 .
we conclude that the number of iterations of the while loop is at most @xmath171 , where @xmath172 .
_ remarks _ + 1 .
a more complicated argument along similar lines can reduce the constant @xmath173 to @xmath174 . here
@xmath175 denotes the spectral radius : @xmath176 .
+ 2 . in practice quadraticbinaryjacobi is not much ( if any ) faster than cubicbinaryjacobi .
its advantage is simply the better worst - case time bound .
a heuristic argument suggests that on average only @xmath177 of the iterations of cubicbinaryjacobi are ugly .
our implementations of cubicbinaryjacobi and quadraticbinaryjacobi are slower than gmp s @xmath103 algorithm ( which is based on stein s binary gcd , as in shallit and sorenson @xcite ) . however , in the next section we use the ideas of our quadraticbinaryjacobi algorithm to get an @xmath1 algorithm .
we do not see how to modify the algorithm of shallit and sorenson to do this .
( @xmath178 in our notation ) to decide whether to perform an interchange .
this makes it difficult to construct a recursive @xmath1 algorithm like algorithm halfbinaryjacobi . ]
algorithm halfbinaryjacobi is a modification of algorithm half - gb - gcd from @xcite .
the main differences are the following : 1 .
binary division with positive ( not centered ) quotient is used ; 2 .
the algorithm returns an integer @xmath132 such that if @xmath17 are the inputs , @xmath180 the output values defined by theorem [ theorem2 ] , then @xmath181 3 . at steps [ hbgcd : step4 ] and
[ hbgcd : step27 ] , we reduce mod @xmath182 ( resp .
@xmath183 ) instead of mod @xmath184 ( resp .
@xmath185 ) , so that we have enough information to correctly update @xmath186 at steps [ hbgcd : step10 ] , [ hbgcd : step17 ] , [ hbgcd : step21 ] and [ hbgcd : step25 ] ; 4 .
we have to `` cut '' some harmless iterations in two ( step [ hbgcd : stepm ] ) .
@xmath187 with @xmath161 , and @xmath188 two integers @xmath189 and a @xmath190 matrix @xmath191 return @xmath192 [ hbgcd : step2 ] @xmath193 @xmath194 , @xmath195 [ hbgcd : step4 ] @xmath196 [ hbgcd : step5 ] @xmath197 , @xmath198 [ hbgcd : step6 ] @xmath199 return @xmath200 [ hbgcd : step9 ] @xmath201 [ hbgcd : step10 ] @xmath202 @xmath203 [ hbgcd : step12 ] [ hbgcd : step13 ] @xmath204 @xmath205 [ hbgcd : stepm ] @xmath157 @xmath206 [ hbgcd : step17 ] @xmath207 [ hbgcd : stepa2 ] @xmath208 [ hbgcd : step19 ] [ hbgcd : step20 ] @xmath209 [ hbgcd : step21 ] @xmath210 @xmath211 [ hbgcd : step23 ] @xmath212 @xmath213 [ hbgcd : step25 ] @xmath214 [ hbgcd : step11 ] @xmath215 [ hbgcd : step27 ] return @xmath216 , @xmath217 [ [ remarks . ] ] remarks .
+ + + + + + + + the matrix @xmath218 occurring at step [ hbgcd : step19 ] is just @xmath219 , where @xmath220 is given by equation ( [ eq : um ] ) .
similarly , the matrix @xmath218 occurring at step [ hbgcd : step23 ] is @xmath221 . in practice ,
steps [ hbgcd : step13][hbgcd : step20 ] can be omitted ( so the algorithm becomes a fast version of cubicbinaryjacobi ) this variant is simpler and slightly faster on average .
[ theorem2 ] let @xmath222 be the inputs of algorithm halfbinaryjacobi , and @xmath223 the corresponding outputs . if @xmath224 , then : @xmath225 proof ( outline ) .
we prove the theorem by induction on the parameter @xmath25 .
the key ingredient is that if we reduce @xmath17 mod @xmath184 in step [ hbgcd : step4 ] , then the gb sequence of @xmath226 matches that of @xmath17 , for the terms computed by the recursive call at step [ hbgcd : step5 ] .
this is a consequence of ( * ? ? ?
* lemma 7 ) ( which also holds for binary division with positive quotient ) .
it follows that in all the binary divisions with inputs @xmath227 in that recursive call , @xmath140 and @xmath228 match modulo @xmath229 the corresponding values that would be obtained from the full inputs @xmath17 ( otherwise the corresponding binary quotient @xmath230 would be wrong ) . since here we reduce @xmath17 mod @xmath182 instead of mod @xmath184 , @xmath140 and @xmath228 now match modulo @xmath231 instead of modulo @xmath229 the values that would be obtained from the full inputs @xmath17 , where @xmath232 since @xmath233 . at step [ hbgcd : step10 ]
, @xmath186 depends only on @xmath234 and @xmath235 , at step [ hbgcd : step17 ] it depends on @xmath236 and @xmath237 , and at step [ hbgcd : step21 ] on @xmath237 and @xmath238 . since @xmath239 and @xmath240 at step [ hbgcd : step21 ]
correspond to some @xmath140 and @xmath228 , it follows that @xmath239 and @xmath240 agree mod @xmath138 with the values that would be computed from the full inputs , and thus the correction @xmath186 is correct .
this proves by induction that @xmath241 .
now we prove that @xmath242 .
if there is no harmless iteration , @xmath242 is a consequence of the proof of theorem 1 in @xcite . in case
there is a harmless iteration , first assume that @xmath243 at step [ hbgcd : stepm ] .
the new values @xmath244 at step [ hbgcd : stepa2 ] correspond to @xmath93 successive ugly iterations , which yield @xmath245 .
thus @xmath246 : we did not go too far , and since we are computing the same sequence of quotients as algorithm quadraticbinaryjacobi , the result follows .
now if @xmath247 , we would go too far if we performed @xmath248 ugly iterations , since it would give @xmath249 , thus @xmath250 , and @xmath251 would exceed @xmath25 .
this is the reason why we `` cut '' the harmless iteration at @xmath252 ( step [ hbgcd : stepm ] ) .
the other invariants are unchanged
. finally we can present our @xmath1 algorithm fastbinaryjacobi , which computes the jacobi symbol by calling algorithm halfbinaryjacobi .
the general structure is similar to that described in @xcite for several asymptotically fast gcd algorithms .
@xmath32 with @xmath161 jacobi symbol @xmath8 @xmath41 , @xmath42 @xmath253 @xmath254 @xmath255 @xmath256 , @xmath42 @xmath47 * then * return @xmath48 * else * return @xmath49 daireaux , maume - deschamps and valle @xcite prove that , for the positive binary division , the average increase of the most significant bits is @xmath257 bits / iteration ( which partly cancels an average decrease of two least significant bits per iteration ) ; compare this with only @xmath258 bits / iteration on average for the centered division . and
@xmath259 respectively . ]
we have implemented the different algorithms in c ( using @xmath260-bit integers ) and in gmp ( using multiple - precision integers ) , as well as in maple / magma ( for testing purposes ) .
for @xmath261 the maximum number of iterations of algorithm cubicbinaryjacobi is @xmath260 , with @xmath262 and @xmath263 .
the number of iterations seems to be @xmath107 for @xmath264 : see table 1 .
this is plausible because , from heuristic probabilistic arguments , we expect about half of the iterations to be good , and experiments confirm this .
for example , if we consider all admissible @xmath265 , the cumulated number of iterations is @xmath266 for @xmath267 calls , i.e. , an average of @xmath268 iterations per call ( max @xmath269 ) ; the cumulated number of good , bad and ugly iterations is @xmath270% , @xmath271% , and @xmath272% respectively .
for @xmath273 , a random sample of @xmath274 pairs @xmath86 gave @xmath275 iterations per call ( max @xmath276 ) , with @xmath277 , @xmath278% , and @xmath279 for good , bad and ugly respectively .
these ratios seem to be converging to the heuristically expected @xmath280 , @xmath281 , and @xmath281 .
when we consider all admissible @xmath265 , the maximum number of iterations of quadraticbinaryjacobi is @xmath282 when @xmath283 , @xmath284 , the cumulated number of iterations is @xmath285 ( @xmath286 per call ) , the cumulated number of good , bad and harmless iterations is @xmath287% , @xmath288% , and @xmath289% respectively . for @xmath273 , a random sample of @xmath274 pairs @xmath86 gave @xmath290 iterations per call ( max @xmath291 ) , with @xmath292 , @xmath293% , and @xmath294 for good , bad and harmless respectively .
these ratios seem to be converging to the heuristically expected @xmath295 , @xmath296 , and @xmath297 .
we have also compared the time and average number of iterations for huge numbers , using the fast gcd algorithm in gmp , say ` gcd ` which implements the algorithm from @xcite and an implementation of the algorithm from @xcite , say ` bgcd ` . for inputs of one million @xmath260-bit words , ` gcd ` takes about 45.8s on a 2.83ghz core 2 , while ` bgcd ` takes about 48.3s and 32,800,000 iterations : this is in accordance with the fact proven in @xcite that each step of the binary gcd discards on average two least significant bits , and adds on average about 0.05 most significant bits .
our algorithm ` bjacobi ` ( based on algorithms 3.13.2 ) takes about 83.1s and 47,500,000 iterations ( for a version with steps [ hbgcd : step13][hbgcd : step20 ] of algorithm 3.1 omitted in the basecase routine ) , which agrees with the theoretical drift of @xmath298 bits per iteration . the break - even point between the @xmath103 implementation of the jacobi symbol in gmp 4.3.1 and
our @xmath179 implementation is about @xmath299 words , that is about @xmath300 bits or about @xmath301 decimal digits ( see fig .
[ fig1 ] ) .
weilert @xcite says : `` _ _ we are not able to use a gcd calculation in @xmath302 $ ] similar to the binary gcd algorithm @xmath303 because we do not get a corresponding quotient sequence in an obvious manner _ _ '' . in a sense we filled that gap for the computation of the jacobi symbol , because we showed how it can be computed using a binary gcd algorithm without the need for a quotient sequence .
we showed how to compute the jacobi symbol with an asymptotically fast time bound , using a binary gcd algorithm without the need for a quotient sequence .
our implementation is faster than a good @xmath103 implementation for numbers with bitsize @xmath304 .
our subquadratic implementation is available from http://www.loria.fr/~zimmerma/software/#jacobi .
binary division with a centered quotient does not seem to give a subquadratic algorithm ; however we can use it with the `` cubic '' algorithm ( which then becomes provably quadratic ) since then we control the sign of @xmath17 . for a better quadratic algorithm
, we can choose the quotient @xmath31 so that @xmath305 , by replacing @xmath31 by @xmath306 if necessary : experimentally , this gains on average @xmath307 bits per iteration , compared to @xmath308 for the centered quotient , and @xmath309 for the positive quotient . in comparison ,
stein s `` binary '' algorithm gains on average @xmath310 bits per iteration @xcite@xcite .
the authors thank steven galbraith who asked them about the existence of an @xmath179 algorithm for the jacobi symbol , arnold schnhage for his comments and a pointer to the work of his former student andr weilert , damien stehl who suggested adapting the binary gcd algorithm , and marco bodrato and niels mller for testing our implementation .
we also thank inria for its support of the anc `` quipe associe '' .
the first author acknowledges the support of the australian research council .
richard p. brent , twenty years analysis of the binary euclidean algorithm , in _ millennial perspectives in computer science : proceedings of the 1999 oxford microsoft symposium in honour of professor sir antony hoare _
( edited by j. davies , a. w. roscoe and j. woodcock ) , palgrave , new york , 2000 , 4153 .
http://wwwmaths.anu.edu.au/~brent/pub/pub183.html benot daireaux , vronique maume - deschamps and brigitte valle , the lyapunov tortoise and the dyadic hare , _ proceedings of the 2005 international conference on analysis of algorithms _ , dmtcs proc .
ad ( 2005 ) , 7194 . | the best known algorithm to compute the jacobi symbol of two @xmath0-bit integers runs in time @xmath1 , using schnhage s fast continued fraction algorithm combined with an identity due to gauss .
we give a different @xmath1 algorithm based on the binary recursive gcd algorithm of stehl and zimmermann . our implementation which to our knowledge is the first to run in time @xmath1
is faster than gmp s quadratic implementation for inputs larger than about @xmath2 decimal digits . |
in this supplementary material , we detail the proof of observation 4 of the main text .
we again label in the stinespring dilation @xmath99 of the quantum channel @xmath1 the input by @xmath3 , the output by @xmath5 and the environment by @xmath6 , but also keep a system @xmath18 holding a purification of the input .
we have to show that if @xmath100 , then also @xmath101 if @xmath100 indeed holds true , then we find for the classical and quantum entanglement - assisted capacities @xmath102 and @xmath103 , respectively , @xmath104 we also know that @xmath105 where the maximum is taken over pure states @xmath106 .
one way of expressing the right hand side , so the mutual information , in our notation of input @xmath3 , output @xmath5 , environment @xmath6 , and purification @xmath18 is @xmath107 where @xmath36 is a mixed state that is shared between @xmath5 and @xmath18 that is obtained as @xmath108 . from the subaddivity of the von - neumann entropy
, it follows that for the input state that achieves the maximum in eq .
( [ b ] ) , @xmath109 since at the same time , by definition @xmath110 , we find that there exists an input pure state @xmath111 shared between @xmath18 , @xmath5 , and @xmath6 , with @xmath112 , @xmath113 being shared between @xmath18 and @xmath3 , such that @xmath114 the latter equality also implies that @xmath115 which in turn means @xmath116 . in the final step ,
it is the aim to construct an input to the channel that certifies that the single shot quantum capacity @xmath117 .
based on the above properties of the channel , this can easily be done . by virtue of the schmidt decomposition
, there exists a unitary @xmath118 supported on @xmath18 such that @xmath119 where for convenience the tensor factors have been indicated .
so the input @xmath120 will give rise to the output as in eq .
( [ out ] ) . from this input to the original problem we can construct a new input which achieves the desired bound : take @xmath121 yielding the output @xmath122 which is a tensor product between @xmath18 and @xmath6 .
this means that @xmath123 , which is that was to be shown . | entangled inputs can enhance the capacity of quantum channels , this being one of the consequences of the celebrated result showing the non - additivity of several quantities relevant for quantum information science . in this work ,
we answer the converse question ( whether entangled inputs can ever render noisy quantum channels have maximum capacity ) to the negative : no sophisticated entangled input of any quantum channel can ever enhance the capacity to the maximum possible value ; a result that holds true for all channels both for the classical as well as the quantum capacity .
this result can hence be seen as a bound as to how `` non - additive quantum information can be '' . as a main result
, we find first practical and remarkably simple computable single - shot bounds to capacities , related to entanglement measures . as examples ,
we discuss the qubit amplitude damping and identify the first meaningful bound for its classical capacity .
how much information can one transmit reliably through a quantum channel such as a tele - communication fiber ?
this basic question is , despite much progress in recent years @xcite , still surprisingly wide open .
some suitable encoding and decoding is necessary , needless to say , but the optimal achievable rates can still not be expressed in a computable closed form . for classical information ,
the hope that the single - shot capacity would be sufficient to arrive at that goal was corroborated by many examples of channels for which this is in fact true @xcite .
alas , it was finally found to be unjustified with the celebrated result @xcite on the non - addivity of several quantities that are in the center of interest in quantum information science @xcite .
in particular , entangled inputs help and do increase the classical information capacity .
this result showed that the question of finding capacities of quantum channels is more complicated than what one might have anticipated . in the case of quantum information transmission
, a similar situation has been known to be true already for a long time : in general one must regularize the single - shot expression , given by the coherent information , in order to attain the quantum capacity @xcite . to contribute to fixing the coordinate system of channel capacities , this insight begs for a resolution of the following question : to what extent can entanglement help then ? is the mentioned result rather an academic observation , manifesting itself in small violations of additivity in high physical dimensions ?
an interesting question in this context is the following : _ can suitably entangled inputs render noisy quantum channels take their maximum possible capacity or make them even perfect ? _ this would be the other extreme , where the non - additivity serves as a resource to overcome the noisiness of channels . in this work
, we answer this question to the negative : for all quantum channels , no matter how elaborate the entangled coding over many uses of the channel might be , one can never achieve the maximum possible capacity if this is not already true on the single - shot level .
this observation holds true both for the classical as well as the quantum capacity .
we show this by introducing new upper bounds to these capacities which can be evaluated on the single - shot level and are computable , which constitute a main result of this work .
we connect questions of capacities to those of entanglement measures of systems and their environments .
these bounds are useful in their own right , which will be shown by means of an example of amplitude - damping channel .
_ notation and setting .
_ we start our discussion by fixing the notation and clarifying some basic concepts that will be used later on .
we consider general quantum channels of arbitrary finite dimension , @xmath0 , modeling any general noisy quantum evolution .
@xmath1 is hence an arbitrary trace - preserving completely positive map .
such a channel can always be written in terms of a _ stinespring dilation _ as @xmath2labeling the input by @xmath3 , associated with the hilbert space @xmath4 , the output by @xmath5 and the environment by @xmath6 , equipped with hilbert spaces @xmath4 and @xmath7 , respectively .
@xmath8 is an isometry mapping the input on @xmath3 onto a quantum state on @xmath5 and @xmath6 .
the classical information capacity , or short _ classical capacity _ , of a quantum channel is the rate at which one can reliably send classical information .
it is related to the _ holevo-@xmath9 _ @xcite or the _ single - shot classical capacity _ of that channel ,
@xmath10where the maximum is taken over probability distributions and states , as the asymptotic regularization @xmath11 the trivial maximum value that the capacity can possibly take is given by the _
maximum output entropy _ of the quantum channel , @xmath12we will say that whenever this bound is saturated , so when @xmath13 , that the channel has _ maximum capacity _ , giving rise to the maximum that is trivially possible .
of course , this notion includes the situation of a _ perfect quantum channel _ that has maximum output entropy of @xmath14 .
the _ quantum capacity _ of a quantum channel , in turn , is related to the rate at which one can reliably send quantum information through a quantum channel .
writing @xmath15 calculated in the state @xmath16 , @xmath17 , as a state on @xmath18 , @xmath5 , @xmath6 , and where @xmath19 , with @xmath20 being again the isometry of @xmath1 , mapping @xmath3 to @xmath5 and @xmath6 , and @xmath21 being a state vector on @xmath18 and @xmath3 .
the quantum capacity is then @xmath22 again , referred to as _ maximum _ if @xmath23 .
_ main result .
_ we can now formulate the main result .
[ perfect ] every quantum channel that is noisy ( in the sense that the single - shot classical capacity is not the maximum output entropy ) can not be made having maximum capacity under the help of any sophisticated entangled input .
so if there is a gap to the maximum possible single - shot capacity , this gap will be preserved in the asymptotic limit , independent of @xmath24 : no entangled input can overcome this limitation .
the single - shot classical capacity may be non - additive , as has been shown in ref .
@xcite . yet
, entanglement can only help to some extent , and can , in particular , not make any imperfect channel perfect .
_ upper bounds for classical capacities .
_ in order to show this result and the equivalent one for the quantum capacity , we make use of upper bounds to channel capacities , starting with the classical capacity . the bounds forming the tools of the argument will be provided by quantities that capture the entanglement between a system and its environment in a dilation of the channel .
we first show what properties a general quantity @xmath25 , defined on bipartite quantum systems , should have . in order to be entirely clear
, we will always give the tensor factors with respect to which an entanglement measure will be taken .
for example , @xmath26 would be the quantity evaluated for @xmath27 with respect to the split @xmath28 .
two properties will be important : 1 .
@xmath29 has the property that @xmath30for every bipartite state @xmath27 defined on @xmath24 copies of a @xmath31-dimensional quantum system , labelled @xmath32 and @xmath33 , @xmath34 denoting the respective reduction .
2 . @xmath29 is _
faithful_. that is , @xmath35 for bipartite states @xmath36 on @xmath3 and @xmath5 if and only if @xmath36 is entangled with respect to this split . here
, @xmath37 denotes the _ entanglement of formation _ @xcite .
as it turns out , for any quantity satisfying property 1 , the following bound holds true : [ cc ] for any quantum channel @xmath1 and any quantity @xmath29 that satisfies the condition 1 .
we find the single - shot upper bound @xmath38 the argument leading to this bound is remarkably simple : starting from eq .
( [ cc ] ) , and defining @xmath39 with reductions @xmath40 , @xmath5 being formed by @xmath32 and @xmath6 by @xmath41 , we find , using the msw - correspondence @xcite , @xmath42using subadditivity , and hence , using property 1 , @xmath43which is the above single - shot bound of observation [ cc ] .
this bound is to be compared with the msw expression @xcite for the holevo-@xmath9 itself , @xmath44this is very similar , except that now the entanglement of formation takes the role of the quantity @xmath29 .
this indeed leads also to the conclusion of observation [ perfect ] for the classical capacity : @xmath45 achieves the maximum upper bound @xmath46 if and only if @xmath47 achieves it .
this is because @xmath9 achieves it if and only if @xmath48for the maximizing @xmath36 in @xmath46 , which means that @xmath49 has to be separable .
now , if @xmath29 is also faithful , i.e. , it satisfies property 2 , then we can see that also @xmath50 achieves @xmath46 iff the optimal @xmath49 is separable @xcite , which proves observation [ perfect ] .
below we shall provide a list of quantities , most of them satisfying both of the postulates .
_ identifying candidates for suitable entanglement measures .
_ this result , needless to say , leaves the question of finding entanglement measures exhibiting the above properties 1 . and 2 .
i.e. we need at least one such measure to prove the claim . moreover , any computable measure satisfying 1
. will give rise to a useful bound for capacity . _
( a ) the entanglement measure @xmath51 : _ define as in ref .
@xcite @xmath52where the infimum is performed over all kraus operators @xmath53 acting in @xmath5 only , satisfying @xmath54 and @xmath55 .
this is a computable single - shot quantity .
we denote the convex hull of this function with @xmath51 , @xmath56where @xmath57 , and which is an `` entanglement measure '' in its own right ( it is at least a monotone under one - way locc ) .
we claim that this function has the right properties .
the quantity @xmath51 has the properties 1 and 2 .
in fact , the validity of property 2 is easily shown : every separable state will have a convex combination in terms of products , for each of which @xmath58 will vanish . in turn , if a state is entangled , then there must in any convex combination be at least an entangled and hence correlated term , which will be detected by @xmath58 . to show property 1 ,
we can make use of a result of ref .
@xcite : for a pure tripartite state @xmath36 shared by @xmath3 , @xmath5 , and @xmath50 , a duality relation gives rise to @xmath59we now use the steps of ref .
@xcite iteratively . for a mixed four - partite state @xmath36 on @xmath3 , @xmath5 , @xmath50 , and @xmath60 , the optimal decomposition for @xmath61 in terms of pure states being @xmath62 , for each @xmath63 we have @xmath64 .
hence @xmath65arriving at property 1 .
this gives rise to a computable bound .
explicitly , it reads @xmath66with @xmath57 , as a single maximization .
a lower bound to this is @xmath67 which is usually less tight , but much simpler to compute .
_ ( b ) variants of the relative entropy of entanglement : _ the measure proposed in ref.@xcite is _ superadditive _ and not larger than the entanglement of formation , implying property 1 . it is also shown to be faithful in ref .
@xcite , which is property 2 . _
( c ) squashed entanglement : _ the _ squashed entanglement _
@xcite @xmath68 is also known to be superadditive and is bounded from above by the entanglement of formation , so qualifies as a bound for the same reason .
it is not easily computable , however , as it is based on a construction involving a state extension the dimension of which is not bounded .
however a lower bound to squashed entanglement was provided in ref .
@xcite : @xmath69 in terms of the trace - norm distance to the set of separable quantum states @xmath27 with respect to the split @xmath70 . _
( d ) distillable entanglement : _ a not efficiently computable but in instances practical bound is provided by the locc or ppt _ distillable entanglement _ with respect to @xmath70 .
( note that either version of distillable entanglement does not satisfy property 2 . )
$ ] .
the chosen kraus operators delivering good bounds for @xmath71 are given by @xmath72 , @xmath73 , @xmath74 for @xmath75 ( dashed ) and @xmath76 ( solid ) .
any choice for kraus operators yields a bound , and with the given choice ( determined via numerical explorations ) one obtains a particularly tight bound.,width=207 ] _ example : the amplitude qubit damping channel . _ to find any non - trivial bound for the capacity of the amplitude damping channel has been an open problem for some time @xcite .
the methods proposed here give rise to such bounds .
the kraus operators of @xmath77 are given by @xmath78for @xmath79 $ ] .
the isometry @xmath8 of this qubit channel maps @xmath80to bound the correlation measure @xmath81 , any choice for @xmath82 and for @xmath83 giving rise to a positive operator valued measure amounts to a valid bound .
this gives rise to the bound depicted in fig .
2 @xcite for @xmath84 $ ] .
note that it is _ significantly _ tighter than the trivial bound @xmath85 , which here takes the value @xmath86 .
it is easy to see that for @xmath84 $ ] there always exists an input diagonal in the computational basis that yields an output @xmath87 with unit entropy . for @xmath88
the channel becomes the perfect channel with @xmath89 .
( the _ entanglement assisted classical information capacity _
@xcite is also a crude upper bound , but yields values even larger than @xmath86 for @xmath90 $ ] ) .
we have hence established a first non - trivial bound for the amplitude damping channel .
needless to say , the same techniques can be applied to any finite - dimensional quantum channel . _
quantum capacity .
_ indeed , an argument very similar to the above one for the classical capacity of a quantum channel holds true also for the quantum capacity .
we arrive at the following conclusion ( for details , see epaps @xcite ) .
again , entanglement can help to a certain degree , but never uplift channels to the maximum possible value .
[ qperfect ] for every quantum channel for which the single - shot quantum capacity is not yet already given by the trivial upper bound @xmath91 , the same will hold true for the quantum capacity . _
summary and outlook .
_ in this work , we have investigated the converse question to the additivity problem : how much can entanglement help enhance capacities of quantum channels . in the focus of interest was the question whether entanglement can ever enhance the capacity to its trivial maximum if a single invocation does not yet reach that .
we affirmatively answer that question to the negative , including the quantum and classical capacity . in doing so
, we have established practical computable upper bounds to capacities , relating them to entanglement measures and rendering bounds and witnesses to the latter quantities useful to assess capacities .
there is though an interesting challenge : all the quantities from our list exhibit a sort of monogamy , i.e. , for states which are highly sharable they have to be small , implying that the bounds may become loose .
an example is a channel whose stinespring dilation gives rise to a @xmath92-dimensional anti - symmetric space .
a normalized projector onto this subspace is @xmath92-sharable , which means that for large @xmath92 all our bounds would tend to @xmath93 , while a direct approach of ref .
@xcite shows that the capacity is bounded by a constant independent of @xmath92 .
an open question is therefore how to find a quantity satisfying our property 1 , but that would not necessarily drop for sharable states .
it is the hope that the present work triggers further work on how `` small '' violations of additivity really are in practice and what role entanglement plays after all in quantum communication . _
acknowledgements .
_ we thank m. christandl , a. harrow , m.p .
mller , and a. winter for useful feedback .
fb is supported by a `` conhecimento novo '' fellowship from the brazilian agency fundao de amparo a pesquisa do estado de minas gerais ( fapemig ) .
je is supported by the eu ( qessence , minos , compas ) , the bmbf ( quorep ) , and the euryi .
mh is supported by the polish ministry of science and higher education grant n n202 231937 and by the eu ( qessence ) .
dy is supported by nsfc ( grant 10805043 ) .
part of this work was done in the national quantum information centre of gdask .
fb , je and mh thank the hospitality of the mittag - leffler institute , where part of this work has been done .
99 m. hastings , nature physics * 5 * , 255 ( 2009 ) .
c. king and m. b. ruskai , ieee trans . inf .
theory * 47 * , 192 ( 2001 ) ; c. king , j. math .
phys . * 43 * , 4641 ( 2002 ) ; c. king , quant .
inf . comp . * 3 * , 186 ( 2003 ) ; m. fannes , b. haegeman , m. mosonyi , and d. vanpeteghem , quant - ph/0410195 ; p. w. shor , j. math .
phys . * 43 * , 4334 ( 2002 ) ; m. m. wolf and j. eisert , new j. phys .
* 7 * , 93 ( 2005 ) .
r. horodecki , p. horodecki , and m. horodecki , rev .
mod .
phys . * 81 * , 865 ( 2009 ) .
g. g. amosov , a. s. holevo , and r. f. werner , problems in inf .
trans . *
36 * , 305 ( 2000 ) .
p.
w. shor , comm . math .
phys . * 246 * , 453 ( 2004 ) ; k. m. r. audenaert and s. l. braunstein , ibid . *
246 * , 443 ( 2004 ) .
p.
w. shor , _ the quantum channel capacity and coherent information _ , msri workshop on quantum computation ( 2002 ) .
l. henderson and v. vedral , j. phys .
a * 34 * , 6899 ( 2001 ) .
h. ollivier and w. h. zurek , phys .
rev .
lett .
* 88 * , 017901 ( 2001 ) .
k. matsumoto , t. shimono , and a. winter , comm .
math .
phys .
* 246 * , 427 a. s. holevo , quant - ph/9809023 .
this result may indeed be viewed as the channel analogue of the observation that the entanglement cost @xmath94 is strictly positive for every entangled state @xcite .
d. yang , m. horodecki , r. horodecki , and b. synak - radtke , phys .
rev .
lett . * 95 * , 190501 ( 2005 ) .
m. piani , phys .
rev .
lett . * 103 * , 160504 ( 2009 ) .
m. christandl and a. winter , j. math .
phys .
* 45 * , 829 ( 2004 ) ; r. r. tucci , quant - ph/9909041 . c. h. bennett , p. w. shor , j. a. smolin , and a. v. thapliyal , quant - ph/0106052
. f. g. s. l. brandao , m. christandl , and j. yard , arxiv:1010.1750 .
d. leung and g. smith , private communication ( 2010 ) .
with @xmath95 , one finds @xmath96 still , the quantity @xmath97 is not convex .
this is numerically evaluated by optimizing over mixed - state ensembles @xmath57 using global optimization both based on the routine @xmath98 in matlab with randomly sampled initial conditions and simulated annealing .
e. m. rains , ieee trans .
inf .
th . * 47 * , 2921 ( 2001 ) ; k. m. r. audenaert et al , phys .
rev .
lett . * 87 * , 217902 ( 2001 ) .
m. christandl , n. schuch , and a. winter , phys .
rev .
lett . * 104 * 240405 ( 2010 ) .
see epaps . |
this paper studies the statistics of the extremes of a gaussian field whose correlations decay logarithmically with the distance .
the model is related to the process introduced by bacry and muzy @xcite ( see also barral and mandelbrot @xcite ) and is similar to the _ logarithmic random energy model _ or _ log - rem _ studied by carpentier and le doussal @xcite , and fyodorov and bouchaud @xcite .
another important log - correlated model is the two - dimensional discrete gaussian free field .
the statistics of the extremes of log - correlated gaussian fields are expected to resemble those of i.i.d .
gaussian variables or _ random energy model _ ( rem ) and at a finer level , those of branching brownian motion .
in fact , log - correlated fields are conjectured to be the critical case where correlations start to affect the statistics of the extremes .
the reader is referred to the works of carpentier and le doussal @xcite ; fyodorov and bouchaud @xcite ; and fyodorov , le doussal and rosso @xcite for physical motivations of this fact .
the analysis for general log - correlated gaussian field is complicated by the fact that , unlike branching brownian motion , the correlations do not necessarily exhibit a tree structure .
the approach of this paper is in the spirit of the seminal work of derrida and spohn @xcite who studied the extremes of branching brownian motion using the gibbs measure .
the method of proof presented here is robust and applicable to a large class of nonhierarchical log - correlated fields .
the model studied here has the advantages of having a graphical representation of the correlations , a continuous scale parameter and no boundary effects ( cf .
section [ sec1.1 ] ) which make the ideas of the method more transparent .
even though correlations are not tree - like for general log - correlated models , such fields can often be decomposed as a sum of independent fields acting on different scales .
the main results of the paper are theorem [ thm : low - temperature ] on the correlations of the extremes and theorem [ thm : main4 ] on the statistics of the gibbs weights .
the results show that , in effect , the statistics of the extremes of the log - correlated field are the same as those of branching brownian motion at the level of the gibbs measure , as conjectured by carpentier and le doussal @xcite . the method of proof is outlined in section [ sect : outline ] .
the proof of the first theorem is based on an adaptation of a technique of bovier and kurkova @xcite originally developed for hierarchical gaussian fields such as branching brownian motion . for this purpose
, we need to introduce a family of log - correlated gaussian models where the variance of the fields in the scale - decomposition depends on the scale .
the free energy of the perturbed models is computed using ideas of daviaud @xcite .
the second theorem on the poisson dirichlet statistics of the gibbs weights is proved using the first theorem on correlations and general spin glass theory results .
following @xcite , we consider the half - infinite cylinder @xmath3_\sim , y \in\r_+^ * \bigr\},\ ] ] where @xmath0_\sim$ ] stands for the unit interval where the two endpoints are identified .
we write @xmath4 for the distance on @xmath0_\sim$ ] .
the following measure is put on @xmath5 : @xmath6 for @xmath7 , _ the variance parameter _ , there exists a random measure @xmath8 on @xmath5 that satisfies : for any measurable set @xmath9 in @xmath10 , the random variable @xmath11 is a centered gaussian with variance @xmath12 ; for every sequence of disjoint sets @xmath13 in @xmath14 , the borel @xmath15-algebra associated with @xmath16 , the random variables @xmath17 are independent and @xmath18 let @xmath19 be the probability space on which @xmath8 is defined , and let @xmath20 be the law of @xmath8 .
the space @xmath19 is endowed with the @xmath15-algebras @xmath21 generated by the random variables @xmath11 , for all the sets @xmath9 at a distance greater than @xmath22 from the @xmath23-axis .
the reader is referred to @xcite for the existence of the probability space @xmath24 .
the subsets needed for the definition of the gaussian field are the cone - like subsets @xmath25 of @xmath5 , @xmath26 where @xmath27 for @xmath28 and @xmath29 otherwise .
see figure [ fig : cones ] for a depiction of the subsets .
observe that , by construction , if @xmath30 , then @xmath25 and @xmath31 intersect exactly above the line @xmath32 .
the gaussian process @xmath33_\sim)$ ] is defined using the random measure @xmath8 , @xmath34_\sim.\ ] ] by properties ( i ) and ( ii ) of @xmath8 listed above , the covariance between @xmath35 and @xmath36 is given by the integral over @xmath37 of the intersection of @xmath38 and @xmath31 , @xmath39 = \int_{a_u(x)\cap a_u(x ' ) } \theta(\d s,\,\d y).\ ] ] and @xmath40 for @xmath41 .
the variance of the variables is given by the integral over @xmath42 of the lighter gray area above @xmath43 , and the covariance by the integral over the intersection of the subsets , the darker gray region . ]
the paper focuses on a discrete version of @xmath44 .
let @xmath45 , and take @xmath41 .
define the set @xmath46 the notation @xmath47 and @xmath48 will be used equally depending on the context . for a given @xmath49 ,
the @xmath50-correlated gaussian field is the collection of gaussian centered random variables @xmath51 for @xmath52 , @xmath53 a compelling feature of this construction is that a _ scale decomposition _ for @xmath54 is easily obtained from property ( ii ) above .
indeed , it suffices to write the variable @xmath55 as a sum of independent gaussian fields corresponding to disjoint horizontal strips of @xmath5 .
the @xmath56-axis then plays the role of the scale .
the covariances of the field are computed from ( [ eqn : corr ] ) by straightforward integration ; see also figure [ fig : cones ] .
[ lem : covariance ] for any @xmath57 , @xmath58 & = & \sigma^2 ( \log n + 1-\log2),\qquad x \in\mathcal x_n , \\
\e [ x_xx_{x ' } ] & = & \sigma^2\bigl ( \log \bigl(1/ \bigl\vert x - x ' \bigr\vert\bigr)-\log2\bigr),\qquad x \neq x ' \in\mathcal x_n.\end{aligned}\ ] ] similar constructions of log - correlated gaussian fields using a random measure on cone - like subsets are also possible in two dimensions ; see , for example , @xcite . without loss of generality , the results of this section are stated for the variance parameter @xmath59 .
the points where the field is unusually high , _ the extremes _ or _ the high points _ , can be studied using a minor adaptation of the arguments of daviaud for the two - dimensional discrete gaussian free field @xcite .
we denote by @xmath60 the cardinality of a finite set @xmath61 .
[ thm : main2 ] let @xmath62 be the set of @xmath63-high points .
then for any @xmath64 , @xmath65 moreover , for all @xmath66 there exists a constant @xmath67 such that @xmath68 for @xmath49 large enough .
the technique of daviaud is based on a tree approximation introduced by bolthausen , deuschel and giacomin @xcite for the discrete two - dimensional gaussian free field . there
, the technique is used to obtain the first order of the maximum .
the same argument applies here .
theorem [ thm : main2 ] and simple gaussian estimates yield @xmath69 the important feature of theorem [ thm : main2 ] and equation ( [ eqn : first order ] ) is that they are identical to the results for @xmath49 i.i.d .
gaussian variables of variance @xmath70 . in other words ,
the above observables of the high points are not affected by the correlations of the field .
the i.i.d .
case is called the _ random energy model _ ( rem ) in the spin glass literature .
the starting point of the paper is to understand to which extent i.i.d .
statistics is a good approximation for more refined observables of the extremes of log - correlated gaussian fields . to this end
, we turn to tools of statistical physics which allow for a good control of the correlations .
first , consider the _ partition function _
@xmath71 of the model ( @xmath72 stands for the inverse - temperature ) , @xmath73 and the _ free energy _ @xmath74 theorem [ thm : main2 ] is used to compute the free energy of the model .
[ cor : main3 ] let @xmath75 .
then , for all @xmath76 @xmath77 the free energy is the same as for the rem with variance @xmath70 .
in particular , the model undergoes _ freezing _ above @xmath78 in the sense that the quantity @xmath79 is constant .
more importantly , consider the _
normalized gibbs weights _ or _ gibbs measure
_ @xmath80 by design , the gibbs measure concentrates on the high points of the gaussian field .
the first main result of the paper is to achieve a control of the correlations at the level of the gibbs measure .
precisely , with spin glasses in mind , we consider the normalized covariance or _ overlap _ @xmath81 clearly , @xmath82 and @xmath83 .
moreover , the overlap @xmath84 is equal to the normalized correlations @xmath85
/ \e [ x_x^2 ] $ ] plus a term that goes to zero as @xmath49 goes to infinity . a fundamental object , that records the correlations of high points , is the _ distribution function of the overlap _ sampled from the gibbs measure .
namely , denote by @xmath86 the product measure on @xmath87 .
let @xmath88 be sampled from @xmath86 .
write for simplicity @xmath89 for @xmath90
. the averaged distribution function of the overlap is @xmath91 , \qquad 0\leq q \leq1.\ ] ] the first result is the analogue of results of derrida and spohn for the gibbs measure of branching brownian motion ( see equation ( 6.19 ) in @xcite ) , of chauvin and rouault on branching random walks @xcite and of bovier and kurkova on derrida s _ generalized random energy models _ ( grem ) @xcite .
it had been conjectured for nonhierarchical log - correlated gaussian field by carpentier and le doussal ; see page 16 in @xcite .
[ thm : low - temperature ] for @xmath92 , @xmath93= \cases { \displaystyle\frac{\beta_c}{\beta},&\quad $ \mbox{for $ 0 \le q <
1$,}$\vspace*{2pt } \cr 1 , & \quad $ \mbox{for $ q=1$.}$}\ ] ] this result is the same as for the rem model @xcite .
it is therefore consistent with rich statistics of extremes consisting of many high values order one away of each other and whose correlations are either very high or close to @xmath1 .
this result is in expectation .
the typical behavior of the random variable @xmath94 for @xmath95 small in terms of @xmath72 should be exponentially small in @xmath72 rather than @xmath96 . to see this , at the heuristic level , it is informative to consider the i.i.d .
case where the same phenomenon occurs . consider @xmath49 i.i.d .
gaussian random variables @xmath97 of variance @xmath70 ordered in a decreasing way . in this case , @xmath98 if @xmath99 .
the following inequality is easily verified : @xmath100 in particular , since the gap @xmath101 is of order one in the limit and since the density of points at distance @xmath23 from the maximum is bounded by @xmath102 for @xmath103 large enough ( see @xcite for a precise statement in terms of extremal process ) , the typical behavior of @xmath104 is expected to be exponentially small in @xmath72 .
we remark also that for @xmath105 the free energy contains all information about the two - overlap distribution .
indeed , since the free energy in corollary [ cor : main3 ] is differentiable for every @xmath76 including @xmath78 , we have by the convexity of the free energy that the derivative of the limit is the limit of the derivatives . hence @xmath106 \bigr)=f'(\beta).\ ] ] the first equality is by gaussian integration by part .
it follows that@xmath107=0 $ ] for @xmath105 .
in particular , since the correlations are positive , the overlap of two sampled points is @xmath1 almost surely for every @xmath105 . in the case of @xmath108 , the first moment of the two - overlap distribution is strictly greater than @xmath1 , therefore more information is needed to determine the distribution .
one way to proceed would be to obtain enough expectations of functions of @xmath89 to determine the distribution .
this can be done by adding parameters to the field and consider the appropriate derivative of the free energy of the perturbed model .
this is similar in spirit to the @xmath109-spin perturbations for the sherrington
kirkpatrick model in spin glasses ; see , for example , @xcite .
it turns out that this kind of pertubative approach pioneered by bovier and kurkova in @xcite for gaussian fields on trees can be generalized to log - correlated fields .
the control of the correlations is achieved by introducing a perturbed version of the model at a specific scale ; cf . section [ sect : perturbed ] . in the present case , the proof is more intricate since the structure of correlations of the gaussian field for finite @xmath49 is not tree - like or _
ultrametric _ as in the cases of branching brownian motion and grem s .
for example , for branching brownian motion , @xmath84 corresponds to the branching time of the common ancestor of two particles at time @xmath110 , @xmath23 and @xmath56 , divided by @xmath110 . because of the branching structure , @xmath111 [ the terminology _ ultrametric _ comes from the fact that the distance induced by the form @xmath112 is ultrametric when ( [ eqn : ultra ] ) holds . ]
the _ parisi ultrametricity conjecture _ in the spin - glass literature states that , even though tree - like correlations might not be present for finite @xmath49 , ultrametric correlations are recovered in the limit @xmath113 for a large class of gaussian fields at the level of the gibbs measure , that is , @xmath114 = 1.\ ] ] it is not hard to see that theorem [ thm : low - temperature ] implies the ultrametricity conjecture for the gaussian field considered , since the overlaps can only take value @xmath1 or @xmath2 .
( in the language of spin glasses , the field is said to admit a _
one - step replica symmetry breaking _ at low temperature . )
the second main result describes the joint distribution of overlaps sampled from the gibbs measure . to this end , for @xmath115
, we denote the product of gibbs measure on @xmath116 by @xmath117 .
we consider the class of continuous functions @xmath118^ { { s(s-1)}/{2}}\to\r$ ] .
we write @xmath119 $ ] for @xmath120 $ ] , that is , the averaged expectation of @xmath121 when @xmath122 is sampled from @xmath117 .
we recall the definition of a poisson dirichlet variable . for @xmath123 ,
let @xmath124 be the atoms of a poisson random measure on @xmath125 of intensity measure @xmath126 .
a _ poisson
dirichlet variable _ @xmath127 of parameter @xmath128 is a random variable on the space of decreasing weights @xmath129 with @xmath130 and @xmath131 which has the same law as @xmath132 where @xmath133 stands for the decreasing rearrangement .
[ thm : main4 ] let @xmath108 and @xmath134 be a poisson dirichlet variable of parameter @xmath135 .
denote by @xmath136 the expectation with respect to @xmath127 . for any continuous function @xmath118^{{s(s-1)}/{2}}\to\r$ ] of the overlaps of @xmath137 points ,
@xmath138 = e \biggl [ \sum_{k_1\in\n,\ldots , k_s\in\n } \xi_{k_1}\cdots\xi_{k_s } f(\delta_{k_lk_{l ' } } ) \biggr].\ ] ] it is important to stress that , as in the case of branching brownian motion and unlike the rem , it is not the collection @xmath139 _ per se _ that converges to a poisson dirichlet variable .
rather , the result suggests that the poisson
dirichlet weights are formed by the sum of the gibbs weights of high points that are arbitrarily close to each other because the continuity of the function @xmath140 naturally identifies points @xmath141 for which @xmath84 tends to @xmath2 in the limit @xmath113 . in the theory of spin glasses ,
these clusters of high points are often called _
pure states_. for more on the connection with spin glasses , the reader is referred to @xcite where the pure states are constructed explicitly for mean - field models .
bolthausen and kistler have studied a family of models called _ generalized grems _ for which the correlations are not ultrametric @xcite for finite @xmath49 . by construction , the overlaps of these models can only take a finite number of values ( uniformly in @xmath49 , the number of variables ) .
they compute the free energies and the gibbs measure and prove the parisi ultrametricity conjecture for these .
bovier and kurkova @xcite have obtained the distribution of the gibbs measure for gaussian fields , called the crems , where the values of the overlaps are not a priori restricted . their analysis is restricted to models with ultrametric correlations and include the case of branching brownian motion . the works of bolthausen , deuschel and zeitouni @xcite , bramson and zeitouni @xcite and ding @xcite establish the tightness of the recentered maximum of the two - dimensional discrete gaussian free field .
we expect that their method can be applied to the gaussian field we consider .
we note that fang and zeitouni @xcite have studied a branching random walk model where the variance of the motion is time - dependent .
this model is related to the simpler grem model of spin glasses and to the crem of bovier and kurkova .
the family of log - correlated gaussian fields introduced in section [ sec2.2 ] is akin to these hierarchical models , where the scale parameter replaces the time parameter .
the proof is split in three steps , and each can be adapted ( with different correlation estimates ) to other log - correlated gaussian fields .
the gaussian field we study has a graphical representation of its correlations as well as no boundary effect which help in illustrating the method . in this section ,
we define a family of gaussian fields for which the variance parameter @xmath15 is scale - dependent . it can be seen as the grem analogue for the nonhierarchical gaussian field considered here .
we restrict ourselves to the case where @xmath15 takes two values , which is the one needed for the proof of theorem [ thm : low - temperature ] .
however , the construction and the results can hold for any finite number of values .
fix @xmath142 .
we introduce a scale ( or time ) parameter @xmath110 by defining for any @xmath143 $ ] , @xmath144 observe that for any fixed @xmath23 , the process @xmath145 has independent increments and is a martingale for the filtration @xmath146 , @xmath147= x_x(s)\qquad \mbox{for $ t > s$}.\ ] ] this is a consequence of the defining property ( ii ) of the random measure @xmath8 .
the parameters of the family of perturbed models are @xmath128 where @xmath148 and @xmath149 with @xmath150 , @xmath151 . for the sake of clarity and to avoid repetitive trivial corrections , it is assumed throughout the paper that @xmath152 and @xmath153 are integers .
the gaussian field @xmath154 is defined from the field @xmath54 as follows : @xmath155 the construction is depicted in figure [ fig : perturbed - daviaud1 ] .
we write @xmath156 for the field @xmath157 .
the dependence on @xmath158 and @xmath159 will sometimes be dropped in the notation of @xmath160 for simplicity . .
] consider the partition function @xmath161 of the perturbed model @xmath162 and the free energy @xmath163 the log number of high points can be computed for the gaussian field @xmath160 using daviaud s technique recursively .
the free energy is then obtained by doing an explicit sum on these high points .
this is the object of sections [ sec : perturbed - maximum ] and [ sect : free energy ] .
the result is better expressed in terms of the free energy of the rem with @xmath49 i.i.d .
gaussian variables of variance @xmath164 , @xmath165 corollary [ cor : main3 ] follows from the next result with the choice @xmath166 . [ prop : perturbed - free - energy ] let @xmath167 .
then : * _ case _ 1 : if @xmath168 , @xmath169 * _ case _ 2 : if @xmath170 , @xmath171 where the convergence holds almost surely and in @xmath172 .
the expressions are identical to the free energy of a grem with two levels . in case 1
, it is reduced to a rem .
the conditions can be rewritten by defining a piecewise linear function of slopes @xmath173 and @xmath174 on the intervals @xmath175 $ ] , @xmath176 $ ] , respectively . in case 1 , this function fails to be concave .
however , it is easily verified that the effective parameters define the concave hull of the function .
the reader is referred to @xcite and @xcite for more details on the concavity conditions which is very general for the family of grem models . in case 1
there is one critical value for @xmath72 , and in case 2 there are two critical values for @xmath72 corresponding to the respective @xmath177 of the two effective parameters @xmath178 . in case 1 , the critical @xmath72 is @xmath179 , whereas the two critical @xmath72 s are @xmath180 and @xmath181 in case 2 . on the strip
@xmath182 $ ] where @xmath41 . ]
the proof of theorem [ thm : low - temperature ] relies on determining the overlap distribution of the original model from the free energy of the perturbed ones .
this approach has been used by bovier and kurkova in the case of the grem - type models @xcite . for @xmath183 and @xmath184 ,
consider the field @xmath185 defined in ( [ eqn : perturbed y ] ) with the choice of parameters @xmath186 ; see figure [ fig : perturbed - daviaud2 ] .
( recall that , for the sake of clarity , it is assumed that @xmath187 and @xmath153 are integers . ) the original gaussian field @xmath188 is recovered at @xmath189 . note that if @xmath190 , the parameters correspond to the first case of proposition [ prop : perturbed - free - energy ] and if @xmath191 , to the second .
the field @xmath160 can also be represented as follows : @xmath192 the proof of the next lemma is a simple integration and is postponed to the ; see appendix [ proof - covariance - acc - perturbed ] . [
lem : covariance - acc - perturbed ] fix @xmath57 , and @xmath184 .
let @xmath193 .
then , for @xmath194 @xmath195 = \e[\tilde x_xx_x]= ( 1 -\alpha ) \log n , \qquad x\in\mathcal { x}_\varepsilon,\ ] ] and , for @xmath196 , @xmath197=\cases { \bigl(q\bigl(x , x'\bigr)-\alpha \bigr ) \log n + o_n(1 ) , & \quad $ \mbox{if } \alpha < q\bigl(x , x ' \bigr ) \le1,$ \vspace*{2pt}\cr 0 , & \quad $ \mbox{if } 0 \le q\bigl(x , x'\bigr ) \le\alpha,$}\ ] ] where @xmath198 is a term uniformly bounded in @xmath49 , and we recall that @xmath199 .
this result and a gaussian integration by parts yield an important lemma .
[ lem : perturbed - ipp ] for all @xmath184 , we have @xmath200,\ ] ] where @xmath201 stands for a term that goes to @xmath1 as @xmath49 goes to @xmath202 . fix @xmath41 and @xmath184 .
note that @xmath203 is a gaussian vector of @xmath204 variables .
therefore , gaussian integration by parts ( see lemma [ lem : gaussianibp ] ) yields , for all @xmath194 , @xmath205 & = & - \sum _ { x'\in\mathcal{x}_\varepsilon } { \mathbb e}[\tilde x_x x_{x ' } ] { \mathbb e } \biggl[\frac{\ee^{\beta(x_x+x_{x ' } ) } } { ( \sum_{z\in \mathcal{x}_\varepsilon } \ee^{\beta x_z } ) ^2 } \biggr ] \\ & & { } + { \mathbb e}[\tilde x_xx_x ] { \mathbb e}\biggl [ \frac{\ee^{\beta x_x}}{\sum_{z\in\mathcal{x}_\varepsilon } \ee ^{\beta x_z } } \biggr].\end{aligned}\ ] ] lemma [ lem : covariance - acc - perturbed ] and elementary manipulations imply @xmath206 \\ & & \qquad= \sum_{x , x'\in\mathcal x_\varepsilon } \biggl ( \int_{\alpha}^{1 } \mathbf{1}_{\{q(x , x')\le s\ } } \,\d s \biggr ) { \mathbb e}\bigl [ g_{\beta , n}(x ) g_{\beta , n}\bigl(x'\bigr ) \bigr ] + o
\biggl(\frac{1}{\log n } \biggr ) \\ & & \qquad= \int_{\alpha}^{1 } { \mathbb e}\bigl [ g_{\beta , n}^{\times2 } \{q_{12 } \le s \ } \bigr ] \,\d s + o \biggl(\frac{1}{\log n } \biggr),\end{aligned}\ ] ] which concludes the proof of the lemma .
proof of theorem [ thm : low - temperature ] fix @xmath207 .
write @xmath208 for the partition function ( [ eq : partition - function - perturbed ] ) for the choice @xmath209 . direct differentiation and equation ( [ eqn : y rep ] ) give @xmath210,\ ] ] which , together with lemma [ lem : perturbed - ipp ] , yields @xmath211 observe that @xmath212 is a convex function of @xmath22 . moreover , by proposition [ prop : perturbed - free - energy ] , @xmath213 converges .
the limit , that we denote @xmath214 , is also convex in the parameter @xmath22 .
in particular , by a standard result of convexity ( see , e.g. , proposition i.3.2 in @xcite ) , at every point of differentiability , the derivative of the limit equals the limit of the derivative @xmath215 \\[-8pt ] \eqntext{\mbox{$\forall u$ where $ u\mapsto f^{(u , \alpha)}(\beta)$ is differentiable}.}\end{aligned}\ ] ] we show @xmath214 is differentiable at @xmath189 .
the derivative can be computed by proposition [ prop : perturbed - free - energy ] . for @xmath22 small enough
, @xmath72 is larger than all critical @xmath72 s .
thus @xmath216 from this , it is easily verified that @xmath214 is differentiable at @xmath189 and @xmath217 equations ( [ eqn : convex 1 ] ) , ( [ eqn : convex 2 ] ) and ( [ eqn : convex 3 ] ) together imply @xmath218 therefore , any weak limit @xmath219 must satisfy @xmath220 for any point of continuity @xmath221 , since @xmath219 is nondecreasing . if there exists @xmath123 such that @xmath222 , there would be a contradiction with ( [ eqn : weak conv ] ) , since by right - continuity and monotonicity of @xmath219 we could find @xmath223 such that @xmath224 this proves that any weak limit @xmath219 of @xmath225 is the same and equals @xmath226 on @xmath227 .
the subsequential limits being the same , this proves in particular convergence of the sequence to the desired distribution function . in this section
, the link between theorems [ thm : low - temperature ] and [ thm : main4 ] is explained .
the technique , inspired from the study of spin glasses in particular @xcite , is general and is of independent interest to prove convergence to poisson
dirichlet statistics .
the first step is to find a good space for the convergence of @xmath228 .
let @xmath103 be the compact metric space of @xmath229 covariance matrices with @xmath2 on the diagonal endowed with the product topology on the entries .
for a given @xmath49 , consider the mapping @xmath230 where for @xmath231 @xmath232 consider the probability measure @xmath233 on @xmath234 .
the push - forward of this probability measure under the above mapping defines a random element of @xmath103 that we denote @xmath235 . since each point is sampled independently from the same measure , the law of @xmath235 is _ weakly exchangeable _ , that is , for any permutation @xmath236 of a finite number of indices , @xmath237 the sequence of random matrices @xmath238 is tight by prokhorov s theorem since the space @xmath103 is a compact metric space .
hence , there exists a subsequence @xmath239 that converges weakly .
denote the subsequential limit by @xmath240 .
observe that @xmath240 is also weakly exchangeable since the mappings on @xmath103 induced by a finite permutation is continuous .
therefore , by the representation theorem of dovbysh and sudakov @xcite , @xmath240 is constructed like @xmath235 by sampling from a random measure .
precisely , the theorem states that there exists a random probability measure @xmath241 on a hilbert space @xmath242 , with law @xmath243 and corresponding expectation @xmath136 , such that the random matrix @xmath240 has the same law as the gram matrix of a sequence of vectors @xmath244 that are sampled under @xmath245 .
[ in other words , the vectors @xmath244 are i.i.d . conditionally on @xmath241 . ]
the equality in law can be expressed as follows : for any continuous function @xmath140 on @xmath103 , @xmath246= e\mu_{\beta}^{\times\infty}\bigl[f \bigl(v_l \cdot v_l'\bigr)\bigr].\ ] ] note that , since @xmath247 , the random measure @xmath248 is supported on the unit ball .
the first consequence of theorem [ thm : low - temperature ] is that for any subsequential limit @xmath241 , @xmath249 & = & \lim_{n\to\infty } { \mathbb e}\bigl [ g_{\beta , n}^{\times2 } \ { q_{12 } \le q \ } \bigr ] \nonumber \\[-8pt ] \\[-8pt ] \nonumber & = & \frac{\beta_c}{\beta } 1_{[0,1)}(q)+ 1_{\ { 1\}}(q).\end{aligned}\ ] ] the first equality is obtained by bounding @xmath250}(q_{ll'})$ ] by continuous functions on @xmath251 above and below and by applying ( [ eqn : mu ] ) . in view of equations ( [ eqn :
mu ] ) and ( [ eqn : x mu ] ) , we see the random measures @xmath241 as limit points of @xmath252 . the main ingredient to prove poisson
dirichlet statistics is a general property of the gibbs measure @xmath253 of centered gaussian fields known as the ghirlanda
guerra identities .
they were introduced in @xcite and were proved in a general setting by panchenko @xcite .
[ thm : gg ] let @xmath241 be a subsequential limit of @xmath254 inthe sense of ( [ eqn : mu ] ) . then for any @xmath255 and any continuous functions @xmath256^{s(s-1)/2}\to\r$ ] @xmath257 & = & \frac{1}{s } e \mu_\beta^{\times2 } [ v_1\cdot v_2 ] e \mu_\beta^{\times s } \bigl[f(v_l \cdot v_{l ' } ) \bigr ] \nonumber \\[-8pt ] \\[-8pt ] \nonumber & & { } + \frac{1}{s}\sum _ { k=2}^s e \mu_\beta^{\times s } \bigl[v_1\cdot v_k f(v_l\cdot v_{l ' } ) \bigr ] . \nonumber \ ] ] recall that we write @xmath117 for the product measure on @xmath116 .
also for @xmath258 , the overlaps @xmath259 , @xmath260 , are denoted @xmath251 . in a similar way
, we write @xmath261 for the field @xmath262 of the first point sampled from @xmath228 . it is shown in @xcite that , for any @xmath72 where the free energy @xmath263 is differentiable , the following concentration holds : @xmath264=0.\ ] ] note that by corollary [ cor : main3 ] , differentiability holds at all @xmath72 for the gaussian field considered . since the function @xmath140 is bounded , ( [ eqn : sa ] ) implies @xmath265- \e g_{\beta , n}[x_1 ] \e g_{\beta , n}^{\times s } \bigl[f(q_{ll ' } ) \bigr ] \bigr)=0.\vadjust{\goodbreak}\ ] ] the two terms can be evaluated by gaussian integrations by part ( see lemma [ lem : gaussianibp ] ) , @xmath266= 1- \e g_{\beta , n}^{\times 2}[q_{12 } ] + o \biggl(\frac{1}{\log n } \biggr)\ ] ] and @xmath267 \nonumber \\ & & \qquad = -s \e g_{\beta , n}^{\times s+1 } \bigl[q_{1,s+1 } f(q_{ll'})\bigr]+ \sum_{1 \le k \le s } \e g_{\beta , n}^{\times s } \bigl[q_{1k } f(q_{ll ' } ) \bigr ] \\ & & \qquad\quad{}+ o
\biggl(\frac { 1}{\log n } \biggr).\nonumber\end{aligned}\ ] ] finally recalling ( [ eq : gg - double+ ] ) and assembling ( [ eq - gg - ipp1+])([eq - gg - ipp2 + ] ) yields the ghirlanda guerra identities ( see equation ( 16 ) in @xcite ) , @xmath268 \nonumber \\ & & \qquad=\frac{1}{s } \e g_{\beta , n}^{\times2 } [ q_{12 } ] \e g_{\beta , n}^{\times s } \bigl[f(q_{ll ' } ) \bigr ] + \frac{1}{s}\sum_{k=2}^s\e g_{\beta , n}^{\times s } \bigl[q_{1k } f(q_{ll ' } ) \bigr]\\ & & \qquad\quad { } + o_n(1).\nonumber \ ] ] [ note that the term for @xmath269 cancels with the 1 since @xmath270 . ] in particular , for any subsequential limit @xmath241 of @xmath271 in the sense of ( [ eqn : mu ] ) , one obtains ( [ eqn : pre gg ] ) by taking the limit @xmath113 and applying the definition of convergence in the sense of ( [ eqn : mu ] ) . equation ( [ eqn : x mu ] ) and the ghirlanda
guerra identities imply that @xmath241 is atomic .
[ cor : atomic ] let @xmath241 be a subsequential limit of @xmath254 in the sense of ( [ eqn : mu ] ) . then there exist random weights @xmath272 with @xmath273 , @xmath274 and orthonormal vectors @xmath275 such that @xmath276 moreover , from ( [ eqn : x mu ] ) , @xmath277=1- \frac { \beta_c}{\beta}$ ] .
let @xmath244 be a sequence sampled from @xmath278 . from @xmath244 , we reconstruct @xmath241 up to isometry . for
a fixed @xmath279 consider the sequence @xmath280 .
this is a sequence of 0 s and 1 s by ( [ eqn : x mu ] ) .
we first show that , almost surely , for every @xmath281 , there exists @xmath282 such that @xmath283 ; in particular , since all vectors are in the unit ball , @xmath284 and @xmath285 . for this
, we proceed as in lemma 1 in @xcite .
write @xmath286 .
in other words , @xmath287 is @xmath2 if @xmath288 for @xmath289 , otherwise it is @xmath1 .
denote for short @xmath290 $ ] .
equation ( [ eqn : pre gg ] ) implies @xmath291 \\ & & \qquad= \frac{\alpha}{s}e\mu_\beta^{\times s}\{v_1\cdot v_l=0 , 2\leq l\leq s\ } + \frac{1}{s}\sum _ { l=2}^se\mu_\beta^{\times s } \{v_1\cdot v_l=0 , 2\leq l\leq s\ } \\ & & \qquad=\frac{s-1+\alpha}{s}e\mu_\beta^{\times s}\{v_1\cdot v_l=0 , 2\leq l\leq s\}= \frac{(s-1+\alpha)\cdots(1+\alpha ) \alpha } { s ! } , \ ] ] where the last equality is obtained by induction .
the last term goes to @xmath1 as @xmath292 since @xmath221 , hence @xmath293 from which we deduce that , @xmath243-a.s . , @xmath294 and then that , for @xmath241-almost all @xmath295 , @xmath296 since the vectors @xmath297 are i.i.d .
@xmath241-sampled , it follows that , @xmath243-a.s .
, for @xmath241-almost all @xmath295 , @xmath298 , thus @xmath299 as claimed . by the reasoning above , a vector that is sampled once in @xmath244
is sampled infinitely many times @xmath245-a.s . moreover , since the vectors are conditionally i.i.d .
, for @xmath281 , the following limit exists and must be nonzero : @xmath300 in particular , every sampled vector @xmath297 is an atom a.s .
and its weight is measurable with respect to @xmath244 .
moreover , if @xmath301 , then @xmath302 @xmath303-a.s . therefore the atoms are orthogonal .
it remains to consider the different atoms without repetitions and reorder the weights .
let @xmath304 , @xmath305 where @xmath306 , @xmath307 where @xmath308 , and so forth . by construction , @xmath309 are orthonormal vectors .
( the collection is not necessarily infinite at this point . )
we can assign to each vector @xmath310 its weight @xmath311 by ( [ eqn : weight ] ) .
the collection can then be ordered in decreasing order to get the result .
the fact that @xmath277=1- \frac{\beta_c}{\beta}$ ] is straightforward from ( [ eqn : x mu ] ) .
to finish the proof of theorem [ thm : main4 ] , it remains to show that the random weights @xmath127 are distributed like a poisson dirichlet variable of parameter @xmath312 .
in fact , the parameter is already determined by corollary [ cor : atomic ] , since for a poisson dirichlet variable @xmath313 of parameter @xmath23 , @xmath314=1-x$ ] holds ; see , for example , corollary 2.2 in @xcite .
this will also imply that for any converging sequence of @xmath315 in the sense of ( [ eqn : mu ] ) , the limit is the same .
in particular , it implies convergence of the whole sequence by compactness . to prove the poisson
dirichlet statistics of the weights @xmath127 , we use the following characterization theorem of the law ; see @xcite , page 22 for details .
define for all @xmath316 the joint moments of the weights @xmath317 the collection of @xmath318 , @xmath316 , determines the law of a random mass - partition , that is , a random variable on ordered sequences @xmath319 with @xmath320 . if @xmath127 is a poisson dirichlet variable , it is shown in @xcite , proposition 1.2.8 , that the moments satisfy the recursion relations @xmath321 \\[-8pt ] \nonumber & & { } + \sum _ { 2\leq l \leq m } \frac{n_l}{s}s(n_1+n_l , n_2 , \ldots , n_{l-1 } , n_{l+1},\ldots , n_m),\end{aligned}\ ] ] where @xmath322 .
it is not hard to verify that all moments @xmath318 ( and thus the law of @xmath127 ) are determined by recursion from @xmath323 and the identities ( [ eqn : gg ] ) .
it turns out that these identities are satisfied by @xmath127 defined by theorem [ thm : gg ] and corollary [ cor : atomic ] .
[ thm : pd ] let @xmath127 be a random mass - partition satisfying the assumptions of corollary [ cor : atomic ] .
the moments @xmath318 of @xmath127 satisfy ( [ eqn : gg ] ) for any @xmath316 and any @xmath324 . in particular , @xmath127 has the law of a poisson dirichlet variable of parameter @xmath325 . to deduce ( [ eqn : gg ] ) from ( [ eqn : pre gg ] ) , we follow @xcite , pages 2425 .
the set @xmath326 can be decomposed into the disjoint union of sets @xmath327 with @xmath328 for all @xmath329 .
consider the functions @xmath330 given by @xmath331 and define @xmath332 .
then elementary manipulations imply ( [ eqn : gg ] ) .
note that the second term on the right - hand side of ( [ eqn : pre gg ] ) yields the last two terms of ( [ eqn : gg ] ) .
in this section , the log - number of high points at a given level is computed for the perturbed models introduced in section [ sect : outline ] .
the focus is on the gaussian field introduced in section [ sect : perturbed ] , though the technique applies to any perturbed model with a finite number of parameters .
the free energies of the models are computed in section [ sect : free energy ] .
let @xmath333 be the gaussian field introduced in section [ sect : perturbed ] .
recall the notation and the two choices of parameters in proposition [ prop : perturbed - free - energy ] : @xmath334 \\[-8pt ] \nonumber & & \mathit{case}\ 2\dvtx\qquad \sigma_1\geq\sigma_2 . \ ] ] define also as before @xmath167 .
[ prop : perturbed - maximum ] @xmath335 where @xmath336 [ prop : perturbed - highpoints ] let @xmath337 be the set of @xmath63-high points .
then , for all @xmath338 , @xmath339 where in case 1 , @xmath340 and in case 2 , @xmath341 moreover , for any @xmath342 , there exists @xmath343 such that @xmath344 the proof of case 1 is by a union bound , @xmath345 which goes to zero by a gaussian estimate ; see lemma [ lem : gaussian ] . for case 2
, we construct a gaussian field with hierarchical correlations that dominates @xmath160 at the level of the covariances .
the result will follow by comparison using slepian s lemma .
notice that if @xmath346 , the corresponding cone - like sets for @xmath347 and @xmath348 in @xmath349 intersect between the lines @xmath350 and @xmath351 .
therefore the covariance of the variables satisfies , writing @xmath352 , @xmath353 & = & \sigma_2 ^ 2 \int_{\ell}^{\varepsilon^{\alpha } } \frac { y-\ell}{y^2 } \,\d y + \sigma_1 ^ 2 \biggl ( \int_{\varepsilon^{\alpha } } ^{1/2 } \frac{y-\ell}{y^2 } \,\d y + \int_{1/2}^{\infty } \frac{1/2-\ell } { y^2 } \,\d y \biggr ) \\ &
\geq & \sigma_1 ^ 2 \biggl ( \log\frac{1/2}{\varepsilon^{\alpha } } -1 \biggr).\end{aligned}\ ] ] by applying the same reasoning when @xmath354 , one obtains the following lower bound for the covariance : @xmath355\geq\cases{0 , & \quad $ \mbox{if $ \bigl\|x - x'\bigr\| > \varepsilon^{\alpha},$}$\vspace*{2pt } \cr \displaystyle\sigma_1 ^ 2 \biggl(\log\frac{1/2}{\varepsilon^{\alpha}}-1 \biggr ) , & \quad $ \mbox{if $ \varepsilon < \bigl\|x - x'\bigr\| \leq\varepsilon^{\alpha}.$}$ } \ ] ] equation ( [ eqn : cov bound ] ) is used to construct a gaussian field @xmath356 .
define the map @xmath357 where @xmath358 is the unique @xmath359 such that @xmath360 .
( if @xmath361 , there are two possibilities for @xmath56 .
we take the right point . ) the pre - image of @xmath362 under @xmath236 are exactly the points in @xmath363 that are at a distance less than @xmath364 from @xmath56 .
one can think of @xmath358 as the _ ancestor _ of @xmath23 at the scale @xmath365 .
consider the following gaussian variables @xmath366 \\[-8pt ] \nonumber & & \bigl(g^{(2)}_{x } , x\in\mathcal x_{\varepsilon}\bigr ) \qquad \mbox{i.i.d .
gaussians of variance $ \sigma_2 ^ 2(1-\alpha ) \log n+2\sigma_1 ^ 2$.}\hspace*{-35pt}\end{aligned}\ ] ] these two families are also assumed independent .
then , the field @xmath356 is defined , using the map @xmath236 above and the gaussian random variables @xmath367 , by @xmath368 this construction and equation ( [ eqn : cov bound ] ) directly imply the following comparison lemma .
[ lem : comparison ] @xmath369&=&\e \bigl[y_x^2\bigr]\qquad \mbox{$\forall x\in\mathcal x_\varepsilon$ } , \nonumber \\[-8pt ] \\[-8pt ] \nonumber \e[\tilde y_x \tilde y_y]&\leq&\e[y_x y_y]\qquad \mbox{$\forall$ $ x\neq y$ , $ x , y\in\mathcal x_\varepsilon$.}\end{aligned}\ ] ] the following corollary is a straightforward consequence of the above lemma and slepian s lemma ; see corollary 3.12 in @xcite .
[ cor : bound max ] for any @xmath370 , @xmath371 the gaussian field @xmath356 is almost identical to a grem model with two levels with parameters @xmath123 and @xmath372 ; see , for example , @xcite .
in fact the only aspect different from an exact grem are the terms of order one in the variances of the gaussian random variables @xmath367 s .
however , these do not affect the first order of the maximum .
the proof of proposition [ prop : perturbed - maximum ] is concluded by the following standard grem result . the proof of the lemma is not hard and is omitted for conciseness .
the reader is referred to theorem 1.1 in @xcite where a stronger result on the maximum is given and to @xcite , lecture 9 , for more details on the free energy and on the log - number of high points of a two - level grem .
[ lem : grem ] let @xmath356 be the gaussian field constructed above .
then @xmath373 where @xmath374 is defined in proposition [ prop : perturbed - maximum ] .
the goal is to get an upper bound in probability ) for @xmath375 where @xmath376 . in case 1 ,
a first moment computation gives the result .
indeed , a gaussian estimate ( see lemma [ lem : gaussian ] ) gives @xmath377 = n \p(y_1 \ge\sqrt{2}\gamma\log n ) \le c n^{\mathcal{e}^{(\vec { \sigma } , \alpha)}(\gamma)},\ ] ] where @xmath378 .
therefore , by markov s inequality , for any @xmath66 , @xmath379 in case 2 , if @xmath380 the same argument gives the correct bound .
it remains to bound the case @xmath381 .
the argument is essentially an explicit comparison with a 2-level grem . for the scale @xmath128 ,
define @xmath382 a first moment computation yields , for any @xmath383 and any @xmath66 , @xmath384 similarly , a union bound gives @xmath385 recall that , for any @xmath386 , we denote by @xmath387 the closest point in @xmath388 , hence @xmath389 .
we define for all @xmath49 and @xmath390 , @xmath391 the parameter @xmath392 will be fixed later and will depend on @xmath393 . using a union bound together with lemma [ lem : lem12 ] , we obtain , for all @xmath390 , @xmath394 we also consider the events giving the log - number of high points at scale @xmath128 .
precisely , we divide @xmath395 $ ] in intervals of size @xmath396 where @xmath397 will be fixed later .
define @xmath398 , for @xmath399 and @xmath400 , \qquad 1 \le i \le m.\ ] ] by ( [ eq : controlscalealpha ] ) , the events @xmath401 are such that @xmath402 therefore , by ( [ eq : annu ] ) and ( [ eq : bni ] ) , we are reduced to estimate @xmath403 which is smaller than @xmath404.\ ] ] we split the set @xmath405 into the possible value of the field at scale @xmath128 @xmath406 the term in ( [ eqn : to bound ] ) can then be bounded above by @xmath407.\ ] ] if @xmath408 , note that @xmath409 satisfies @xmath410 where @xmath411 moreover , if @xmath412 , the maximum is attained at @xmath413 , thus @xmath414 for all @xmath415 $ ] . for @xmath416 , one gets @xmath417 \\ & & \qquad= \e\biggl [ \sum_{x\in\mathcal x_\varepsilon } \mathbf{1}_{\{y_x\geq \sqrt { 2}\gamma\log n , y_{\pi(x)}(\alpha)\in i^{(i)}\ } } ; a_{n,\nu}^ { c } \cap b_{n , i}^ { c } \biggr ] \\ & & \qquad\leq\e\biggl [ \sum_{x\in\mathcal x_\varepsilon } \mathbf{1}_{\ { y_x - y_x(\alpha)\geq\sqrt{2}(\gamma-\eta_i-\nu ) \log n , y_{\pi ( x)}(\alpha)\geq\sqrt{2}\eta_{i-1}\log n \ } } ; b_{n , i}^ { c } \biggr ] \\ & & \qquad \leq cn^{\mathcal{e}_1(\eta_{i-1 } ) + \rho/2 } n^{1-\alpha } n^{- { ( \gamma-\eta_i-\nu)^2}/{(\sigma_2 ^ 2(1-\alpha ) ) } } \\ & & \qquad= c n^{\rho/2}n^{1- { ( \eta_{i-1})^2}/{(\sigma_1 ^ 2\alpha)}- { ( \gamma-\eta_i-\nu)^2}/{(\sigma_2 ^ 2(1-\alpha))}},\end{aligned}\ ] ] where the last inequality follows by the definition of @xmath418 the independence of the field at different scales and a gaussian estimate . since @xmath419 for
all @xmath420 $ ] , the last term is smaller than @xmath421 by taking @xmath392 small enough and @xmath397 large enough , but fixed . for @xmath422 , a similar argument gives also the bound @xmath423 . putting this back in ( [ eqn : to bound ] ) shows that the term goes to @xmath1 as @xmath113 as desired .
the proof of the lower bound is two - step recursion .
two lemmas are needed .
the first is a generalization of the lower bound in daviaud s theorem ; see theorem [ thm : main2 ] or @xcite . [
lem : recurrence ] let @xmath424 .
suppose that the parameter @xmath15 is constant on the strip @xmath0_\sim\times[\varepsilon^{\alpha '' } , \varepsilon^{\alpha'}]$ ] , and that the event @xmath425 is such that @xmath426 for some @xmath427 , @xmath428 and @xmath429 .
let @xmath430 then , for any @xmath431 such that @xmath432 and any @xmath433 , there exists @xmath343 such that @xmath434 we stress that @xmath431 may be such that @xmath435 .
the second lemma , which follows , serves as the starting point of the recursion and is analogous to lemma 8 in @xcite .
[ lem : init ] for any @xmath436 such that @xmath437 , there exists @xmath438 and @xmath439 such that @xmath440 we first conclude the proof of the lower bound in proposition [ prop : perturbed - highpoints ] using the two above lemmas .
proof of the lower bound of proposition [ prop : perturbed - highpoints ] let @xmath63 such that @xmath338 .
choose @xmath441 such that @xmath442 .
it will be shown that for some @xmath443 @xmath444 by lemma [ lem : init ] , for @xmath445 arbitrarily close to @xmath1 , there exists @xmath438 and @xmath446 , such that @xmath447 observe that we have @xmath448 .
moreover , let @xmath449 lemma [ lem : recurrence ] is applied from @xmath436 to @xmath128 .
for any @xmath450 with @xmath451 and any @xmath452 , there exists @xmath453 such that @xmath454 therefore , lemma [ lem : recurrence ] can be applied from @xmath128 to @xmath2 for any @xmath450 with @xmath451 .
define similarly @xmath455 then , for any @xmath456 with @xmath457 and @xmath458 , there exists @xmath459 such that @xmath460 recalling that @xmath448 , equation ( [ eqn : lower- ] ) follows from ( [ eqn : lower prob ] ) if it is proved that @xmath461 for an appropriate choice of @xmath450 [ in particular such that @xmath451 ] .
it is easily verified that , for a given @xmath63 , the quantity @xmath463 is maximized at @xmath464 plugging these back in ( [ eqn : e_1 ] ) shows that @xmath465 provided that @xmath466 with @xmath436 small enough ( depending on @xmath63 ) .
furthermore , since @xmath467 we obtain @xmath468 , which completes the proof in the case @xmath469 .
if @xmath470 , the condition @xmath471 will be violated as @xmath436 goes to zero . in this case , for @xmath390 , pick @xmath472 such that @xmath473 .
the first term in @xmath474 corresponds to @xmath475 evaluated at @xmath476 for @xmath477 . in particular , @xmath478 . from ( [ eqn : e_2 ] ) , this shows that @xmath479 note that @xmath480 is strictly positive if and only if @xmath481 .
this concludes the proof of ( [ eqn : lower- ] ) .
proof of lemma [ lem : recurrence ] let @xmath431 such that @xmath432 and @xmath441 such that @xmath482 .
pick @xmath483 such that @xmath484 since @xmath485 , there exists @xmath486 such that @xmath487 for @xmath488 ( which will be fixed later ) , we set @xmath489 observe that the @xmath490 s and the @xmath491 s satisfy @xmath492 , and @xmath493 .
consider the sets @xmath494 given by @xmath495 for @xmath496 .
note that only half of the @xmath497 s in @xmath498 s are considered .
also , to each @xmath497 we consider the points @xmath499 in @xmath500 that are close to @xmath497 . by analogy with a branching process
, these points can be thought of as the _ children _ of @xmath501 .
the reason for these two choices is that the cones corresponding to the variables @xmath502 and @xmath503 do not intersect below the line @xmath504 if @xmath505 ; see figure [ fig : intersect - daviaud ]
. now consider , the sets of high points of @xmath494 , @xmath506 and @xmath507 where @xmath508 such that @xmath509 and @xmath510 .
furthermore , with these definitions and the choice of @xmath511 in ( [ eq : rec2 ] ) and ( [ eq : rec1 ] ) , we have for large @xmath49 @xmath512 it is thus sufficient to find a bound for @xmath513 to prove the lemma . for events @xmath514 to be defined in ( [ eqn : c ] ) , we use the elementary bound @xmath515 which applied recursively gives @xmath516 the last term has the correct bound by assumption .
it remains to bound the ones appearing in the sum . on the event @xmath517
, there exist at least @xmath518 high @xmath519-branches @xmath520 , these are branches that satisfy @xmath521 for @xmath522 .
select the first @xmath518 such @xmath523-branches , and denote them by @xmath524 , for all @xmath525 .
consider the set @xmath526 , the children of @xmath527 at level @xmath528 : @xmath529 .
it holds @xmath530 where @xmath531 and @xmath532 .
a crucial point is that @xmath533 is not equal to @xmath534 since @xmath535 in general .
however , it turns out that their value must be very close since the variance of the difference is essentially a constant due to the logarithmic correlations .
precisely , let @xmath536 \\[-8pt ] \nonumber & & \hspace*{90pt}{}\ge\sqrt{2 } \nu\frac{(\overline{\gamma}-\gamma ' ) ( 1-\varsigma)}{k } \log n \biggr\}\end{aligned}\ ] ] for @xmath390 which is fixed and will be chosen small later . by lemma [ lem : lem12 ] of the , @xmath537 , for every @xmath538 , and any @xmath539 , @xmath540 such that @xmath541 .
therefore , a gaussian estimate ( see lemma [ lem : gaussian ] ) , together with the union - bound give @xmath542 for all @xmath538 and some @xmath543 .
it remains to bound the first term appearing in the sum of ( [ eqn : sum events ] ) .
on @xmath544 , @xmath545 can be replaced by @xmath546 in ( [ eqn : zeta ] ) , making a small error that depends on @xmath392 .
namely , one has @xmath547 , where @xmath548 note that conditionally on @xmath549 , the @xmath550 s are i.i.d .
moreover , since the @xmath551 s are independent of @xmath549 , they are also independent of each other .
lemma [ lem : ld ] of the guarantees that the sum of the @xmath552 can not be too low .
observe that @xmath553 = \p\biggl(z \ge\sqrt{2 } ( 1+\nu ) \frac { ( \overline{\gamma}-\gamma')(1-\varsigma)}{k } \log n \biggr),\ ] ] where @xmath554 is a centered gaussian with variance @xmath555 . by a gaussian estimate , lemma [ lem : gaussian ] , @xmath553 \ge\exp\biggl\ { -\frac{1}{k } \frac { ( 1 + 2\nu)^2 ( \overline{\gamma}-\gamma')^2 ( 1-\varsigma)^2}{\sigma^2 ( \alpha''-\alpha ' ) } \log n \biggr\},\ ] ] where @xmath556 has been replaced by @xmath557 to absorb the @xmath558 term in front of the exponential .
consequently , using elementary manipulations , @xmath559 \bigr)\\ & & \hspace*{38pt } { } \le \frac{2n_{\ell+1}}{n^{(\alpha''-\alpha')/k } } - n_\ell n^{- ( { 1}/{k})({(1 + 2\nu)^2 ( \overline{\gamma}-\gamma')^2 ( 1-\varsigma ) ^2}/{(\sigma^2 ( \alpha''-\alpha ' ) ) } ) } \biggr\ } \\ & & \qquad\subset \biggl\ { \biggl{\vert}\sum_{j=1}^{n_\ell } \bigl(\tilde{\zeta}_j - { \mathbb e}[\tilde{\zeta}_j ] \bigr ) \biggr{\vert}\ge\frac{1}{2 } n_\ell n^{-({1}/{k})({(1 + 2\nu)^2 ( \overline{\gamma}-\gamma')^2 ( 1-\varsigma)^2}/{(\sigma^2 ( \alpha ' ' -\alpha ' ) ) } ) } \biggr\},\end{aligned}\ ] ] provided @xmath560 that is @xmath561 fix @xmath392 small enough such that ( [ eq : rec5 ] ) is satisfied .
write for short @xmath562 then , taking @xmath563 and @xmath564 in lemma [ lem : ld ] , we get @xmath565 by the form of @xmath566 in ( [ eqn : nl ] ) , @xmath567 can be taken large enough so that @xmath568 for some @xmath569 and all @xmath570 .
this completes the proof of the lemma .
proof of lemma [ lem : init ] take @xmath571 in such a way that @xmath572 .
consider the set @xmath573 and the event @xmath574 the parameters @xmath575 , @xmath576 and @xmath577 will be chosen later as a function of @xmath436 . by splitting the probability on the event @xmath9 , @xmath578\\ & & \qquad\quad{}+\p \bigl(a^c\bigr),\end{aligned}\ ] ] where the second inequality is obtained by restricting to the set @xmath579 .
first we prove that the definition of @xmath9 yields a super - exponential decay of the first term for @xmath575 and @xmath576 depending on @xmath580 .
the variables @xmath581 , @xmath582 , are i.i.d .
gaussians of variance @xmath583 .
write for simplicity @xmath584 for i.i.d .
gaussians random variables with variance @xmath585 . a gaussian estimate
( see lemma [ lem : gaussian ] ) implies @xmath586 therefore @xmath587 \\ & & \qquad\leq\p\biggl(\sum_{i=1}^{n^\delta } \bigl ( \mathbf{1}_{\{z_i\geq\sigma _ 1 ( \alpha_0-\alpha')\log n\}}-\p\bigl(z_i\geq\sigma_1 \bigl(\alpha_0-\alpha'\bigr)\log
n \bigr ) \bigr)\\ & & \hspace*{166pt } { } \leq n^{\mathcal e_0}-n^{\delta-({2}/{3})(\alpha_0-\alpha ' ) } \biggr).\end{aligned}\ ] ] lemma [ lem : ld ] in the gives a super - exponential decay of the above probability for the choice @xmath588 and @xmath589 , for example , @xmath590 and @xmath591 .
it remains to show that @xmath592 has super - exponential decay .
we have @xmath593 the second term is easily shown to have the desired decay .
we focus on the first . on the event @xmath594 , @xmath595 since @xmath596
, it is easily checked that for @xmath597 , the above is smaller than @xmath598 .
therefore we choose @xmath577 such that @xmath599 .
finally the left - hand side of ( [ eqn : init ] ) is a gaussian random variable , whose variance is of order @xmath2 .
therefore the probability that it is smaller than @xmath600 is super - exponentially small .
this completes the proof of the lemma .
in this section , we compute the free energy of the perturbed models introduced in section [ sect : perturbed ] .
the free energy @xmath601 is shown to converge in probability to the claimed expression .
the @xmath172-convergence then follows from the fact that the variables @xmath602 are uniformly integrable .
this is a consequence of borell - tis inequality .
( another more specific approach used by capocaccia , cassandro and picco @xcite for the grem models could also have been applied here ; see section 3.1 in @xcite .
indeed , we clearly have @xmath603 therefore , uniform integrability follows if it is proved that @xmath604 $ ] is uniformly bounded .
@xmath605 \bigr)^2 \bigr ] + \frac{1 } { ( \log n ) ^2 } \e\bigl [ \max _ { x \in\mathcal x_{n } } y_x\bigr ] ^2.\ ] ] the first term is bounded by the borell - tis inequality ( see @xcite , page 50 ) @xmath606 which gives @xmath607}{\log n } \biggr)^2 \biggr ] \leq4 \int _ 0^\infty
r \mathrm{e}^{-{r^2}/{(2v_{12})}\log n } \,\d r.\ ] ] the right - hand side goes to zero for @xmath113 .
the term @xmath608 $ ] can be bounded uniformly by comparing with i.i.d .
centered gaussian random variables of variance @xmath609 and using slepian s inequality ; see , for example , @xcite , page 57 .
equivalently , one can reason as follows .
it is easily checked that the probability that the maximum be negative decreases exponentially with @xmath49 .
thus to control the second term it suffices to control @xmath610 it suffices to split the integral in two intervals : @xmath611 and@xmath612 .
the first integral divided by @xmath70 is evidently of order @xmath2 .
the second integral divided by @xmath70 tends to @xmath1 by a union bound and a gaussian estimate .
the almost - sure convergence is straightforward from the @xmath172-convergence and the almost - sure self - averaging property of the free energy @xmath613 this is a standard consequence of concentration of measure ( see @xcite , page 32 ) since the free energy is a lipschitz function of i.i.d .
gaussian variables of lipschitz constant smaller than @xmath614 .
( note that the @xmath347 s can be written as a linear combination of i.i.d .
standard gaussians with coefficients chosen to get the correct covariances . )
it remains to prove that the free energy @xmath615 converges in probability to the claimed expression in proposition [ prop : perturbed - free - energy ] . for fixed @xmath76 and @xmath390
, we prove that @xmath616 first , we introduce some notation and give a preliminary result . for simplicity
, we will write @xmath441 for @xmath617 throughout the proof . for any @xmath618 , consider the partition of @xmath619 $ ] into @xmath397 intervals @xmath620 , where the @xmath621 s are given by @xmath622 moreover for any @xmath623 , any @xmath624 and any @xmath625 , define the random variable @xmath626 and the events @xmath627 the next result is a straightforward consequence of propositions [ prop : perturbed - maximum ] and [ prop : perturbed - highpoints ] .
[ lem : thm3-bonenv ] for any @xmath624 and any @xmath569 , we have @xmath628 define the continuous function @xmath629.\ ] ] using the expression of @xmath441 in proposition [ prop : perturbed - highpoints ] on the different intervals , it is easily checked by differentiation that @xmath630 } p_{\beta } ( \gamma ) = f^{(\vec{\sigma } , \alpha)}(\beta).\ ] ] furthermore , the continuity of @xmath631 on @xmath632 $ ] yields @xmath633 } p_{\beta } ( \gamma)=f^{(\vec{\sigma } , \alpha)}(\beta ) , \qquad m \to\infty.\ ] ] fix @xmath624 large enough and @xmath569 small enough , such that @xmath634 note that for fixed @xmath397 , @xmath635 since @xmath636 is a decreasing function on @xmath632 $ ] .
_ proof of the lower bound _ ( [ eq : tm3-eq1 ] ) .
observe that the partition function@xmath637 associated with the perturbed model satisfies @xmath638 .
therefore on @xmath639 we get @xmath640 this yields on @xmath639 @xmath641 since for @xmath576 in ( [ eq : tm3-choicedelta ] ) @xmath642 the choices of @xmath397 , @xmath576 in ( [ eq : tm3-choicem ] ) and ( [ eq : tm3-choicedelta ] ) give that @xmath643 on @xmath639 for @xmath49 large enough .
therefore , ( [ eq : tm3-eq1 ] ) is a consequence of lemma [ lem : thm3-bonenv ] . _
proof of the upper bound _ ( [ eq : tm3-eq2 ] ) .
observe first that the partition function @xmath644 satisfies on @xmath639 @xmath645 the second term coming from the negative values of the field . since @xmath646 , on @xmath639 and for @xmath49 large enough
, we have using ( [ eq : tm3-choicedelta ] ) @xmath647 thus @xmath648 .
moreover , on @xmath639 the random variable @xmath649 are less than @xmath650 for all @xmath651 .
the two last observations imply by the choice of @xmath576 @xmath652 therefore , on the event @xmath639 , we get @xmath653 } p_{\beta}(\gamma ) + \frac { \sqrt{2 } \beta}{m } + \delta.\ ] ] recalling ( [ eq : sup = free ] ) and since @xmath654 , the choices of @xmath397 and @xmath576 in ( [ eq : tm3-choicem - bis ] ) and ( [ eq : tm3-choicedelta ] ) imply that @xmath655 on @xmath639 for @xmath49 large enough .
therefore ( [ eq : tm3-eq2 ] ) is a consequence of lemma [ lem : thm3-bonenv ] .
[ lem : gaussian ] let @xmath54 be a standard gaussian random variable . for any @xmath656 , we have @xmath657 [ lem : ld ] let @xmath658 be i.i.d .
real valued random variables satisfying @xmath659 = 0 $ ] , @xmath660 $ ] and @xmath661 . then for any , @xmath662 [ lem : gaussianibp ] let @xmath663 be a centered gaussian random vector .
then , for any @xmath664 function @xmath665 , of moderate growth at infinity , we have @xmath666 = \sum _ { i=1}^d { \mathbb e } [ x z_i ] { \mathbb e}\biggl [ \frac{\partial f}{\partial z_i}(z_1,\ldots , z_d ) \biggr].\ ] ] recall that @xmath57 , and @xmath667 . also by definition ,
@xmath668 .
it is clear that @xmath669=e[(\tilde x_x)^2]$ ] , which is the variance of the centered gaussian random variable @xmath670 .
this variance can be computed and equals @xmath671_{\varepsilon}^{\varepsilon^ { \alpha}}= ( 1-\alpha ) \log n.\ ] ] for the covariance , observe that @xmath672 $ ] is equal to the variance of the random variable @xmath673 . if @xmath674 [ i.e. , @xmath675 , then the subsets intersect in between the lines @xmath350 and @xmath676 , thus @xmath677 = \int _ { \ell}^{\varepsilon^ { \alpha } } \frac{y-\ell } { y^2 } \,\d y = [ \log y ] _ { \ell}^{\varepsilon^ { \alpha}}+ \ell \biggl [ \frac{1}{y } \biggr ] _ { \ell}^{\varepsilon^ { \alpha}}= \bigl(q \bigl(x , x'\bigr)- \alpha\bigr ) \log n + o_n(1).\ ] ] finally , if @xmath678 [ i.e. , @xmath679 , then the set @xmath680 is empty and thus @xmath672=0 $ ] .
the following lemma is a key tool to approximate the gaussian field we consider by a tree .
indeed the difference between the contribution to the gaussian field at a certain scale for two points that are close can be explicitly computed by integrating parallelograms ( see figure [ fig : intersect - lemma ] below ) and is shown to be small . [
lem : lem12 ] fix @xmath681 as in lemma [ lem : recurrence ] , @xmath22 such that @xmath682 and @xmath683 .
then for all @xmath684 such that @xmath685 , we have @xmath686 where @xmath687 denotes an upper bound for the @xmath688 s .
writing @xmath689 , we have @xmath690 which completes the proof of the lemma . .
the authors thank yan fyodorov , nicola kistler , irina kurkova and vincent vargas for helpful discussions . o. zindy would like to thank the courant institute of mathematical science and the universit de montral for hospitality and financial support . | we study the statistics of the extremes of a discrete gaussian field with logarithmic correlations at the level of the gibbs measure .
the model is defined on the periodic interval @xmath0 $ ] , and its correlation structure is nonhierarchical . it is based on a model introduced by bacry and muzy [ _ comm .
math .
phys . _
* 236 * ( 2003 ) 449475 ] ( see also barral and mandelbrot [ _ probab .
theory related fields _ * 124 * ( 2002 ) 409430 ] ) , and is similar to the logarithmic random energy model studied by carpentier and le doussal [ _ phys .
rev .
e _ ( 3 ) * 63 * ( 2001 ) 026110 ] and more recently by fyodorov and bouchaud [ _ j .
phys .
a _ * 41 * ( 2008 ) 372001 ] . at low temperature
, it is shown that the normalized covariance of two points sampled from the gibbs measure is either @xmath1 or @xmath2 .
this is used to prove that the joint distribution of the gibbs weights converges in a suitable sense to that of a poisson dirichlet variable . in particular , this proves a conjecture of carpentier and le doussal that the statistics of the extremes of the log - correlated field behave as those of i.i.d .
gaussian variables and of branching brownian motion at the level of the gibbs measure .
the method of proof is robust and is adaptable to other log - correlated gaussian fields . |
since almost all stars are observed as pinpoints of light , measurements of stellar variability correspond to the light from the integrated stellar disk .
consequently , changes due to fine patterns experience a severely reduced amplitude and may no longer be measurable .
photometrically , the nonradial pulsation modes of high - degree ( @xmath7 3 ) have been considered to be almost unobservable from the ground due to cancellation effects .
this simplistic view was challenged by @xcite , who argued that some of the multiple frequencies with small amplitudes detected during the multisite campaign of the @xmath0 sct fg vir @xcite could be high - degree modes . with the advent of photometric measurements from space and the resulting detection of amplitudes of 1 part - per - million ( ppm ) ,
high - degree modes need to be considered .
spectroscopically , high - degree modes can be detected through the changes of line profiles , e.g. , @xcite .
the method requires large amounts of observing time on medium - size telescopes .
it is presently applied to a number of @xmath0 sct stars ( hr 1170 , 4 cvn and ee cam ) , for which a number of high - degree modes are detected .
it is remarkable that with considerably fewer data , high - degree modes in @xmath0 sct stars have been seen before : the _ musicos _ multisite project @xcite presented evidence for the presence of high - degree modes in a number of @xmath0 sct stars such as @xmath8 peg @xcite .
they proposed explanations in terms of prograde sectoral modes of @xmath9 = @xmath10@xmath5@xmath10 up to 20 .
these modes would have frequencies that are about equal in the co - rotating frame of the star , leading to equidistant spacing in the observer s frame of reference . from photometry ,
prograde sectoral modes have recently also been suggested for the star rasalhague @xcite , which is also a rapidly rotating a star seen equator - on .
the authors report @xmath10@xmath5@xmath10 up to 7 . in this paper
we examine whether such modes can be seen in the extensive photometry of the rapidly rotating @xmath0 sct star kic 8054146 and relate the observational results to theoretical models .
both the observational and theoretical aspects are pioneering . in the modeling of this star
, one has to consider that rotation has a strong influence on the stellar structure .
the shape of a rapidly rotating star resembles that of an oblate spheroid rather than a sphere due to the effect of centrifugal force .
in fact , the flattening of some rapidly rotating stars could already be empirically measured by means of long - baseline interferometers ( for a review see * ? ? ?
the oblateness of the stellar surface leads to a latitude dependence of effective gravity and effective temperature .
taken together , these effects are termed gravity darkening @xcite .
this leads to the problem that the observed fundamental parameters in rapidly rotating stars , such as effective temperature , gravity and luminosity are a function of aspect , where the angle @xmath11 between the line - of - sight and the stellar rotation axis is often unknown .
the effects of the stellar inclination angle on the effective temperature and luminosity of rotating stellar models were examined , e.g. , by @xcite .
they found a strong dependence of the fundamental parameters on the aspect .
moreover , it was found by @xcite that for rapid rotation the line - widths become insensitive to rotation , and hence the projected rotational velocity @xmath12 is underestimated .
while the authors found these results for rapidly rotating be stars , the same problem may also apply to a stars .
rotation also has an impact on pulsation .
the oscillation cavities are affected by the changes in stellar structure ( i.e. , the deformation of the star and the decreased mean density ) . moreover ,
the impact of centrifugal and coriolis force on the oscillations increases with the rotation rate .
rotation , in particular , affects acoustic modes propagating in the envelope @xcite . in the interpretation of observed frequencies in rapidly rotating stars another problem
arises because of the effect of rotational splitting ( e.g. , * ? ? ?
* ) . due to the transformation from the co - rotating system ( star s system ) to the observer s system ,
the mode frequencies are shifted by @xmath13 .
for example , for an a type star with an equatorial rotation velocity of 300 km s@xmath2 , the angular rotation frequency @xmath14 ranges around 2.8 cycles day@xmath2 . usually , @xmath5 is not known a priori and , therefore , a large uncertainty about the intrinsic mode frequency ( and its nature ) in the co - rotating system of the star persists . in specific cases ,
however , hints on @xmath5 may be derived from regular patterns .
the rapidly rotating a - star kic 8054146 with its intriguing frequency pattern represents such a case .
the @xmath0 sct star kic 8054146 was studied extensively with the _ kepler _ spacecraft .
7.3 quarters of short - cadence data ( one measurement per 58.8s ) and 11 quarters of long - cadence ( two measurements per hour ) are available .
the data were analyzed by @xcite .
the 349 statistically significant frequencies have amplitudes as low as 2 ppm .
one of the most striking results is that several regularities are present in the values of the frequencies .
we can rule out the possibility that these regularities are accidental agreements because of the high frequency resolution and large number of exact agreements , as shown in figure 4 of @xcite . in particular , the frequency patterns seen in the pressure - mode frequency regions ( frequencies in the 6 to 200 cycles day@xmath2 range ) are related to the dominant low - frequency peaks , interpreted as rotation and gravity modes .
an example is the low - frequency peak near 2.81421 cycles day@xmath2 , which shows up with an almost identical value in a number of frequency sequences up to 160 cycles day@xmath2 .
these sequences , hereafter referred to as s , are the subject of the present investigation .
we illustrate this in the left panel of figure [ fig : echelle ] , which is an echelle diagram of the detected frequencies from paper i , folded with the frequency of 2.81421 cycles day@xmath2 .
although the diagram examines the frequency peaks in the 6 to 200 cycles day@xmath2 range , a few dominant low frequencies were also included in order to match the observed patterns .
rlrrrr rlrrrr & & & + cd@xmath2 & i d & & cd@xmath2 & i d & + & & q2-q11&q2 & q5/6 & q7/8 & & & q2-q11&q2 & q5/6 & q7/8 + & & @xmath15 & @xmath16 & @xmath17 & @xmath17 & & & @xmath15 & @xmath16 & @xmath17 & @xmath17
+ & & + 73.2122 & @xmath18 & 3.3 & 3.0 & 2.8 & 3.4 & 9.8444 & @xmath18:-@xmath19 & & 5.4 & 5.6 & 2.5 + 76.0223 & @xmath20 & 10.8 & 23.1 & 15.5 & 14.0 & 12.6546 & @xmath20-@xmath19 & & 8.6 & 5.0 & 2.5 + 78.8346 & @xmath21 & 4.0 & 3.2 & 3.0 & 4.0 & 15.4668 & @xmath21-@xmath19 & & 5.7 & 7.1 & 2.6 + 81.6487 & @xmath22 & 11.5 & 27.1 & 12.1 & 10.3 & 18.2807 & @xmath22-@xmath19 & & 6.8 & 3.6 & 1.7 + 84.4635 & @xmath23 & 6.6 & 1.4 & 6.9 & 5.4 & 21.0957 & @xmath23-@xmath19 & & 5.0 & 12.1 & 7.0 + 87.2785 & @xmath24 & 10.1 & 26.0 & 17.9 & 8.2 & 23.9118 & @xmath24-@xmath19 & & 8.4 & 5.1 & 1.5 + 90.0931 & @xmath25 & 2.7 & 8.3 & 3.8 & 1.3 & 29.5403 & @xmath26-@xmath19 & & 9.9 & 3.6 & 1.6 + 92.9080 & @xmath26 & 6.7 & 33.4 & 12.6 & 9.7 & 32.3557 & @xmath27-@xmath19 & & 14.1 & 4.8 & 2.8 + 95.7233 & @xmath27 & 2.8 & 8.0 & 2.6 & 1.1 & 142.2021 & @xmath21+@xmath19 & & 6.7 & 2.3 & 1.8 + 98.5356 & @xmath28 & 4.1 & 15.3 & 8.8 & 2.0 & 147.8311 & @xmath23+@xmath19 & & 1.5 & 4.2 & 1.7 + 101.3484 & @xmath29 & 2.0 & 4.0 & 2.7 & 1.5 & & & & & & + & & + 6.9143 & @xmath18:-@xmath30 & & 0.9 & 4.1 & 4.3 & 15.5875 & @xmath20-@xmath31 & & 4.1 & 0.9 & 2.9 + 9.7246 & @xmath20-@xmath30 & & 17.0 & 5.9 & 5.8 & 18.4043 & @xmath21-@xmath31 & & 3.6 & 4.2 & 4.5 + 12.5366 & @xmath21-@xmath30 & & 5.2 & 5.3 & 5.1 & 26.8445 & @xmath24-@xmath31 & & 2.9 & 2.3 & 3.4 + 15.3514 & @xmath22-@xmath30 & & 20.1 & 8.7 & 4.0 & 32.4739 & @xmath26-@xmath31 & & 5.9 & 3.4 & 1.9 + 18.1659 & @xmath23-@xmath30 & & 4.6 & 6.4 & 5.0 & & & & & & + 20.9814 & @xmath24-@xmath30 & & 20.5 & 8.6 & 2.5 & + 23.7879 & @xmath25-@xmath30 & & 14.2 & 10.2 & 10.6 & & & & & & + 26.6104 & @xmath26-@xmath30 & & 32.5 & 6.9 & 5.5 & 34.0188 & @xmath24-@xmath32 & & 0.0 & 5.8 & 2.7 + 29.4306 & @xmath27-@xmath30 & & 13.7 & 20.8 & 17.9 & 39.6483 & @xmath26-@xmath32 & & 11.2 & 3.1 & 2.4 + 32.2390 & @xmath28-@xmath30 & & 16.5 & 6.5 & 3.4 & & & & & & + 147.9467 & @xmath22+@xmath30 & & 7.3 & 2.2 & 3.1 & + 159.2061 & @xmath26+@xmath30 & & 12.4 & 1.9 & 1.7 & & & & & & + & & & & & & 28.5565 & @xmath21-@xmath33 & & 2.8 & 3.5 & 5.1 + & & 34.1854 & @xmath23-@xmath33 & & 0.0 & 8.9 & 10.7 + 2.8142 & @xmath34 & & 117.8 & 18.3 & 12.5 & 39.8156 & @xmath25-@xmath33 & & 3.4 & 5.6 & 3.4 + & & & & & & 45.4447 & @xmath27-@xmath33 & & 1.8 & 3.9 & 3.6 + & & & & & & 51.0704 & @xmath29-@xmath33 & & 1.3 & 3.0 & 2.2 + & & & & & & 129.1130 & @xmath21+@xmath33 & & 0.8 & 2.3 & 3.5 + & & & & & & 134.7418 & @xmath23+@xmath33 & & 0.5 & 3.8 & 3.9 + identification : @xmath24 is the frequency with @xmath5 = 30 in the formula , @xmath35 = 2.8519 + @xmath5 * 2.81421 cycles day@xmath2 .
the numbering scheme of the frequencies , @xmath36 , corresponds to that of paper i.
in the data from quarter 2 of the @xmath38 mission ( hereafter called q2 ) , a sequence of high frequencies appear to be high multiples of a low frequency at 2.814 cycles day@xmath2 with a small zero - point offset .
this discovery is made possible by the relatively large amplitudes associated with the low as well as the high frequencies .
the additional data from q5 to q11 confirm the relationship at lower amplitudes , and provide a considerably increased frequency resolution .
this , in turn , confirms that the high frequencies are not exact multiples of a low frequency , but are equidistantly spaced in frequency with a significant zero - point offset .
table 1 lists the frequencies identified with the basic s sequence and its combinations with the more dominant ( other ) pulsation modes.this also shows that the amplitudes vary in a systematic way from quarter to quarter . to achieve the highest precision ,
we have also calculated a common solution with all the data from q2 to q11 , covering more than two years .
this leads to a high statistical significance for our detections ( @xmath39 4 standard deviations ) . for the s - sequence combinations with the other , dominant modes , however , we are limited by the small , systematic frequency drifts of the dominant pulsation modes . since the s sequence produces side lobes to these pulsation modes
, the frequencies of the side lobes also change . for these a common solution
is not possible ; consequently , we only list the separate solutions up to q7/q8 . term .
the formal error bars are equal to the size of the symbols .
the observed frequencies are equidistant with deviations of only 0.001 cycles day@xmath2 .
the spacing corresponds to a dominant low frequency of 2.81421 cycles day@xmath2 .
the diagram also shows that the frequencies have a significant zero - point offset , @xmath40 . ]
the existence of a significant zero - point offset in frequency is demonstrated in figure [ fig:2 ] , where the exact - multiple hypothesis would lead to systematic frequency deviations .
we have calculated an optimum least - squares fit from the observed frequencies using the weighting from the observed q2-q11 amplitudes .
this fit in cycles day@xmath2 is : @xmath41 where @xmath5 are integers from 25 to 35 .
figure [ fig:2 ] shows the excellent fit of this equation ( dashed line ) with an average residual per single point of only 0.0014 cycles day@xmath2 . for the pulsation frequency @xmath30 at 66.298 cycles
day@xmath2 , the side lobes , displaced by @xmath37 , range from 6.914 to 159.206 cycles day@xmath2 .
the various patterns are illustrated in the right panel of figure [ fig : echelle ] .
these agreements and identifications can not be accidental because of the high frequency resolution ( @xmath42 0.0002 cycles day@xmath2 ) of the data set .
in fact , we can use the frequencies of these side lobes to check the coefficients derived above .
for the side lobes only we obtain @xmath43 where @xmath5 once again are integers from 25 to 35 .
this independent result provides a remarkable agreement with the coefficients derived above from the s sequence .
the offset of 2.8519 cycles day@xmath2 vs. a spacing of 2.81421 cycles day@xmath2 has important astrophysical implications , which will be explored below .
however , we can already reject a simple hypothesis , i.e. , that the s sequence is a fourier series of a non - sinusoidal light - curve shape of @xmath34 . in this hypothesis , exact spacing
is predicted , as observed .
an offset , however , is not possible .
moreover , the additional @xmath40 term also rules against an interpretation of the s sequence as rotational variability caused by small photometric asphericities on the star .
an important clue to the origin of the sequence is provided by the odd - even sequence of the amplitudes . in figure
[ fig : oddeven ] we show the amplitudes for the different multiples of @xmath5 for three different time spans .
these were selected for the following reasons : q2 has the largest amplitudes , during q5 the large amplitudes have already decreased , and from q6 to q11 all the peaks are present with small amplitudes , requiring a long data set for an optimum signal - to - noise ratio .
the diagram demonstrates an odd - even variation of the amplitudes , with high amplitudes occurring only for even values of @xmath5 .
the amplitudes for the odd @xmath5 peaks are only a few ppm .
we note here that despite the small amplitudes the detection of these frequencies is statistically significant due to the low noise of the @xmath38 data , especially at high frequencies . for the s sequence and the side lobes of a dominant frequency at 66.3 cycles day@xmath2 .
from q2 and q8 , all amplitudes associated with the s sequences as well as that of the related low frequency at 2.81421 cycles day@xmath2 , decreased in size .
the figure demonstrates the systematic amplitude differences between odd and even values of @xmath5 . ]
in this hypothesis , the amplitude differences between odd and even @xmath5 values can be explained by alternating @xmath9 = 1 and @xmath44 modes and their different amplitudes due to partial light cancellation over the stellar disk .
this has already been detected in another @xmath0 sct star hd 144277 @xcite .
this would predict a value of the large spacing to be twice the observing frequency spacing , i.e. , near 5.6 cycles day@xmath2 .
while such a spacing is within the uncertainties of the known physical parameters of kic 8054146 , present modeling of the observed frequencies shows that the observed frequencies are much too equidistant ( as already discussed in paper i ) .
recent analysis have shown that a large fraction of a / f stars show a rotational modulation of their light output ( see @xcite , @xcite ) .
this asymmetry could be caused by spots . at first sight
, the spot hypothesis appears very attractive to explain the s sequence .
the reason is that small surface features lead to nonsinusoidal light curves : the equidistantly spaced high frequencies would simply be the higher harmonics of a fourier series .
however , there exist serious problems with this hypothesis .
the high frequencies are not exact multiples of the observed low frequency . in fact
, they are not multiples of _ any _ one or two low frequencies and the measured offset is statistically secure .
the frequencies , therefore , are not the result of a fourier decomposition of an extremely nonsinusoidal light curve .
we have confirmed the mathematical argument by attempting to model simple spots : these models failed to give any consistent light curve .
differential rotation , which may occur at different latitudes , causes two observationally detectable effects .
if the ( hypothetical ) spot has a lifetime of years and is drifting in latitude , then its rotation frequency changes .
this is not observed : over two years of observation the frequencies of the s sequence are constant to better than one part in 10@xmath45 .
furthermore , the observed widths of the fourier peaks agree with constant frequency values .
the lack of additional width or multiple close peaks also argue against multiple spots at slightly different latitudes .
we conclude that the s sequence can not be explained by ( simple ) spots .
an alternative explanation for the odd - even dichotomy of the s sequence amplitudes can be derived based on the dependence of mode visibility on the surface geometry of the pulsation mode .
it is well known that even-@xmath9 modes have a higher visibility than odd-@xmath9 modes , because odd degree modes suffer from stronger averaging over the visible disc @xcite .
consequently , we may interpret the s sequence frequencies as prograde sectoral modes of high spherical degree . due to the rather high @xmath5-value
, the mode frequencies in the star s system are very low and can be associated to high - order g - modes . in the case of kic 8054146 ,
an equator - on view is more likely . with a projected rotation velocity ( @xmath46 ) of approximately 300 km s@xmath2
, the keplerian break - up rate already amounts to @xmath470.7 for an equator - on view ( @xmath48 ) .
if we decrease the inclination angle , @xmath11 , the star soon reaches the critical rotation rate , which makes a low inclination angle unlikely . a near equator - on view results in a low visibility of modes for which @xmath49 is an odd number ( i.e. , north - south asymmetry ) . in all these cases , there is a node line at the rotation equator and if the star is seen equator - on , the integrated light variation across the surface is diminished by the opposite behavior of the variations of the two sides of node line .
the amplitudes fall below the limit of detectability .
an equator - on view in particular favours the visibility of sectoral modes ( @xmath9=@xmath50 ) . due to the mirror effect at zero frequency in the fourier spectra
, both prograde and retrograde modes may populate the same frequency range .
however , for rapidly rotating b stars aprilia et al .
( 2011 ) found that retrograde modes suffer a strong damping while prograde sectoral modes are exceptionally weakly damped . due to the pulsational similarities of a and b stars ,
the argument can be extended to @xmath0 sct stars .
consequently , within the family of sectoral modes , we restrict ourselves to prograde modes . in some sense
our case is similar to the case of rasalhague @xcite , which is also a rapidly rotating a star seen equator - on . in this
star sequences of frequencies following the same law ( i.e. , @xmath40+@xmath13 ) have been observed and were interpreted as high - order g - modes . for rasalhague @xmath40 is very small ( @xmath470.1 - 0.3 cycles day@xmath2 ) , which leads the authors to the assumption of observing dispersion - free low - frequency modes , which they later specify to be equatorial kelvin modes .
these modes are prograde sectoral modes confined to an equatorial waveguide . in the case of kic 8054146 we chose @xmath51
cycles day@xmath2 instead of @xmath52 cycles day@xmath2 , which would define @xmath5 to be higher by one integer than in the other case .
this , however , would associate the frequencies with highest amplitudes to odd-@xmath5 modes .
a requirement for this interpretation would be an inclination angle of significantly less than 90 degrees , which is ruled out due to the break - up rate as detailed already . in addition , there are other important differences between rasalhague and our case .
we do not observe several modes per @xmath5 value , but only one .
furthermore , our frequencies in the observer s system are significantly higher , with much higher @xmath5-values in the range of @xmath5 = 25 - 35 .
consequently , the corresponding surface pattern is characterized by bright and dark sectors .
for example , the angle between two subsequent node lines on the stellar surface amounts to 12@xmath53 for a sectoral mode of @xmath54 .
in the reference system of the star the frequencies of those g - modes causing the s sequence are comparable to the rotation frequency . to obtain mode properties required for the computation of mode visibility in this low - frequency domain , we adopt the so - called traditional approximation ( hereafter : ta ) to describe the effect of rotation on oscillation modes . in the ta ,
terms proportional to the cosine of latitude ( @xmath55 ) are neglected to allow for the separation of radial and horizontal variables . instead of spherical harmonics ,
the latitude - dependence of the eigenfunctions is given by hough functions , @xmath56 , which are the solutions of laplace s tidal equations .
the ta is a standard recipe in geophysics ( see * ? ? ?
* and references therein ) and also gained popularity in studies of low - frequency g - modes in main sequence stars @xcite .
we should note , however , that one of the requirements of the ta is that the departure from sphericity is small , and consequently is not valid in rapidly rotating stars such as kic 8054146 .
therefore , only qualitative conclusions can be made with the ta in the given case .
investigations on the instability of low - frequency g modes within the ta were performed by @xcite , @xcite and @xcite .
the stability of these so - called slow modes is determined mainly by the separation parameter , @xmath57 , and the angular frequency , @xmath58 .
comparisons to methods that avoid the ta were made by @xcite and @xcite .
@xcite used a 2d approach instead of the ta and finds more prograde modes unstable in comparison with @xcite .
@xcite examined the effects of mode coupling on the instability in a spb model and found that , in contrast to zonal and tesseral modes , low - degree prograde sectoral modes in rapidly rotating b stars are almost not affected by damping due to the different dependence of their frequencies on @xmath59 .
this also holds true in the case of high - degree modes and a stars .
the visibility of slow modes based on light amplitudes was studied by @xcite .
@xcite and @xcite extended these considerations to rv amplitudes and examined the prospects of mode identification .
@xcite examined the important effect of non - linear mode coupling on the visibility of slow modes .
however , these authors only examined low - degree modes . in this study
we examine high - degree modes @xcite and neglect mode coupling effects .
our study is based on codes developed by @xcite and @xcite .
the codes were slightly modified to allow for the calculation of high-@xmath9 modes .
the general approach is as follows : instead of the @xmath60 term we have a separation parameter @xmath57 , which is determined by solving laplace s tidal equation for a certain spin parameter , @xmath61 , azimuthal order , @xmath5 , and taking into account boundary conditions at @xmath62 and @xmath63 .
our code uses a relaxation technique and , consequently , requires a start value of @xmath9 . in the limit case of @xmath64 ( i.e. , no rotation ) , @xmath65 and the hough function
is described by a single associated legendre polynomial except for normalization . for convenience , we label modes with the @xmath9-value corresponding to @xmath57(s=0 ) for the remaining paper .
linear non - adiabatic calculations yield the complex f - parameter for each mode .
this parameter relates the bolometric flux perturbation with the radial displacement and depends on m , @xmath57 , @xmath58 and the stellar model . * formally , in the non - adiabatic case , @xmath57 is a complex quantity ( as are the frequency and the spin parameter ) .
however , in @xmath0 scuti stars the nonadiabaticity of pulsations is very weak , so that the imaginary part of the frequency is very small relative to the real part .
therefore , in the expression for the spin parameter , @xmath66 , we use only the adiabatic ( real ) part of the frequency and consider @xmath66 and @xmath57(s ) as real numbers .
the same approach was adopted by dziembowski et al .
( 2007 ) : in the adiabatic part of the pulsation code we interpolate the value of @xmath57(s ) from independently computed solutions of the laplace tidal equations , and adopt this value for the non - adiabatic part . *
the distortion of the surface is taken into account when computing the limb - darkening of the visible hemisphere as described in @xcite .
our calculation is based on a nonlinear law of @xcite as detailed in the appendix of @xcite .
the wavelength - dependent flux derivatives and limb darkening coefficients are computed from a static atmosphere model . the visibility of g -
modes with intrinsic frequencies close to the rotation frequency was first examined by @xcite , based on a truncated expansion of @xmath56 into associated legendre functions . @xcite and @xcite instead performed a two - dimensional integration over the visible hemisphere .
the latter approach is also used in our study .
there still are two uncertainties remaining with our chosen approach : the influence of convection on the frequencies of @xmath0 sct stars and our use of a plane - parallel atmosphere model for such high @xmath9 values .
these might restrict us to qualitative , rather than detailed quantitative , conclusions .
our subsequent calculations are based on a main sequence model computed with the warsaw - new jersey stellar evolutionary code , which was described in @xcite .
this code evolves one - dimensional star models taking into account uniform rotation by accounting for the effect of the horizontally averaged centrifugal force on the stellar structure .
the examined model has a mass of 1.80 m@xmath67 and an equatorial rotational velocity of 272 km s@xmath2 .
this rotation rate is less than 2@xmath68 lower than the measured average @xmath46 and its choice will be detailed later .
the standard solar composition according to the @xcite mixture is adopted as well as rather effective convection ( @xmath69 ) . since the oscillations of the given modes are confined to the equatorial region , the effective temperature of the model
was chosen to resemble the conditions at the equator , which is slightly cooler than the temperature derived from the integrated disk since it also contains parts of the hotter poles .
table [ tab : model ] summarizes the detailed parameters of the equilibrium model .
.parameters of the examined equilibrium model [ cols="<,<,<,<,<,<",options="header " , ] we computed @xmath57 eigenvalues for different spin parameters for prograde sectoral modes with @xmath5 = 140 .
the results for @xmath57 as a function of spin were also checked by comparing them to the asymptotic formulae given by @xcite , which showed good agreement .
high - order prograde sectoral g - modes were then computed using a standard oscillation code by replacing @xmath60 with @xmath57 .
the results of these computations for the given model are shown in figure [ fig : smopt ] .
the normalized growth rate , @xmath70 , is plotted against the frequency as observed in the inertial system of the observer and compared to the detected s sequence frequencies .
full driving corresponds to @xmath71 , while full damping corresponds to @xmath72 . vs. frequency ( in observer s system ) for prograde sectoral high - order g - modes ( black crosses ) .
modes with the same @xmath9 are connected by a line and for a subset the corresponding @xmath9 value is denoted .
the observed s sequence frequencies are shown with vertical red lines and line lengths proportional to the detected amplitudes . ]
it can be seen that for each @xmath9 there is a maximum in @xmath73 .
the frequency of the maximum mainly depends on the rotation velocity . * consequently , to match the maxima in @xmath73 with the observed 2.814 c / d frequency separation pattern of the s sequence , we adjusted the equatorial rotational velocity to 272 km s@xmath2 which corresponds to a rotation frequency of 2.56 c / d ( see figure [ fig : smopt ] ) .
this is no disagreement because the observed frequency separations in the s sequence do not correspond to rotational splitting of modes with equal spherical degrees .
in addition to the excellent fit of the observed frequency spacing , the even - odd pattern is correctly assigned as well .
* these maxima in normalized growth rate , @xmath73 , have a similar nature as the low - frequency maxima in @xmath73 in b stars ( which cause spb type pulsation ) . in a stars , however , this bump is not as prominent and usually no mode gains sufficient driving .
however , as we see in figure [ fig : smopt ] for certain high-@xmath9 modes it is higher than for low degree modes .
the @xmath73-parameter is close to 0.0 at its maximum ( for @xmath74=1.8 and the given mass ) .
taking into account uncertainties of our simplistic model , both in its global parameters and in description of convection and opacities in the envelope , it may be possible to obtain marginally unstable modes .
if we consider for each @xmath5 value only the mode with maximum @xmath73 , we also see that we have a local maximum at @xmath5=30 .
the reason for this curved shape is illustrated in figure [ fig : wcum ] , which shows the contribution of different stellar layers to driving and damping for the mode with highest @xmath73 for the cases
@xmath9=20 , 30 and 40 .
it can be seen that damping near the z opacity bump ( at @xmath755.3 ) and differences in the contribution to driving from the and / partial ionization zone ( at @xmath754.65 and 4.1 , respectively ) are evident for the different modes . but with different @xmath9 .
] since this is only a qualitative examination we refrain from listing the exact mode properties in a table . generally , the parameters of the modes which correspond to the observed counterparts are as follows : the spin parameter ranges from s=0.55 ( at @xmath5=25 ) to s=0.44 ( at @xmath5=35 ) and * the nondimensional frequency , @xmath76 , from 1.24 to 1.53 .
moreover , the radial orders span the range from n=54 ( at @xmath5=25 ) to n=60 ( at @xmath5=35 ) . *
a potential problem remains : for each @xmath9 , we observe only one frequency , but several modes of subsequent radial order are predicted .
why is only one mode excited ?
formally , all these modes are stable , but due to @xmath73 being close to 0 , only a minor contribution to driving is sufficient to excite these modes .
consequently , it may be possible that with increased driving only the modes at maxima become marginally unstable , forming the given equidistant pattern .
we computed the visibility for the subset of modes , which is located at maxima in @xmath73 .
due to lack of atmosphere data for the kepler band we computed the results for the geneva @xmath77 band .
the results are shown in figure [ fig : visibility ] . to obtain similar amplitudes as observed , the intrinsic mode amplitude , @xmath78 ,
was adjusted to 0.0001 , which implies rather small intrinsic amplitudes . in the range between @xmath5=25 and 35 , the ratio between amplitudes of even-@xmath9 to odd-@xmath9 modes
is @xmath472.5 , which is similar to the observations .
moreover , there is only a small general decrease of the given alternating pattern towards higher @xmath9 values , significantly increasing the probability of detecting high - degree modes .
in this paper , we have examined a number of equidistantly spaced frequency series , s , in kic 8054146 with a spacing of 2.814 cycles day@xmath2 .
these are interpreted in terms of prograde sectoral modes as well as a small amplitude modulation of the dominant pressure modes .
the basic sequence is found to be multiples of 2.814 cycles day@xmath2 ( up to 35 ) with a significant zero - point offset .
the spacing and the existance of an offset are similar to the behavior of the reported prograde sectoral mode patterns detected in rasalhague by @xcite , albeit with smaller multiples .
we note again that the equidistant spacing is caused by the observer s frame of reference . for kic 8054146 ,
the measured projected rotational velocity of 300 km s@xmath2 and the corresponding fraction of the keplerian break - up velocity of 0.7 indicate a higher probability of a near equator - on view .
in such a geometrical configuration modes with north - south symmetry , in particular sectoral modes , have the highest chances to be detected .
as detailed in section 4.3 , prograde modes are the most viable candidates among sectoral modes .
the observed odd - even dichotomy of the amplitudes of the s - sequence frequencies can be explained by associating these frequencies to odd and even azimuthal order respectively .
we computed the visibility of prograde sectoral modes with a formalism based on the ta . assuming a fixed intrinsic amplitude for all modes
, we find that the computed amplitudes of prograde sectoral g - modes do not decrease quickly in an equator - on view .
hence , it becomes possible to detect them .
the frequencies and amplitudes of the s sequence can be approximately fitted with a 1.8 solar mass model , which rotates with an equatorial rotation velocity of approximately 272 km s@xmath2 .
a remaining uncertainty with our interpretation is that the modes with the highest normalized growth - rates are still predicted to be stable .
however , the possible instability of these modes is within the uncertainties of the given stellar equilibrium model in terms of opacities , chemical composition , efficiency of envelope convection or stellar mixing processes in general , and also interactions between pulsation and rotation .
generally , if our hypothesis is correct , it would require marginal instability only for the modes at maxima in the normalized growth rate , i.e. , with a specific radial order , while the other modes surrounding this maximum are stable .
this is a very sharp criterion and hence has a low probability of being observed .
consequently , another selection mechanism might also be involved .
although for kic 8054146 some open questions remain , our qualitative theoretical study shows evidence that high - degree prograde sectoral modes may be driven and observed in @xmath0 scuti stars .
it is a pleasure to thank richard townsend , wojtek dziembowski , and luis balona for helpful discussions .
this investigation has been supported by the austrian fonds zur frderung der wissenschaftlichen forschung through project p21830-n16 .
pl and aap acknowledge partial financial support from the polish ncn grants 2011/01/b / st9/05448 and 2011/01/m / st9/05914 .
the authors also wish to thank the _ kepler _ team for their outstanding efforts which have made these results possible .
funding for the _ kepler _ mission is provided by nasa s science mission directorate . , lee , u. & saio , h. 2011 , , 412 , 2265 asplund , m. , grevesse , n. , sauval , a. j. & scott , p. 2009
, ara&a , 47 , 481 balona , l. a. 2011 , , 415 , 1691 , l. a. & dziembowski , w. a. 1999 ,
mnras , 309 , 221 breger , m. , lenz , p. , antoci , v. , et al .
2005 , a&a , 435 , 955 breger , m. , balona , l. a. , lenz , p. , et al .
2011 , , 414 , 1721 breger , m. , fossati , l. , balona , l. , et al .
2012 , , 759.62 ( paper i ) claret , a. 2000 , , 363 , 1081 2002 , , 392 , 151 daszyska - daszkiewicz , j. , dziembowski , w. a. & pamyatnykh , a. a. 2006 , , 77 , 113 daszyska - daszkiewicz , j. , dziembowski , w. a. & pamyatnykh , a. a. 2007 , aca , 57 , 11 dziembowski , w. a. 1977 , aca , 27 , 203 dziembowski , w. a. , daszyska - daszkiewicz , j. & pamyatnykh , a. a. 2007 , mnras , 374 , 248 espinosa lara , f. & rieutord , m. 2011 , a&a , 533 , 43 , t. , zimmerman , j. t. f. , maas , l. r. m. & van haren , h. 2008 , reviews of geophysics , 46 , 2004 gillich , a. , deupree , r. g. , lovekin , c. c. , short , c. i. & toqu , n. 2008 , apj , 683 , 441 kennelly , e. j. , walker , g. a. h. , catala , c. et al .
1996 , a&a , 313 , 571 kennelly , e. j. , brown , t. m. , kotak , r. , et al .
1998 , , 495 , 440 , u. & saio , h. 1997 , , 491 , 839 , u. 2006 , , 365 , 677 , u. 2012 , , 420 , 2387 lignires , f. & georgeot , b. 2009 , , 500 , 1173 , j. d. , townsend , r. h. d. , che , x. , et al .
2010 , , 725 , 1192 pamyatnykh , a. a. 2003 , ap&ss , 284 , 97 pamyatnykh , a. a. , dziembowski , w. a. , handler , g. & pikall , h. 1998 , a&a , 333 , 141 , g. j. 2005 , , 443 , 557 , g. j. 2007 , , 469 , 1057 townsend , r. h. d. 2003a , , 340 , 1020 townsend , r. h. d. 2003b , , 343 , 125 townsend , r. h. d. , owocki , s. p. & howarth , i. d. 2004 , mnras , 350 , 189 townsend , r. h. d. 2005a , , 360 , 465 townsend , r. h. d. 2005b , , 364 , 573 , g. t. 2012 , , 20 , 51 zima , w. 2006 , a&a , 455 , 227 zwintz , k. , lenz , p. , breger , m. , et al .
2011 , a&a , 533 , 133 | this paper examines the 46 frequencies found in the @xmath0 sct star kic 8054146 involving a frequency spacing of @xmath1 2.814 cycles day@xmath2 ( 32.57 @xmath3hz ) , which is also a dominant low - frequency peak near or equal to the rotational frequency .
these 46 frequencies range up to 146 cycles day@xmath2 .
three years of _ kepler _ data reveal distinct sequences of these equidistantly spaced frequencies , including the basic sequence and side lobes associated with other dominant modes ( i.e. , small amplitude modulations ) .
the amplitudes of the basic sequence show a high - low pattern .
the basic sequence follows the equation @xmath4 cycles day@xmath2 with @xmath5 ranging from 25 to 35 .
the zero - point offset and the lack of low - order harmonics eliminate an interpretation in terms of a fourier series of a non - sinusoidal light curve .
the exactness of the spacing eliminates high - order asymptotic pulsation .
the frequency pattern is not compatible with simple hypotheses involving single or multiple spots , even with differential rotation .
the basic high - frequency sequence is interpreted in terms of prograde sectoral modes .
these can be marginally unstable , while their corresponding low - degree counterparts are stable due to stronger damping .
the measured projected rotation velocity ( 300 km s@xmath2 ) indicates that the star rotates with @xmath6 70% of the keplerian break - up velocity .
this suggests a near equator - on view .
we qualitatively examine the visibility of prograde sectoral high - degree g - modes in integrated photometric light in such a geometrical configuration and find that prograde sectoral modes can reproduce the frequencies and the odd - even amplitude pattern of the high - frequency sequence . |
interest in the outskirts of spiral galaxy disks has increased over recent years as these regions are the site of the most recent acquisition of gas for these systems ( e.g. * ? ? ?
* ) , as well as low - level star - formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for recent results ) , making these faint outskirts the interface between the island universes the galaxies themselves and the cosmic web of primordial gas .
the low - level star - formation was first discovered in h@xmath4 emission by @xcite and @xcite . after the launch of the galaxy evolution explorer ( galex , * ? ? ?
* ) , initial anecdotal evidence pointed to ultraviolet disks of spiral galaxies extending much beyond their optical radius @xcite .
subsequent structural searches for these extended ultraviolet ( ) disks by @xcite and @xcite find that some 2030% of spirals posses an disk and 40% of s0s @xcite , making this type of disk common but not typical for spiral and s0 galaxies .
these disk complexes are generally @xmath5 myr old , explaining why most lack h@xmath4 @xcite , as opposed to a top - light imf ( as proposed by * ? ? ?
* ) , and sub - solar but not excessively low metallicities ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* @xmath6 , based on emission lines ) .
additionally , it has been known for some time now that atomic hydrogen ( ) as observed by the 21 cm fine structure line also extends well beyond the optical disk of spiral galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in those few cases where both high - quality and deep galex data are available , a close relation in their respective morphology was remarked upon @xcite .
while we have to wait for the all - sky surveys in to catch up to the coverage of the galex surveys ( e.g. , the wide survey with wsrt / apertif or the wallaby survey with askap ) , we can compare the morphology for those galaxies for which uniform information is available .
the canonical view of the origin of the disks , is that the kennicutt - schmidt law @xcite needs to be extended to low global surface densities of gas and the formation of individual o - stars in the very outskirts of disks @xcite .
the recent accretion of cold gas flows @xcite into the disk is the origin for the young stars ( the fueling rate implied by xuv disks is explored in * ? ? ? * ) and warps@xcite .
the fraction of spirals that have an disk ( @xmath7% ) supports this scenario as the remaining spirals may simply have no current cool gas inflow . in this case
, one would expect and morphology to follow each other reasonably closely for the disks but not for many of the others , as the star - formation in the inner disk is more closely related to the molecular phase @xcite .
however , the existence of disks pose an intriguing alternate possibility for the origin of the atomic hydrogen disk . instead of primordial gas accreting onto the disk , the disk could also be the byproduct of photodissociation of molecular hydrogen on the ` skins ' of molecular clouds by the ultraviolet flux of the young stars in the disk ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this explanation has been explored in the stellar disks of several nearby galaxies @xcite .
@xcite calculate the time - scales ( molecular gas dissociation and re - formation ) involved but these are inconclusive regarding the origin of the disk . in this scenario
, one would expect the and morphologies to follow each other closely in all cases ; in the outer disk , the flux from a few young stars would reach out to large areas of the low - column density gas to dissociate enough hydrogen to form the outer disk .
thus , the low - flux and low column density morphology those defining the limits and extent of the and disks should show a close relation in parameters such as concentration , gini , 20 and the effective radius .
a third explanation is in terms of recent tidal interaction .
a major merger often pulls gas out of the planes of galaxies and triggers star - forming events . some anecdotal evidence ( e.g. , ugc 04862 is a late - stage major merger * ? ? ?
* ) does point to this possible origin . in this scenario
, one would expect that most disks hosting an disk would be seriously tidally disrupted . in this series of papers ,
we have explored the quantified morphology of available maps with the common parameters for visible morphology ; concentration - asymmetry - smoothness , gini and 20 and @xmath0 . in @xcite
, we compare the morphology to other wavelengths , noting that the and ultraviolet morphologies are closely related . in subsequent papers of the series ,
we use the morphology to identify mergers @xcite , their visibility time @xcite and subsequently infer a merger rate from @xcite as well as identify phenomena unique to cluster members @xcite . in this paper
, we explore the morphological link between the and disks in the westerbork survey project ( , * ? ? ?
* ; * ? ? ?
* ) , a survey of several hundred observations of nearby galaxies .
we complement this data with galex images to explore the morphology and typical scales of these maps .
a direct and quantified comparison between the gas and ongoing star - formation morphology could help answer open questions regarding the origin and nature of disks : how do their respective sizes relate ?
are their morphologies closely related in every case ? do their respective morphologies point to a dominant formation mechanism ; gas accretion , photodissociation or tidal ? are disks in morphologically distinct or typical disks ?
are disks embedded in disks that appear to be in an active interaction ? what is the relation between flux and column density in the disks , especially the outer disk ?
the paper is organized as follows ; [ s : morph ] gives the definitions of the quantified morphology parameters we employ , and
[ s : data ] describes the origin the data .
we describe the application of the morphological parameters and the results in [ s : app ] and [ s : analysis ] , and we discuss them in [ s : disc ] .
we list our conclusions and discuss possible future work in [ s : concl ] .
in this series of papers , we use the concentration - asymmetry - smoothness parameters ( cas , * ? ? ? * ) , combined with the gini-@xmath8 parameters from @xcite and one addition of our own @xmath0 .
we have discussed the definitions of these parameters in the previous papers , as well as how we estimate uncertainties for each . here , we will give a brief overview but for details we refer the reader to @xcite .
we select pixels in an image as belonging to the galaxy based on the outer contour and adopt the position from the 2mass catalog @xcite as the central position of the galaxy . given a set of @xmath9 pixels in each object , iterating over pixel @xmath10 with value @xmath11 , position @xmath12 with the centre of the object at @xmath13 these parameters are defined as : @xmath14 with @xmath15 as the radial aperture , centered on @xmath13 containing percentage @xmath16 of the light of the galaxy ( see definitions of @xmath17 in * ? ? ?
* ; * ? ? ?
* ) .. , and we adopt the new , correct values in this paper . ]
we include the @xmath18 , or `` effective radius '' in our catalog as well .
@xmath19 where @xmath20 is the pixel at position @xmath10 in the galaxy s image , after it was rotated @xmath21 around the centre of the galaxy .
@xmath22 where @xmath23 is pixel @xmath10 in a smoothed image .
the type of smoothing ( e.g. , boxcar or gaussian ) has changed over the years .
we chose a fixed 5 " gaussian smoothing kernel for simplicity .
the gini coefficient is defined as : @xmath24 where the list of @xmath9 pixels was first ordered according to value and @xmath25 is the mean pixel value in the image .
@xmath26 where @xmath27 is the second order moment of pixel @xmath10 ; @xmath28 $ ] .
@xmath29 is the second order moment summed over all pixels in the object and @xmath8 is the relative contribution of the brightest 20% of the pixels in the object .
instead of using the intensity of pixel @xmath10 , the gini parameter can be defined using the second order moment : @xmath30 these parameters trace different structural characteristics of a galaxy s image but these do not span an orthogonal parameter space ( see also the discussion in * ? ? ?
we use three data sets for this paper ; radio observations from or , and ultraviolet observations from galex space telescope observations .
the starting dataset here is the 266 observations done as part of the westerbork observations of neutral hydrogen in irregular and spiral galaxies ( , * ? ? ?
* ; * ? ? ?
* ; 339 individual galaxies ) that also have reasonable quality and data .
the observation targets were selected from the uppsala general catalogue of galaxies @xcite , with blue major axis diameters @xmath31 , declination ( b1950 ) @xmath32 20 degrees and flux densities at 21-cm larger than 100 mjy , later lowered to 20 mjy .
observation times were typically 12 hours of integration .
the galaxies satisfying these selection criteria generally have redshifts less than 20000 km / s ( @xmath33 ) .
has mapped the distribution and velocity structure of in several hundreds of nearby galaxies , increasing the number of observations of galaxies by an order of magnitude .
the project provides a uniform database of data - cubes , zeroth - order and velocity maps .
its focus has been on the structure of the dark matter halo as a function of hubble type , the tully - fisher relation and the dark matter content of dwarf galaxies @xcite . until the large all - sky surveys with new instruments
are completed , is the largest , publicly available data - set of resolved observations .
we compiled a catalogue of basic data ( position , radial velocity etc . )
( see for details * ? ? ?
we use the highest available resolution zero - moment maps ( beam size of @xmath3412 `` x 12''/sin(@xmath35 ) ) , from the wow website and converted these to @xmath36 column density maps with j2000 coordinates and of the same size as the postage stamp ( see for details * ? ? ?
* ) . to complement the column density maps
, we retrieved galex @xcite postage stamps from http://skyview.gsfc.nasa.gov , near- and far - ultraviolet ( 13501750 and 17502750 ) .
the majority of observations were taken as part of the ais , mis , ngs , etc .
imaging surveys and hence vary in depth and signal - to - noise .
spatial resolution is 46 " , depending on position in the field - of - view . of the 339 galaxies in
, 266 had reliable information and this cross - section is what we use for the remainder of the paper .
the images with contours overlaid are shown in figures [ f : xuvt1 ] , [ f : xuvt2 ] , and figures [ f : xuv ] [ f : xhi ] and all galaxies in appendix c ( _ electronic edition _ ) . in the sample
, @xcite classified 22 galaxies for the presence and type of disks .
the type 1 disks are characterized by discrete regions out to high radii and type 2 by an anomalously large , uv - bright outer zone of low optical surface brightness .
table [ t : thilker ] lists the classifications for the sample ; 22 from @xcite and a single one from @xcite .
the latter study s lack of overlap is unsurprising as they focus on a more distant sample in complement to the thilker et al.s nearby sample ( @xmath37mpc ) .
we designate an disk with both a type 1 and type 2 as type 1/2 to discriminate in the plots below and mark those galaxies in whisp surveyed by thilker et al .
that do not have an disk as type 0 .
figures [ f : xuvt1 ] and [ f : xuvt2 ] show the overlap between and the disks identified by thilker .
the type 1 collection includes a single merger remnant ( ugc 4862 ) and some disks that appear slightly lopsided , tentatively supporting a tidal origin as one path to generate type 1 disks .
figure [ f : xuvt0 ] shows the contours around those galaxies inspected by thilker et al .
but not found to contain an disk .
there are two instances of clear ongoing mergers as well as a variety of morphologies .
[ [ s : things ] ] since there overlap between sample and the thilker et al classification sample is small , we also compare the survey s and morphologies to their classifications of disks ( table [ t : things ] ) .
we use the morphological information from @xcite , with the concentration definition appropriately amended .
the morphological classifications are based on the public datasets from the nearby galaxy survey ( things , * ? ? ?
* ) , and galex nearby galaxy atlas ( nga , * ? ? ?
* ) , retrieved from mast ( http://galex.stsci.edu ) .
we use the robust - weighted ( ro ) column density maps as these are similar in resolution as the native resolution of for the contour of @xmath38 @xmath39 , approximately the spatial extent of the disk .
table [ t : things ] lists the things galaxies that have disk classifications from @xcite .
the benefits of the sample are that there is an classification for all galaxies , the and data are of comparable spatial resolution , and the observations are of uniform depth .
drawbacks are that is is an equally small sample as the overlap and there is only a single ongoing merger in this sample as it was constructed to be quiescent . .
disks classified by @xcite and one from @xcite for the galaxies in their respective samples . [ cols="<,<,<",options="header " , ] [ t : crit ] in this section we look at the morphological parameters in detail to explore if bona - fide disks can be identified from their morphology alone as computed over an area defined by the disk . for this we can use the 22 galaxies in and the sample , both classified by thilker et al .
figure [ f : fuvmorph ] shows the parameters for the disks identified by @xcite as well as those disks inspected but without an extended component for both samples .
the above concentration - only criterion ( , @xmath40 ) pre - selects a number of the disks but it would miss several bona - fide disks as well as include many false positives . using the information in two ( or possibly more ) parameters is how we identify mergers and may identify ( candidate ) disks on inspection of the distributions of the and parameters ( figures [ f : scarlata_nuv ] and [ f : scarlata_fuv ] ) , there are several combinations of parameters that offer the possibility of pre - selecting disks without too much loss : gini - asymmetry , 20-asymmetry and gini- .
the goal is not to just select only the bona - fide , previously identified disks but also to exclude as many disks , that were previously classified as non- as practical .
this latter criterion immediately excludes gini - asymmetry and gini- combinations .
the combination of 20 and asymmetry in is promising and we define the following criterion to identify disks in an survey ( 6 `` resolution ) : latexmath:[\[\label{eq : things : m20a } sample and we divide the linear relation by a factor two for the survey ( to account for the @xmath3412 '' resolution ) : @xmath42 with dashed lines .
table [ t : crit ] lists the two criteria the simple concentration criterion and the 20-asymmetry one ( adjusted for each survey ) and their success rate ( percentage of all bona - fide disks included ) and contamination rate ( percentage of all classified objects that are non- ) .
the 20-asymmetry criterion selects most but with a contamination of some 55% of non- disks . a future search for disks could therefore use this as a first cut before visual classification .
given the 109 galaxies selected by this criterion ( table [ t : xuvauto ] ) , 61 of these should be bona - fide disks , 23% of our sample of 266 .
figure [ f : himorph ] shows five of the morphological criteria we used or defined in @xcite to identify those disks that are currently undergoing an interaction ( see figure [ f : scarlata_hi2 ] for a similar plot to figure 5 in @xcite for a direct comparison ) .
we mark those galaxies identified by @xcite and those identified as extended uv or disks by the concentration parameters .
based on the morphologies , the @xmath0 and @xmath43 criteria , those disks that host an disk are not interacting in the majority of cases . in the case of the other morphological criteria ( @xmath44 , @xmath45 , @xmath46 ) ,
this distinction is not as clear - cut but disks either selected by thilker et al or by their concentration appear to reside in morphologically very typical disks . at the same time , extended disks are often interacting .
this is in part because of our definition of an interaction is based on concentration but even in the @xmath0 criterion , these are conspicuously separated from the bulk of the disks .
typically , the bona - fide disks selected by an morphological selection of an ongoing merger appear to be type 1 , consistent with the first impression from figures [ f : xuvt0 ] - [ f : xuvt2 ] .
thus , mergers appear to be one avenue to generate a type 1 disk . if we plot these as a sanity check for the sample , the disks identified by @xcite equally appear not to be interacting based on the criterion , but several are based on g - a or 20-g criteria .
several of the non- disks are interactions in a-20 .
these criteria would need to be adjusted for the higher resolution of the sample ( difficult to do without many mergers in the sample , as we noted in * ? ? ?
our first motivation for this study was to explore the possibility if disks could be reliably found using the morphological parameters in and , especially when computed over a wide aperture such as the disk .
based on their morphology parameters in and , disks identified in @xcite rarely stand out from the bulk of the sample , with the exception of 20-a and concentration in . in the
, the 20-a relation appears reasonably successful in identifying disks and the rate in the sample ( 23% ) is consistent with the @xcite and @xcite result , for a survey of late - types such as .
the values of asymmetry and 20 could therefore be used to identify candidate disks in the local universe with or similar quality data or in deep rest - frame observations with _
however , these will have to be computed over a much larger area than commonly used ( e.g. , the de vaucouleur diameter @xmath47 or the petrosian radius ) , such as the one defined by the outer contour .
the quantified and morphologies do not correlate as closely as they did for the sample in @xcite ( their figure 6 ) . in our opinion , this is due to the wider range of environments probed by the survey as well as its lower resolution compared to .
a closer relation may return for the inner stellar disk and possibly at higher spatial resolution .
our second motivation for this work on disks was to explore their origin .
are they the ( by)product of ( a ) cold flow accretion , ( b ) a gravitational interaction or mergers with another galaxy or ( c ) is photo - dissociation of molecular clouds by the regions of the disks responsible for the disks ?
the fact that the majority of the disks do not reside in morphologically distinct disks ( figure [ f : morphfuv ] and [ f : thingsfuv ] ) , precludes the notion that most disks originate from major interaction between galaxies .
few are selected by the previously defined interaction criteria for disks .
there are , however , some obvious exceptions ( figure [ f : xuvt1 ] ) .
so our nave interpretation for the origin of type 1 disks in figure [ f : xuvt1 ] is only partly supported by morphology and only holds for type 1 . since extended ( ) and extended disk ( ) populations as defined by their concentration parameter computed over the
area do not coincide ( figure [ f : morphfuv ] , top left panel ) , a direct link , such as the photo - dissociation origin of the disk , appears less likely .
a closer relation in morphology may well still hold within the inner stellar disk , especially since the photo - dissociation scenario was developed to explain the higher column densities of in this environment @xcite
. however , the existence of both the @xmath3 disks without complementary extended disks as well as extended ( @xmath48 ) disks without a complementary disk appears to be contradictory to the photo - dissociation origin of disks .
two considerations are still in its favor however , @xcite showed that uv emission does not fully trace the lower - lever star - formation in the low - column density environments of the outer disk .
thus , the ionizing stars may be lower - mass main sequence stars and simply remain undetected in imaging due to their extreme low surface - brightness . and with higher resolution and more sensitive and imaging , a closer relation may still return in their quantified morphology ( witness the sample ) .
thus , we can not completely rule out a photo - dissociation scenario explaining the relation between and disks .
similarly , @xcite find their time - scale argument for a photo - dissociation origin of disks to be inconclusive .
this leaves cold flows as a likely origin for most disks .
a number of these type 1 disks may be the result of a later - stage full merger such as ugc 04862 but judging from the morphologies , this is small fraction of the full or sample .
cold flows occur as both a steady trickle along a filament of the cosmic web as well as discrete accretion in the form of the cold gas originally associated with a small satellite .
there are indirect signs that cold flows are indeed the most common cause for disks .
for example , at least three of the nine galaxies with an disk ( m83 , ngc 4736 and ngc 5055 ) are currently cannibalizing a small companion @xcite .
satellites embedded in the gas filament are one of two fueling mechanisms of cold flow ; discrete and trickle cold flows @xcite .
the acquired cold gas accompanying the cannibalized companion is a likely trigger for the formation complexes .
similarly , those disks selected by their extended emission ( figure [ f : xuv ] ) , often appear to have flocculent , low - column density morphologies , which _ may _ point to recent cold accretion .
a few cases ( ugc 7081 , 7651 , and 7853 ) are obviously interacting major mergers with accompanying / tidal features .
in contrast , the disks are nearly always small galaxies with an obvious close companion ( figure [ f : xhi ] ) , which makes a tidal origin likely for their very extended disk .
the quantified morphologies of the and disks are not conclusive evidence that cold flows are responsible for the complexes but certainly suggest this is the most likely origin for most , with a few exception due to major interactions .
based on the quantified morphology of the and maps of the sample of galaxies , we conclude the following : * there are distinct galaxy populations that stand out by their high concentration values in or ( and thoughout the paper ) .
these population do not overlap .
some known disks are in the sample ( figures [ f : morphfuv]-a [ f : morphfuv]-e , [ f : thingsfuv]-a and [ f : thingsfuv]-c).to compute this concentration index , the outer contour is needed to delineate the extent of the disk . * the fact that relative dilute disks ( @xmath3 ) and disks ( @xmath2 ) population do not overlap at all and the lack of close morphological relations in any other parameters together suggest that a common small - scale origin for the and disks , such as a photodissociation of molecular hydrogen scenario , is less likely for disks ( but may still very much hold for the inner disks ) .
* asymmetry and 20 can be used in combination to select disks with reasonable reliability ( 80% included ) but substantial contamination ( 55% ) , when properly calibrated for the survey s spatial resolution ( equation [ eq : things : m20a ] and [ eq : whisp : m20a ] ) and the morphologies computed over a large enough aperture . with this selection , one can find the number of disks in a survey or candidates for visual classification ( figure [ f : fuvmorph ] ) .
* based on the morphology of the disk in which they occur , disks appear not to occur often in tidally disturbed gas disks ( figures [ f : himorph ] and [ f : hithings ] ) . * in a few cases , the disk _ is _ the product of a major merger ; for example ugc 04862 , identified by @xcite and ugc 7081 , 7651 , and 7853 , identified by their large concentration .
this appears to be an avenue to form _ some _ of the type 1 disks . * the morphology and anecdotal evidence for small satellite cannibalism all point to a third mechanism for the origin of most disks ; cold flow accretion . with the emergence of new and refurbished radio observatories in preparation for the future square kilometre array ( ska ; *
* ) , a new window on the 21 cm emission line of atomic hydrogen gas ( ) is opening .
the two ska precursors , the south african karoo array telescope ( meerkat ; * ? ? ?
* ; * ? ? ?
* ) , and the australian ska pathfinder ( askap ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) stand poised to observe a large number of southern hemisphere galaxies in in the nearby universe ( [email protected] ) . in addition , the extended very large array ( evla ; * ? ? ?
* ) and the aperture tile in focus instrument ( apertif ; * ? ?
* ; * ? ? ?
* ) on the westerbork synthesis radio telescope ( wsrt ) will do the same for the northern hemisphere .
the surveys conducted with these new and refurbished facilities will add new , high - resolution , observations on many thousands of galaxies . in the case of wallaby ( koribalski et al .
_ in preparation _ )
, these will be of similar quality spatial resolution as the survey . combining these with existing observations from
, we can explore the relation between the morphology of the atomic hydrogen and ultraviolet light for much greater samples and in much greater detail .
an application of the asymmetry-20 identification of disks in the nearby galaxy atlas @xcite or a sample similar to @xcite could reveal additional examples but the combination with deep observations can prove the link with cold flows for the majority of disks .
the authors would like to thank d. thilker for discussions at the cloister see - on conference , the audience at leiden observatory for encouragement for this work and the anonymous referee for his or her comments that improved the science and writing of this paper .
the authors acknowledge the us of the hyperleda database http://leda.univ-lyon1.fr , 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 , and the westerborg on the web ( wow ) project ( http://www.astron.nl/wow/ ) .
the westerbork synthesis radio telescope is operated by the netherlands institute for radio astronomy astron , with support of nwo . based on observations made with the nasa galaxy evolution explorer .
galex is operated for nasa by the california institute of technology under nasa contract nas5 - 98034 .
r. j. , 2002 , in astronomical society of the pacific conference series , vol .
276 , seeing through the dust : the detection of hi and the exploration of the ism in galaxies , taylor a. r. , landecker t. l. , willis a. g. , eds .
, p. 288
a. , boissier s. , madore b. f. , seibert m. , joe y. h. , boselli a. , wyder t. k. , thilker d. , bianchi l. , rey s .- c .
, rich r. m. , barlow t. a. , conrow t. , forster k. , friedman p. g. , martin d. c. , morrissey p. , neff s. g. , schiminovich d. , small t. , donas j. , heckman t. m. , lee y .- w . , milliard b. , szalay a. s. , yi s. , 2007 , , 173 , 185 a. , madore b. f. , boissier s. , thilker d. , bianchi l. , snchez contreras c. , barlow t. a. , conrow t. , forster k. , friedman p. g. , martin d. c. , morrissey p. , neff s. g. , rich r. m. , schiminovich d. , seibert m. , small t. , donas j. , heckman t. m. , lee y .- w . , milliard b. , szalay a. s. , wyder t. k. , yi s. , 2007 , , 661 , 115 s. , bailes m. , bartel n. , baugh c. , bietenholz m. , blake c. , braun r. , brown j. , chatterjee s. , darling j. , deller a. , dodson r. , edwards p. g. , ekers r. , ellingsen s. , feain i. , gaensler b. m. , haverkorn m. , hobbs g. , hopkins a. , jackson c. , james c. , joncas g. , kaspi v. , kilborn v. , koribalski b. , kothes r. , landecker t. l. , lenc e. , lovell j. , macquart j .- p .
, manchester r. , matthews d. , mcclure - griffiths n. m. , norris r. , pen u .-
phillips c. , power c. , protheroe r. , sadler e. , schmidt b. , stairs i. , staveley - smith l. , stil j. , taylor r. , tingay s. , tzioumis a. , walker m. , wall j. , wolleben m. , 2007 , publications of the astronomical society of australia , 24 , 174 s. , feain i. j. , gupta n. , 2009 , in astronomical society of the pacific conference series , vol .
407 , astronomical society of the pacific conference series , d. j. saikia , d. a. green , y. gupta , & t. venturi , ed . , pp .
446+ s. , taylor r. , bailes m. , bartel n. , baugh c. , bietenholz m. , blake c. , braun r. , brown j. , chatterjee s. , darling j. , deller a. , dodson r. , edwards p. , ekers r. , ellingsen s. , feain i. , gaensler b. , haverkorn m. , hobbs g. , hopkins a. , jackson c. , james c. , joncas g. , kaspi v. , kilborn v. , koribalski b. , kothes r. , landecker t. , lenc a. , lovell j. , macquart j .-
, manchester r. , matthews d. , mcclure - griffiths n. , norris r. , pen u .-
l . , phillips c. , power c. , protheroe r. , sadler e. , schmidt b. , stairs i. , staveley - smith l. , stil j. , tingay s. , tzioumis a. , walker m. , wall j. , wolleben m. , 2008 , experimental astronomy , 22 , 151 d. c. , fanson j. , schiminovich d. , morrissey p. , friedman p. g. , barlow t. a. , conrow t. , grange r. , jelinsky p. n. , milliard b. , siegmund o. h. w. , bianchi l. , byun y .- i .
, donas j. , forster k. , heckman t. m. , lee y .- w .
, madore b. f. , malina r. f. , neff s. g. , rich r. m. , small t. , surber f. , szalay a. s. , welsh b. , wyder t. k. , 2005 , , 619 , l1
d. , gabany r. j. , crawford k. , zibetti s. , majewski s. r. , rix h .- w .
, fliri j. , carballo - bello j. a. , bardalez - gagliuffi d. c. , pearrubia j. , chonis t. s. , madore b. , trujillo i. , schirmer m. , mcdavid d. a. , 2010 , , 140 , 962 g. r. , wong o. i. , kim j. h. , hanish d. j. , heckman t. m. , werk j. , bland - hawthorn j. , dopita m. a. , zwaan m. a. , koribalski b. , seibert m. , thilker d. a. , ferguson h. c. , webster r. l. , putman m. e. , knezek p. m. , doyle m. t. , drinkwater m. j. , hoopes c. g. , kilborn v. a. , meyer m. , ryan - weber e. v. , smith r. c. , staveley - smith l. , 2009 , , 695 , 765 p. , 1973
, uppsala general catalogue of galaxies .
acta universitatis upsaliensis .
nova acta regiae societatis scientiarum upsaliensis - uppsala astronomiska observatoriums annaler , uppsala : astronomiska observatorium , 1973 c. , carollo c. m. , lilly s. , sargent m. t. , feldmann r. , kampczyk p. , porciani c. , koekemoer a. , scoville n. , kneib j .-
, leauthaud a. , massey r. , rhodes j. , tasca l. , capak p. , maier c. , mccracken h. j. , mobasher b. , renzini a. , taniguchi y. , thompson d. , sheth k. , ajiki m. , aussel h. , murayama t. , sanders d. b. , sasaki s. , shioya y. , takahashi m. , 2007 , , 172 , 406 d. a. , bianchi l. , boissier s. , gil de paz a. , madore b. f. , martin d. c. , meurer g. r. , neff s. g. , rich r. m. , schiminovich d. , seibert m. , wyder t. k. , barlow t. a. , byun y .-
i . , donas j. , forster k. , friedman p. g. , heckman t. m. , jelinsky p. n. , lee y .- w . , malina r. f. , milliard b. , morrissey p. , siegmund o. h. w. , small t. , szalay a. s. , welsh b. y. , 2005 , , 619 , l79 d. a. , bianchi l. , meurer g. , gil de paz a. , boissier s. , madore b. f. , boselli a. , ferguson a. m. n. , muoz - mateos j. c. , madsen g. j. , hameed s. , overzier r. a. , forster k. , friedman p. g. , martin d. c. , morrissey p. , neff s. g. , schiminovich d. , seibert m. , small t. , wyder t. k. , donas j. , heckman t. m. , lee y .- w . , milliard b. , rich r. m. , szalay a. s. , welsh b. y. , yi s. k. , 2007 , , 173 , 538 d. a. , hoopes c. g. , bianchi l. , boissier s. , rich r. m. , seibert m. , friedman p. g. , rey s .- c . , buat v. , barlow t. a. , byun y .-
i . , donas j. , forster k. , heckman t. m. , jelinsky p. n. , lee y .- w .
, madore b. f. , malina r. f. , martin d. c. , milliard b. , morrissey p. f. , neff s. g. , schiminovich d. , siegmund o. h. w. , small t. , szalay a. s. , welsh b. y. , wyder t. k. , 2005 , , 619 , l67 j. m. , 2002 , in astronomical society of the pacific conference series , vol .
276 , seeing through the dust : the detection of hi and the exploration of the ism in galaxies , taylor a. r. , landecker t. l. , willis a. g. , eds . , pp .
84+ m. a. w. , oosterloo t. a. , van cappellen w. a. , bakker l. , ivashina m. v. , van der hulst j. m. , 2008 , in american institute of physics conference series , vol . 1035 , the evolution of galaxies through the neutral hydrogen window , r. minchin & e. momjian , ed . , pp . | extended ultraviolet ( ) disks have been found in a substantial fraction of late - type s0 , spiral and irregular galaxies .
similarly , most late - type spirals have an extended gas disk , observable in the 21 cm radio line ( ) . the morphology of galaxies can be quantified well using a series of scale - invariant parameters ; concentration - asymmetry - smoothness ( cas ) , gini , 20 , and @xmath0 parameters . in this series of papers
, we apply these to column density maps to identify mergers and interactions , lopsidedness and now disks . in this paper , we compare the quantified morphology and effective radius ( @xmath1 ) of the westerbork observations of neutral hydrogen in irregular and spiral galaxies project ( ) maps to those of far - and near - ultraviolet images obtained with , to explore how close the morphology and scales of and in these disks correlate .
we find that disks do not stand out by their effective radii in or .
however , the concentration index in appears to select some disks . and
known disks can be identified via a criterion using asymmetry and 20 ; 80% of disks are included but with 55% contamination .
this translates into 61 candidate disk out of our 266 galaxies , 23% consistent with previous findings .
otherwise , the and morphology parameters do not appear closely related .
our motivation is to identify disks and their origin .
we consider three scenarios ; tidal features from major mergers , the typical extended disk is a photo - dissociation product of the regions and both and features originate in cold flows fueling the main galaxy .
we define extended and disks based on their concentration ( @xmath2 and @xmath3 respectively ) , but that these two subsamples never overlap in the sample .
this appears to discount a simple photo - dissociation origin of the outer disk .
previously , we identified the morphology space occupied by ongoing major mergers .
known disks rarely reside in the merger dominated part of morphology space but those that do are type 1 .
exceptions , disks in ongoing mergers , are the previously identified ugc 4862 and ugc 7081 , 7651 , and 7853 .
this suggests cold flows as the origin for the complexes and their surrounding structures .
[ firstpage ] |
the transfer of phonons across the interface between two solids is crucial to the operation of a range of novel devices from semiconductor nanostructures , quantum - cascade lasers , vertical structures , superlattices , quantum dot arrays to the cryogenic phonon - mediated detectors of elementary particles . in such structures non - radiative transitions and
hot - electron thermalization gives rise to fluxes of highly energetic phonons spanning the whole of the brillouin zone .
in early studies @xcite , phonons were regarded as bulk elastic continuum waves so that transmission and reflection of energy at the interface between two materials were a consequence of acoustic mismatch .
such a macroscopic continuum theory describes generic low - frequency properties but does not account for important features such as phonon dispersion , the debye cutoff and weak bonding at the interface , for which a lattice - dynamical model is required @xcite .
most early investigations assume that the solid - solid interface possesses in - plane translational invariance , whereas , it is known that disorder - induced scattering can play a significant role @xcite . within a lowest order perturbation theory @xcite
, it is predicted that disorder can induce a strong frequency dependence in phonon transport coefficients with a power - law crossover determined by interface conditions .
the aim of this paper is to go beyond this elastic continuum perturbative approach , by formulating an exact lattice - description valid for terahertz phonon transport , which includes features such as phonon dispersion and singularities in the phonon density of states . to formulate the problem of phonon transmission ( reflection ) we consider two identical semi - infinite leads of uniform cross - section attached to a scattering region. the two semi - infinite perfect crystals are treated as waveguides for incident and scattered phonons whose dynamics is described by the lattice equations of motion .
the geometry analyzed is shown in fig.1 , where the scattering region consists of a single disordered atomic plane .
disorder at the planar interface is introduced microscopically as correlated or uncorrelated random variation of the on - site atomic masses and we assume that there are no significant anharmonic effects . to calculate transport properties of such a system , we use methods specifically developed to investigate electronic transport through phase - coherent structures @xcite .
transport coefficients are calculated as a function of phonon frequency within the landauer - buttiker formalism @xcite by employing an exact recursive green function technique @xcite .
in what follows we present the results for two lattice dynamical models : an fcc lattice with central - force nearest - neighbour interactions and a scalar square lattice . for the former the atomic masses
@xmath0 are located at equilibrium positions @xmath1 , where @xmath2 sum to an even number and @xmath3 is the nearest - neighbour spacing , whereas for the square lattice @xmath4 is the lattice constant and @xmath5 . in the latter case
we do not distinguish between phonon polarizations and hence , are not concerned about mode conversion due to scattering . in the harmonic approximation the linearized equations of motion of an atom located at @xmath6
is @xmath7 where @xmath8 is the atomic displacement vector , @xmath9 and @xmath10 is taken from the first coordination sphere .
the indices @xmath11 label the @xmath12 components of displacements for the fcc model and should be dropped for the scalar model .
all atoms in the semi - infinite leads are of unit mass , whereas the random masses at the interface have mean value unity and standard deviation @xmath13 . for simplicity we allow the masses to fluctuate only along a single direction , which is chosen to be the [ 101 ] and [ 100 ] cubic crystallographic directions for the fcc and square lattice respectively .
henceforth , we refer to this as the disorder - line and denote masses at different positions along the line by @xmath14 .
thus , even for a three dimensional lattice , due to the residual translational invariance , scattering is confined in the plane fixed by the disorder - line and the angle with respect to the interface determined by the conserved component of the incident phonon momentum . from this point of view
our scattering problem is essentialy two dimensional .
an ensemble of @xmath15 positive masses , where @xmath15 is the number of atoms along the direction of the disorder - line in the lead cross - section , is generated as follows : first we introduce the correlated random numbers @xmath16 where @xmath17 and @xmath18 are gaussian random numbers with zero mean value and @xmath19 .
@xmath20 is taken commensurate with the phonon wavenumbers determined by the boundary conditions in the transverse direction and @xmath21 denotes the correlation length .
random masses @xmath14 along the disorder - line with mean value unity and standard deviation @xmath13 are obtained from @xmath22 by the linear transformation @xmath23 in order to isolate the effect of correlations , we compare phonon transmission ( reflection ) coefficients for such correlated interfacial disorder , with those of an uncorrelated disorder configuration obtained by randomly shuffling the above set .
in the absence of inelastic scattering , phonon transport through an arbitrary scattering region can be described by the scattering matrix @xmath24 @xcite , which depends on the conserved phonon frequency and yields probabilities for all scattering transitions within a system described by a dynamical matrix @xmath25 . if the scattering region is connected to external reservoirs by phonon waveguides with open channels labelled by quantum numbers @xmath26 , then @xmath24-matrix elements @xmath27 are defined such that @xmath28 is the outgoing flux of phonons along channel @xmath26 , arising from a unit flux incident along channel @xmath29 .
the scattering matrix defined this way is unitary , i.e. , @xmath30 , and since @xmath31 , where @xmath32 is real , @xmath33 is symmetric .
if @xmath34 belong to the left ( right ) lead then the corresponding @xmath24-matrix elements are left ( right ) reflection amplitudes which we denote @xmath35 , whereas , if @xmath34 belong to the right ( left ) and left ( right ) lead , the @xmath24-matrix elements are left ( right ) transmission amplitudes denoted @xmath36 .
it is evident that for an @xmath37-mode waveguide three distinct transmittances can be defined , @xmath38 to isolate the frequency dependence of the overall phonon transmittance it is natural to normalize @xmath39 by the density of states . in the presence of
finite - width leads this means averaging @xmath39 over all open scattering channels @xmath37 and is equivalent to the angle - averaged transmission coefficient , @xmath40 which is related to the corresponding reflection coefficient by @xmath41 in what follows the @xmath24-matrix is obtained by solving dyson s equation for the green function at the surfaces of the two leads @xcite ( see fig.1 for definition of l , r ) , namely @xmath42 @xmath43 where @xmath44 is the surface green function of the semi - infinite leads @xmath45 and @xmath46 an effective dynamical matrix @xmath47 the latter describes the effective ( renormalized ) coupling between the surfaces of the two leads , obtained by projecting out the internal degrees of freedom of the scatterer using the recursive green function technique .
the green function @xmath44 , the spectrum and structure of the allowed wavequide modes at fixed phonon density of states , as well as @xmath33 are calculated using a general algorithm developed in ref . @xcite .
we present results for the average transmission coefficient @xmath48 where the subscript c denotes a configurational averaging . in fig.2(a),(b )
the frequency dependence of the overall transmission coefficient @xmath48 averaged over 10 realizations of the interfacial mass - disorder for the fcc lattice and over 500 for the scalar square lattice is shown .
the width of the disorder - line is 29 and 100 nearest - neighbour spacings respectively , which corresponds to @xmath49 and @xmath50 degrees of freedom . for the fcc model periodic boundary conditions in the lateral direction ( disorder - line )
were used , whereas for the scalar model we adopted fixed - end boundary conditions .
the boundary conditions set the precise frequencies at which propagating waveguide modes become available in the external leads , but apart from features which scale inversely with the number of such modes do not affect the shape of the transmittance curves .
figure 2 shows that for both fcc and scalar lattices , phonon transmittances ( reflectances ) follow the same universal behaviour . at sufficiently low frequencies ,
the phonon transmittance for correlated disorder is lower than for uncorrelated interfacial disorder with the same mean mass and standard deviation @xmath13 . at low frequencies the results of lattice - dynamical calculations in fig.2(a),(b )
are in good agreement with perturbation theory @xcite . to model the disorder in @xcite a thin layer of dissimilar material sandwiched between irregular boundaries was introduced .
the density and elastic constants of this layer were assumed to differ from those of the bulk on either side of the layer . to make a comparison with the perturbative analysis we consider the case of no fluctuation in the elastic constants and a line of random masses at the location of the idealized interfacial plane @xmath51 , i.e. , @xmath52 , with @xmath53 a random function of lateral coordinate with zero mean value .
apart from lattice discreteness which plays no significant role in the long - wavelength limit there is no difference between our model of site - to - site mass disorder and the model of interfacial roughness of ref .
the following relation is useful when comparing the results @xmath54 for such a two dimensional model the perturbation theory @xcite with uncorrelated disorder @xmath55 and @xmath56 , where k is the phonon wavevector , yields @xmath57 while for correlated disorder at sufficiently high frequencies so that @xmath58 the perturbative result is @xmath59 more generally for all @xmath60 , fig.2(c ) shows the perturbative results for @xmath61 and @xmath62 . equations ( [ per1 ] ) and ( [ per2 ] ) are in good agreement with the exact lattice - dynamical calculation .
they both show lower transmittance for correlated disorder over a wide range of frequencies until the faster power dependence of phonon transmittance for uncorrelated disorder causes the two curves to cross at a frequency @xmath63 . according to the perturbative calculations @xmath63 should be close to half the maximum phonon frequency @xmath64 independent of disorder characteristics , correlation length and strength . despite the similarities ,
this is not evident in the lattice - dynamical calculations .
indeed , the crossing point is quite close to a singularity in the phonon density of states and thus using an elastic continuum model at frequencies of order @xmath63 is an oversimplification .
another simplifying assumption of the theory in ref.@xcite is the neglect of multiple scattering .
this gives the power law @xmath65 ( rayleigh scattering ) with subsequent decrease by one ( two ) power(s ) for correlated disorder in 2d ( 3d ) .
a log - log plot of phonon reflectances showed that both lattice dispersion and multiple phonon scattering effects result in a change of the power exponents to a slightly different value .
however , the drop by approximately one power for our 2d scattering problem in the case of correlated disorder prevails in all cases .
the effect is attributed to the restrictions in the phase volume available for the scattered states , due to the correlation - induced finite width of the disordered spectral distribution .
phonon transmittances over the whole frequency spectrum for the scalar model for two different values of disorder strength @xmath13 are shown in fig.3 .
in addition to the low frequency behaviour discussed in sec .
iv , the transmittance shows a cusp at a frequency corresponding to a van hove singularity in the phonon density of states and rapidly drops to zero at the maximum frequency of the phonon spectrum @xmath64 . at the van hove singularity
the difference between correlated and uncorrelated disorder realizations is much less pronounced .
all these features are also seen for the fcc model but the exact behaviour at van hove singularities is less clear due to the finite - size effects described in the previous section . to interpret the results we concentrate on the scalar model .
the general expression for the transmission amplitude from incident channel @xmath29 to channel @xmath26 for systems with nearest - neighbour interactions has been derived in @xcite .
the explicit expression for the scalar model with periodic boundary conditions is @xmath66 here @xmath67 is the component of the group velocity in _ x_-direction along the lead axis of the @xmath26-th channel mode , m labels sites at the interface ( along lateral direction ) , and @xmath68 has been defined in eq.([green ] ) .
the matrix w is given by @xcite @xmath69 in formulae ( [ amplit],[w ] ) , @xmath70 , as a result of transverse quantization with periodic boundary conditions , while @xmath71 denotes the wavenumber along the lead axis for phonons of fixed frequency @xmath60 of state @xmath26 ( @xmath26-th channel ) . using the complete orthonormalized set of functions @xmath72 describing the profile of modes in the transverse direction , @xmath73 can be written @xmath74 by substitution of eq.([w]),([green1 ] ) to ( [ amplit ] ) one obtains @xmath75 thus , to calculate the channel - to - channel transmittance @xmath76 or generally the overall transmittance @xmath77 one needs to consider the non - trivial statistically averaged product of two green functions , @xmath78 or @xmath79 . in
what follows , we shall demonstrate that generic features can be captured by considering the case of a regular array of atoms with mass @xmath80 located at the interface , thereby forming a linear(planar ) defect in a two(three ) dimensional system . in this case , the full green function @xmath81 is of the form @xmath82 where @xmath83 is the green function of the infinite ideal lead at coinciding arguments , and @xmath84 is the relative mass difference at the location of the defects . substituting this result into eq.([trans2 ] ) and ( [ amplit1 ] ) yields for the transmittance @xmath85 eq.([total ] ) exhibits the following features : ( i)@xmath86 equals unity for an ideal interface with @xmath87 .
( ii)at low frequencies both @xmath88 , @xmath89 and therefore @xmath90 , where @xmath91 is numerical coefficient , in full agreement with the results of perturbation theory @xcite and formula ( [ per2 ] ) .
( iii)at the end of the phonon spectrum @xmath92 since propagating channels are closing with @xmath93 approaching zero at the brillouin zone boundaries .
this effect is analogous to the total reflection by a single mass defect in one dimension .
( iv)at the van hove singularity all channels are open and eq.([total ] ) reduces to @xmath94 which suggests that for @xmath95 the dependence on disorder strength is much weaker than at low frequencies . also , @xmath96 is small for a large number of contributing channels thereby yielding the characteristic cusp .
( v)we also note that eq.([total ] ) has been derived for the [ 100 ] orientation of the interface but since the phonon group velocity along the lead axis is proportional to @xmath93 the transmittance is in general anisotropic . in fig.4(a ) , we plot the phonon transmittance given by formula ( [ total ] ) for selected values of @xmath97 along with the numerical results for correlated disorder realizations .
notice that in spite of the differences in the type of interfacial defect the curves look very similar .
a natural way to explain this result is to view the interface for correlated disorder as a mosaic of clusters of average diameter @xmath98 .
suppose that within a particular cluster @xmath99 all sites are occupied by defects with the same mass and characteristic relative mass difference @xmath100 .
as long as the phonon wavelength is smaller than the correlation length , i.e. , @xmath58 , the overall transmittance can be considered as the average over all clusters with local transmittances given by eq.([total ] ) . in fig.4(b ) such an average is plotted for clusters with excess mass fully balanced by clusters with mass deficit .
the relative mass difference is taken from a gaussian distribution with width equal to the disorder strength @xmath13 .
clearly , this procedure gives a better fit to the correlated disorder curve , which suggests that the phonon transmittance across a correlated disordered interface can be split to a statistically averaged mean field contribution induced by mosaic grains and a fluctuating component .
the latter is controlled by the rapid residual variations in the relative mass difference at the cluster boundaries as well as the fluctuations in cluster sizes .
the qualitative agreement between correlated disorder realizations and the mosaic picture prompts us to ask whether there exists a simple model for the case of uncorrelated disorder . to explore this possibility
, we examine a two - colour disorder model where linear interfacial defects are built up from atoms with relative mass difference @xmath97 which alternates sign every @xmath101 sites .
this model is motivated by the observation that uncorrelated disorder contains all spatial harmonics with equal probability .
in what follows , we compare the exact transmittance for uncorelated disorder with that of the two - colour model , averaged over many values of @xmath102 . in fig.5
, we plot the mean arithmetic value of the phonon transmittances for @xmath103 taken after averaging over 500 two - colour disorder realizations given by a gaussian distribution with zero mean and @xmath104 .
the remarkable agreement with the calculations for the uncorrelated disorder over a wide range of frequencies suggests that the mean field contribution for this case can be calculated by picking up few elements from the basis set of the two - colour disorder model . finally , we compare the results of our calculation with the diffuse mismatch model _ dmm _ introduced in @xcite and extensively used in studies of kapitza conductance @xcite .
application of the dmm at a boundary between two solids with identical acoustic properties gives a @xmath105 phonon transmittance independent of frequency .
it can be easily demonstrated that the kapitza conductance @xmath106 for our model is larger than that predicted by the dmm .
it is also less than the radiation limit which in this case would coincide with acoustic mismatch theory ( total transmission ) .
moreover , @xmath106 is larger for uncorrelated disorder than for correlated .
finally , one notes that the results of @xcite which are valid for phonon transmission through the ideal interface between two different fcc crystals a transmittance which is frequency - independent over the whole spectrum up to a cut - off frequency corresponding to maximum phonon frequency in one of the crystals , whereas , in our model the phonon transmittance is a strong function of frequency .
in summary , generic properties of elastic phonon transport at a disordered interface were studied by computing the exact scattering properties of a single disordered atomic layer sandwiched between identical waveguides .
the results show that phonon transmittance is a strong function of frequency and the disorder correlation length . at frequencies lower than the van hove singularity the transmittance at a given frequency increases as the correlation length decreases . at low frequencies ,
this is reflected by different power - laws for phonon conductance across correlated and uncorrelated disordered interfaces which are in approximate agreement with perturbation theory of an elastic continuum @xcite .
we also investigated simple mosaic and two - colour models of the interface and showed that phonon transmittance through interfaces with correlated and uncorrelated disorder can be understood in terms of phonon conductances of structures with regular interfaces . | generic properties of elastic phonon transport at a disordered interface are studied .
the results show that phonon transmittance is a strong function of frequency and the disorder correlation length . at frequencies lower than the van hove singularity the transmittance at a given frequency increases as the correlation length decreases . at low frequencies ,
this is reflected by different power - laws for phonon conductance across correlated and uncorrelated disordered interfaces which are in approximate agreement with perturbation theory of an elastic continuum .
these results can be understood in terms of simple mosaic and two - colour models of the interface . |
negative index materials ( nims ) were investigated theoretically by veselago in @xcite and the existence of such materials was confirmed by shelby , smith , and schultz in @xcite .
the study of nims has attracted a lot attention in the scientific community thanks to their many potential applications .
mathematically , the study of nims faces two difficulties .
first , the equations modeling nims have sign changing coefficients ; hence the ellipticity and the compactness are lost in general . secondly , the localized resonance , i.e. the fields blow up in some regions and remain bounded in some others as the loss goes to 0 , might appear .
three known applications of nims are superlensing and cloaking using complementary media and cloaking a source via anomalous localized resonance ( alr ) . superlensing using complementary media
was suggested by veselago in @xcite for a slab lens ( a slab of index @xmath1 ) using the ray theory .
later , cylindrical lenses in the two dimensional quasistatic regime , the veselago slab lens and cylindrical lenses in the finite frequency regime , and spherical lenses in the finite frequency regime were studied by nicorovici , mcphedran , and milton in @xcite , pendry in @xcite , and ramakrishna and pendry in @xcite respectively for dipoles sources . superlensing using complementary media for arbitrary objects in the acoustic and electromagnetic settings
was established by nguyen in @xcite for related schemes . cloaking using complementary media
was suggested and investigated numerically by lai et al . in @xcite .
this was established rigorously for related schemes by nguyen in @xcite for the quasi - static regime and later extended by nguyen and ( h. l. ) nguyen in @xcite for the finite frequency regime . cloaking a source via alr
was discovered by milton and nicorovici in @xcite for constant radial symmetric plasmonic structures in the two dimensional quasi - static regime .
their work has root from @xcite ( see also @xcite ) where the localized resonance was observed and established for such a setting .
later , cloaking a source via alr was studied by milton et al . in @xcite ,
bouchitte and schweizer in @xcite , ammari et al . in @xcite ,
kohn et al in @xcite , , nguyen and ( h. l. ) nguyen in @xcite in which special structures were considered due to the use of the separation of variables or the blow up of the power was investigated via spectral theory or variational method . in @xcite ,
nguyen investigated cloaking a source via alr for a class of complementary media called the class of doubly complementary media for a general core - shell structure . in these works ,
the blow up of the power , the localized resonance , and the cloaking effect are studied .
the reader can find a recent survey on the mathematical aspects for nims in @xcite . in this paper
, we add to the list of applications of nims a new one namely cloaking * an arbitrary object via alr*. more precisely , we propose schemes for this type of cloaking and provide the analysis for them . before stating the result
, we introduce notations : given @xmath2 and @xmath3 , we denote @xmath4 the open ball in @xmath5 centered at @xmath6 and of radius @xmath7 ; when @xmath8 , we simply denote @xmath9 by @xmath10 . the first result in this paper
, whose proof is given in section [ sect - proof ] , confirms that an annular plasmonic structure @xmath0 cloaks small but finite size objects close to it in the quasistatic regime .
more precisely , we have [ thm - cs-1 ] let @xmath11 , @xmath12 , @xmath13 , and @xmath14 .
set @xmath15 and @xmath16 and let @xmath17 be a symmetric uniformly elliptic matrix - valued function defined in @xmath18 .
define @xmath19 i & \mbox { otherwise } , \end{array}\right . \quad
\mbox { and } \quad s_\delta = \left\{\begin{array}{cl } -1 - i \delta & \mbox { in } b_{r_2 } \setminus b_{r_1 } , \\[6pt ] 1 & \mbox { otherwise}. \end{array}\right.\ ] ] given @xmath20 with @xmath21 and @xmath22 , let @xmath23 be respectively the unique solution to the equations @xmath24 and @xmath25 for any @xmath26 , there exists @xmath27 , which depends only on @xmath28 and @xmath29 , such that if @xmath30 then , for @xmath2 , @xmath31 where @xmath32 is a positive constant independent of @xmath33 , @xmath34 , @xmath35 , @xmath36 , and @xmath37 . in @xmath38 ( the red region )
cloaks arbitrary objects located nearby in @xmath18 .
the medium on the left is equivalent to the homogeneous medium @xmath39 on the right for sources far away from the plasmonic structure.,width=377 ] here and in what follows , @xmath40 for an observer outside @xmath41 , the medium @xmath42 in @xmath41 looks like @xmath39 in @xmath41 : the object @xmath17 in @xmath18 is cloaked .
no condition is imposed on @xmath17 ; any symmetric uniformly elliptic matrix @xmath17 in @xmath18 is allowed .
the cloak @xmath0 in @xmath38 is * independent * of the object .
it is interesting to note that the plasmonic structure @xmath0 in @xmath38 can used in lens devices using nims , see @xcite . in the two dimensional finite frequency regime , we obtain the following result whose proof is given in section [ sect - proof ] .
[ thm - cs-1-k ] let @xmath11 , @xmath43 , @xmath13 , and @xmath14 .
set @xmath15 and @xmath44 .
let @xmath17 be a symmetric uniformly elliptic matrix - valued function and let @xmath45 be a bounded real function both defined in @xmath18 such that @xmath17 is piecewise @xmath46 and @xmath47 is bounded below by a positive constant .
define @xmath48 \big ( i , r_2 ^ 4/|x|^4 \big ) & \mbox { in } b_{r_2 } \setminus b_{r_1 } , \\[6pt ] \big ( i , r_3 ^ 2/r_1 ^ 2 \big ) & \mbox { in } b_{r_1 } \setminus { \cal c } , \\[6pt ] \big ( i , 1\big ) & \mbox { otherwise } , \end{array}\right .
\quad \mbox { and } \quad s_\delta = \left\{\begin{array}{cl } -1 - i \delta & \mbox { in } b_{r_2 } \setminus b_{r_1 } , \\[6pt ] 1 & \mbox { otherwise}. \end{array}\right.\ ] ] given @xmath20 with @xmath21 ,
let @xmath49 be respectively the unique outgoing solution to the equations @xmath50 and @xmath51 for any @xmath26 , there exists @xmath52 , which depends only on @xmath28 and @xmath29 such that if @xmath53 then @xmath54 for some positive constant @xmath32 independent of @xmath33 , @xmath34 , @xmath35 , @xmath36 , and @xmath37 .
recall that a function @xmath55 @xmath56 for some @xmath2 which is a solution to the equation @xmath57 in @xmath58 is said to satisfy the outgoing condition if @xmath59 for an observer outside @xmath41 , the medium @xmath60 in @xmath41 looks like @xmath61 in @xmath41 : the object @xmath62 in @xmath18 is cloaked .
no condition other than the standard ones is imposed on @xmath17 and @xmath47 .
the cloak is * independent * of the object .
it is interesting to note that the plasmonic structure @xmath63 in @xmath38 can used in lens devices using nims , see @xcite . in section [ sect - general ] ,
we show in two and three dimensions that for a class of doubly complementary media , introduced in @xcite ( see definition [ def - dcm ] in section [ sect - general ] ) , an arbitrary object can be disappeared if it is small and located close to the plasmonic structure ( see theorem [ thm - main - g ] in section [ sect - general ] ) .
therefore , a medium in this class becomes a cloaking device which is * independent * of the object .
cloaking an object via alr is related to but different from cloaking using complementary suggested in @xcite and rigorously established in @xcite for related schemes .
both types of cloaking use the concept of complementary media to design cloaking devices .
nevertheless , in the cloaking using complementary media approach , one cloaks an object by using its complementary to cancel the effect of light ; hence the cloaking device depends on the object . cloaking an object via alr and cloaking a source via alr share some similar figures but have some different characters . in the two dimensional quasistatic regime
, one can use the plasmonic structure @xmath0 in @xmath38 as a cloaking device in both settings .
more general , doubly complementary media are used as a cloaking device in both types of cloaking ( see @xcite and section [ sect - general ] ) .
nevertheless , concerning cloaking a source via alr , the cloaking effect is * relative * in the sense that the source is cloaked after being renormalized so that the power remains finite . ]
( see e.g. @xcite ) .
concerning cloaking an object via alr , the cloaking effect is * not relative * in the sense that one does not need to renormalize the power .
in fact , the power is finite in the setting of cloaking an object via alr considered here ( see remarks [ rem1-power ] and [ rem2-power ] ) .
milton and nicorovici in @xcite questioned that whether or not the structure @xmath0 in @xmath38 would cloak small objects nearby in two dimensional quasistatic regime .
various numerical simulations on this effect were reported by bruno and lintner in @xcite ; even though the cloaking effect is still mysterious and questionable . in this paper ,
the guest of milton and nicorovici was confirmed ( theorem [ thm - cs-1 ] ) and similar phenomena are observed and analyzed in the finite frequency regime in both two and three dimensions ( theorems [ thm - cs-1-k ] and [ thm - main - g ] ) .
as mentioned , the plasmonic structures used in theorem [ thm - cs-1 ] and [ thm - cs-1-k ] can be used in lens devices using nims .
more generally , it is established in @xcite that doubly complementary media can act like lenses . in this paper
, we showed that they might become a cloak for small but finite size objects near by .
thus the results presented here show that the modification given in @xcite from the suggestions in @xcite is necessary to ensure the superlensing property ; otherwise , a lens can become a cloak ( see section [ sect - cm - superlensing ] ) .
in the same spirit , we as well show that it is necessary to modify the scheme of cloaking using complementary media suggested in @xcite as done in @xcite ( see section [ sect - cm - cloaking ] ) .
the analysis in this paper is on one hand based on the use of reflecting and removing singularity techniques introduced in @xcite and on the other hand involves a new type of three spheres inequality with partial data " and the use of conformal maps in two dimensions .
the rest of the paper is organized as follows . in section [ sect - general ]
, we present a scheme of cloaking an object via alr for a class of doubly complementary media and provide their analysis .
the main result of this section is theorem [ thm - main - g ] . in section [ sect-2d ] , we first provide the proof of theorems [ thm - cs-1 ] and [ thm - cs-1-k ] in section [ sect - proof ] .
we then show that the modified schemes considered in @xcite and @xcite for superlensing and cloaking using complementary media are necessary in sections [ sect - cm - superlensing ] and [ sect - cm - cloaking ] .
let @xmath64 and let @xmath65 be a real uniformly elliptic symmetric matrix - valued function and @xmath66 be a real function bounded below and above by positive constants both defined in @xmath5 ( @xmath67 ) , i.e. , @xmath65 is symmetric and for some @xmath68 @xmath69 assume that @xmath70 and @xmath71 the assumption is used for the uniqueness of outgoing solutions . in what follows , in the case @xmath72 ,
any matrix - valued function considered is assumed to satisfy .
let @xmath73 be smooth bounded simply connected open subsets of @xmath5 , and set , for @xmath74 , @xmath75 1 & \mbox { in } { \mathbb{r}}^d \setminus ( \omega_2 \setminus \omega_1 ) .
\end{array}\right.\ ] ] given @xmath76 with compact support and @xmath77 , and @xmath78 , let @xmath79 be the unique outgoing solution to @xmath80 in this section , we discuss the behavior of @xmath81 as @xmath82 when the medium inherits the doubly complementary property .
to this end , we first recall the definition of reflecting complementary media and doubly complementary media introduced in @xcite and @xcite respectively .
[ def - geo ] let @xmath83 be smooth bounded simply connected open subsets of @xmath5 .
the media @xmath84 in @xmath85 and @xmath86 in @xmath87 are said to be _ reflecting complementary _ if there exists a diffeomorphism @xmath88 such that @xmath89 , @xmath90 @xmath91 and the following two conditions hold : 1 ) there exists an diffeomorphism extension of @xmath92 , which is still denoted by @xmath92 , from @xmath93 for some @xmath94 ; 2 ) there exists a diffeomorphism @xmath95 such that @xmath96 , @xmath97 , and @xmath98 here and in what follows , if @xmath99 is a diffeomorphism and @xmath100 and @xmath101 are respectively a matrix - valued function and a complex function , we use the following standard notations @xmath102 conditions and are the main assumptions in definition [ def - geo ] . the key point behind
this requirement is roughly speaking the following property : if @xmath103 is a solution of @xmath104 in @xmath105 and if @xmath106 is defined in @xmath85 by @xmath107 , then @xmath108 in @xmath85 , @xmath109 on @xmath110 ( see e.g. lemma [ lem - to ] ) . here and in what follows , @xmath111 denotes the outward unit vector on the boundary of a smooth bounded open subset of @xmath5 .
hence @xmath112 in @xmath85 by the unique continuation principle .
conditions 1 ) and 2 ) are mild assumptions .
introducing @xmath113 makes the analysis more accessible , see @xcite and the analysis presented in this paper .
we are ready to recall the definition of doubly complementary media : [ def - dcm ] the medium @xmath114 is said to be _ doubly complementary _ if for some @xmath115 , @xmath84 in @xmath85 and @xmath116 in @xmath117 are reflecting complementary , and @xmath118 for some @xmath92 and @xmath113 coming from definition [ def - geo ] .
the reason for which media satisfying are called doubly complementary media is that @xmath86 in @xmath38 is not only complementary to @xmath84 in @xmath85 but also to @xmath84 in @xmath119 , a subset of @xmath120 ( see figure [ fig1 ] ) .
the key property behind definition [ def - dcm ] is as follows .
assume that @xmath121 is a solution of with @xmath122 and @xmath123 in @xmath124 .
set @xmath107 and @xmath125 . then @xmath126 satisfy the equation @xmath127 in @xmath85 , @xmath128 on @xmath110 , and @xmath129 on @xmath130 ( see e.g. lemma [ lem - to ] )
this implies @xmath131 in @xmath85 .
is doubly complementary : @xmath86 in @xmath87 ( the red region ) is complementary to @xmath132 in @xmath85 ( the grey region ) and @xmath133 with @xmath134 in @xmath135 ( the blue grey region).,width=226 ] taking @xmath11 and @xmath136 , and letting @xmath92 and @xmath113 be the kelvin transform with respect to @xmath137 and @xmath138 , one can verify that the media considered in theorems [ thm - cs-1 ] and [ thm - cs-1-k ] with @xmath139 are of doubly complementary property .
theorems [ thm - cs-1 ] and [ thm - cs-1-k ] reveal that doubly complementary media might cloak small but finite size arbitrary objects nearby .
[ rem - dcm ] given @xmath84 in @xmath5 and @xmath140 , it is not easy in general to verify whether or not @xmath60 is doubly complementary
. nevertheless , given @xmath141 and @xmath84 in @xmath85 , it is quite easy to choose @xmath84 in @xmath142 such that @xmath143 is doubly complementary .
one just needs to choose diffeomorphisms @xmath92 and @xmath113 as in definition [ def - geo ] and define @xmath144 in @xmath87 and @xmath145 in @xmath146 .
the following result established in ( * ? ? ?
* theorem 1.1 ) ( see also ( * ? ? ?
* theorem 2 and corollary 2 ) ) provide an interesting property of doubly complementary media .
[ pro - recall ] let @xmath147 , @xmath72 , @xmath148 , @xmath76 with compact support and @xmath149 , and let @xmath150 be the unique outgoing solution of .
assume that @xmath143 is doubly complementary .
then @xmath151 where @xmath152 is the unique outgoing solution of @xmath153 here @xmath154 ( g_*f_*a , g_*f_*\sigma ) & \mbox { in } \omega_3 .
\end{array}\right.\ ] ] the cloaking property for a class of doubly complementary media is given in the following [ thm - main - g ] let @xmath155 , @xmath156 .
set @xmath157 assume that @xmath143 is doubly complementary with @xmath158 , @xmath159 , and @xmath160 . let @xmath17 be a symmetric uniformly elliptic matrix - valued function and @xmath47 be a real function bounded above and below by positive constants both defined in @xmath161 .
assume that @xmath65 is @xmath46 in @xmath162 with @xmath163 .
define @xmath164 ( a , \sigma ) & \mbox { otherwise}. \end{array}\right.\ ] ] let @xmath72 , @xmath148 , @xmath76 with @xmath165 and let @xmath166 be the unique outgoing solution of @xmath167 for any @xmath26 , there exists a positive constant @xmath168 depending only on @xmath169 , @xmath28 , and the elliptic and the lipschitz constant of @xmath65 in @xmath162 such that if @xmath170 then @xmath171 for some positive constant @xmath32 independent of @xmath33 and @xmath34 where @xmath152 is the unique outgoing solution of .
the geometry of the cloak is given in figure [ fig2 ] .
for an observer outside @xmath41 , the medium @xmath60 in @xmath41 looks like @xmath172 in @xmath41 , which is independent of @xmath62 in @xmath18 : the object @xmath62 in @xmath18 is cloaked .
for example , by choosing the medium @xmath84 in @xmath120 in such a way that @xmath173 and @xmath174 in @xmath124 , i.e. , @xmath175 and @xmath176 in @xmath120 then @xmath177 : the object @xmath62 in @xmath178 disappeared in comparison with the homogeneous medium @xmath179 .
the assumption that @xmath65 is @xmath46 in @xmath162 is required for the use of three spheres inequality . which has red boundary .
the plamonic structure is is charaterized by @xmath180 for some diffeomorphism @xmath181 .
the region @xmath182 is the rectangle containing the blue object .
any object inside this region ( for example the blue one ) is cloaked.,width=226 ] the rest of this section containing two subsections is devoted to the proof of theorem [ thm - main - g ] .
some useful lemmas are presented in the first subsection and the proof of theorem [ thm - main - g ] is given in the second subsection .
we first recall a three spheres inequality which is an immediate consequence of ( * ? ? ?
* theorem 2 ) .
[ lem - minhloc ] let @xmath183 , @xmath184 , @xmath185 , and let @xmath100 be a lipschitz uniformly elliptic symmetric matrix - valued function defined in @xmath186 , and @xmath187 .
assume that @xmath188 , depending only on the elliptic and the lipschitz constants of @xmath100 such that @xmath189 where , for a smooth manifold @xmath190 of @xmath5 with or without boundary , @xmath191 here @xmath192 is a positive constant depending on the elliptic and the lipschitz constants of @xmath100 , and the constants @xmath193 @xmath194 @xmath195 , but independent of @xmath196 . in lemma [ lem - minhloc ] , the constant @xmath197 is independent of @xmath198 and @xmath199 and there is no requirement on the smallness of @xmath200 and @xmath201 .
this is different from the standard three sphere inequalities obtained previously in the literature ( see , e.g. , @xcite ) and plays an important role in our analysis .
applying lemma [ lem - minhloc ] , we can establish [ lem-3s-1 ] let @xmath147 , @xmath202 and set @xmath203 .
let @xmath100 and @xmath101 be respectively a lipschitz symmetric uniformly elliptic matrix - valued function and a real bounded function both defined in @xmath204 .
let @xmath205 and @xmath206 be such that @xmath207 define @xmath208 . for any @xmath209 , there exists a constant @xmath210 depending only on the lipschitz and the elliptic constants of @xmath100 such that if @xmath211 then , for some positive constant @xmath212 independent of @xmath196 and @xmath33 , @xmath213 lemma [ lem-3s-1 ] can be considered as a type of three sphrees inequality with partial data " since no information of @xmath196 on @xmath214 is required .
lemma [ lem-3s-1 ] plays an important role in the proof of theorem [ thm - main - g ] .
* we first consider the case @xmath123 .
fix @xmath215 such that @xmath216 on @xmath217 and @xmath218 .
let @xmath219 be such that @xmath220 @xmath221 = \varphi v , \quad [ a \nabla w \cdot \nu ] = \varphi a \nabla v \cdot \nu \mbox { on } \gamma.\ ] ] here and in what follows , on @xmath222 , @xmath223 : = u |_{{\mathbb{r}}^d \setminus d } - u |_{d}$ ] for an appropriate function @xmath224 ; similar notation is used for @xmath225 $ ] on @xmath222 .
since @xmath226 it follows that @xmath227 here and in what follows @xmath192 denotes a positive constant independent of @xmath196 .
define @xmath228 w & \mbox { in } b_{4r/5 } \setminus d. \end{array}\right.\ ] ] note that @xmath229 since @xmath216 in @xmath230 , it follows from , , and that @xmath231 , @xmath232 and @xmath233 set @xmath234 , @xmath235 , @xmath236 , and @xmath237 .
applying lemma [ lem - minhloc ] , we have @xmath238 with @xmath239 for some @xmath240 independent of @xmath196 , @xmath7 , @xmath241 , and @xmath242 . from
, we derive from lemma [ lem - compact ] below that , for some positive constant @xmath243 , @xmath244 with @xmath245 or @xmath201 .
it follows from that @xmath246 note that @xmath247 can be chosen arbitraly close to @xmath248 by taking @xmath249 and large @xmath241 in .
the conclusion in the case @xmath123 follows from , , and by the definition of @xmath250 in ; in particular , @xmath251 in @xmath252 .
we next consider the case @xmath205 .
let @xmath253 be such that @xmath254 then @xmath255 the conclusion in the general case now follows from the case @xmath123 by applying the previous case for @xmath256 .
the proof is complete .
@xmath257 in the proof of lemma [ lem-3s-1 ]
, we use the following result [ lem - compact ] let @xmath147 , @xmath2 , @xmath100 and @xmath101 be respectively a lipschitz symmetric uniformly elliptic matrix - valued function and a real bounded function both defined in @xmath10 .
let @xmath258 be such that @xmath259 then , for some positive constant @xmath260 , independent of @xmath196 , @xmath261 * proof .
* the second inequality of is a consequence of the trace theory .
it remains to establish the first inequality of .
we first prove that @xmath262 by contradiction .
in this proof , @xmath192 denotes a positive constant independent of @xmath196 and @xmath263 .
suppose that it is not true .
there exists some sequence @xmath264 such that @xmath265 and @xmath266 multiplying the equation of @xmath267 by @xmath268 ( the conjugate of @xmath267 ) , integrating in @xmath10 , and using , we obtain @xmath269 thus @xmath270 is bounded in @xmath271 . without loss of generality , one may assume that @xmath270 converges to @xmath196 weakly in @xmath271 and strongly in @xmath272 to @xmath196 for some @xmath258 .
it is clear from and that @xmath273 by the unique continuation principle , we have @xmath274 this contradicts the fact @xmath275 .
hence is established .
the first inequality of is now a consequence of after multiplying the equation of @xmath196 by @xmath276 and integrating in @xmath10 .
@xmath257 we next recall a stability result of established in ( * ? ? ?
* lemma 2.1 ) .
[ lem1 ] let @xmath277 , @xmath72 , @xmath278 , @xmath279 , @xmath280 with support in @xmath281 . for @xmath282
, there exists a unique outgoing solution @xmath283 of .
moreover , @xmath284 for some positive constant @xmath32 independent of @xmath285 and @xmath34 .
we end this section by recalling the following change of variables formula ( * ? ? ?
* lemma 2 ) which is used several times in this paper .
[ lem - to ] let @xmath286 be three smooth bounded open subsets of @xmath5 .
let @xmath287^{d \times d}$ ] , @xmath288 and let @xmath99 be a diffeomorphism from @xmath289 onto @xmath290 .
assume that @xmath291 and set @xmath292 .
then @xmath293 for some @xmath294 , if and only if @xmath295 assume in addition that @xmath296 on @xmath297 .
then @xmath298 the proof uses the reflecting and removing localized singularity techniques introduced in @xcite . applying lemma [ lem1 ] ,
we have @xmath299 here and in what follows in this proof , @xmath192 denotes a positive constant independent of @xmath34 and @xmath33 .
set @xmath300 define @xmath301 where @xmath92 and @xmath113 come from the definition of doubly complementary media .
applying lemma [ lem - to ] , we obtain , since @xmath302 in @xmath303 ( the medium is of doubly complementary property ) @xmath304 @xmath305 this implies @xmath306 and @xmath307 applying lemma [ lem - to ] , we have , since @xmath308 and @xmath309 in @xmath303 and @xmath123 in @xmath41 , @xmath310 which yields @xmath311 recall that @xmath312 set @xmath313 . applying lemma [ lem-3s-1 ] with @xmath314 and @xmath315 for @xmath316
, we derive from , , and that if @xmath170 and @xmath241 is sufficiently large then @xmath317 define @xmath318 and @xmath319 , and set @xmath320 u_{2 , \delta } + u_\delta - u_{1 , \delta } & \mbox { in } o , \\[6pt ] u_{2 , \delta } & \mbox { in } o_2 .
\end{array}\right.\ ] ] then @xmath321 for all @xmath322 and @xmath323 is an outgoing solution of the equation @xmath324 note that @xmath325 is uniformly elliptic and @xmath326 is bounded above and below by positive constants .
it follows that @xmath327 \le c_r \big ( \|[{\cal u}_\delta ] \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \big).\ ] ] by the definition of @xmath328 , we have @xmath329 \|_{h^{1/2}(\partial o_2 ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o_2 ) } = \|u_\delta - u_{1 , \delta } \|_{h^{1/2}(\partial o_2 ) } + \|a \nabla ( u_\delta - u_{1 , \delta } ) \cdot \nu\|_{h^{-1/2}(\partial o_2)}\ ] ] and @xmath330 \|_{h^{1/2}(\partial b_{r_3 } ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial b_{r_3 } ) } \\[6pt ] = \|u_{2 , \delta } - u_{1 , \delta } \|_{h^{1/2}(\partial b_{r_3 } ) } + \|a \nabla ( u_{2 , \delta } - u_{1 , \delta } ) \cdot \nu\|_{h^{-1/2}(\partial b_{r_3})}. \end{gathered}\ ] ] since @xmath331 , we derive from , , and that @xmath332 \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \le c \delta^{\alpha } \| u_\delta \|_{h^1(b_{r_3})}.\ ] ] it follows from and that , for @xmath333 , @xmath334 since @xmath335 . since @xmath336 , it follows that @xmath337 since @xmath338 , we derive from that , for @xmath2 , @xmath339 which yields , by , @xmath340 \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \le c \delta^{\alpha - 1/ 2 } \| f\|_{l^2}.\ ] ] hence @xmath328 is bounded in @xmath341 . without loss of generality , one may assume that @xmath342 weakly in @xmath343 as @xmath82 for any @xmath2 ; moreover , @xmath344 is the unique outgoing solution to the equation @xmath345 hence @xmath346 in @xmath5 .
since the limit is unique , the convergence holds for the whole family @xmath347 as @xmath82 . in other words , @xmath348 in @xmath349 . to obtain the rate of convergence ,
let us consider the equation of @xmath350 .
we have @xmath351 as in , we have @xmath352 \le c_r \big ( \|[{\cal u}_\delta ] \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \big).\ ] ] the conclusion now follows from and .
@xmath257 [ rem1-power ] the power of @xmath81 defined by @xmath353 is finite in the setting considered in theorem [ thm - main - g ] by .
it is showed in @xcite that resonance takes place for reflecting complementary media .
various conditions on the stability of the helmholtz equations with sign changing coefficients were given there and the necessity of the reflecting complementary property of media for the occurrence of the resonance was discussed .
this section containing three subsections is devoted to cloaking an object via alr in two dimensions and related problems . in the first section
, we present the proofs of theorems [ thm - cs-1 ] and [ thm - cs-1-k ] . in the second section
, we make some comments on superlensing using complementary media .
more precisely , we show that a modifications as in @xcite from the suggestion in @xcite is necessary : without the modification , a lens can become a cloaking device ( see proposition [ pro - lensing ] ) .
finally , we make some comments on cloaking using complementary media .
we prove that without the modification suggested in @xcite , cloaking effect might not be achieved for the schemes proposed in @xcite ( see proposition [ pro - cloaking ] ) .
[ lem-3s-2 ] let @xmath358 , @xmath359 , and define @xmath360 where @xmath361 is used in this paper : for @xmath354 with @xmath362 , @xmath363 , we define @xmath364 . ]
. set @xmath365 and @xmath366 .
let @xmath101 be a real bounded function defined in @xmath367 , @xmath368 , and @xmath369 be such that @xmath370 for any @xmath209 , there exists @xmath371 depending only on @xmath7 such that if @xmath372 , then , for some neighborhood @xmath373 of @xmath374 and for some positive constant @xmath192 both independent of @xmath196 and @xmath33 , @xmath375 * proof . *
the proof of lemma [ lem-3s-2 ] is similar to the one of lemma [ lem-3s-1 ] . for the convenience of the reader
, we present the proof .
we first consider the case @xmath123 in @xmath367 .
for the simplicity of notations , set @xmath376 .
we first assume that @xmath241 is sufficiently large such that @xmath377 ; this is possible since @xmath378 .
fix @xmath379 such that @xmath216 in @xmath380 and @xmath381 , and define @xmath382 .
set @xmath383 , the disk centered at @xmath384 and of radius @xmath385 , and extend @xmath101 by @xmath248 in @xmath386 .
we still denote the extension of @xmath101 by @xmath101 for simplicity of notations .
let @xmath387 be the unique solution of the system @xmath388 @xmath221 = \varphi_m v , \quad \mbox { and } \quad [ \partial_\nu w ] = \varphi_m \partial_\nu v \mbox { on } \gamma.\ ] ] since @xmath389 it follows that @xmath390 here and in what follows @xmath192 denotes a positive constant independent of @xmath196 .
define @xmath391 w & \mbox { in } o \setminus \hat d. \end{array}\right.\ ] ] then @xmath392 , @xmath393 and @xmath394 set @xmath395 , @xmath396 , @xmath397 , and @xmath398 . applying lemma [ lem - minhloc ] , we have @xmath399 where @xmath400 for some @xmath240 independent of @xmath196 , @xmath401 , and @xmath241 .
here we use the fact that , by lemma [ lem - compact ] , @xmath402 with @xmath403 , or @xmath201 . the conclusion in the case @xmath123 follows from , , , and the definition of @xmath250 in by noting that @xmath404 ( hence @xmath251 in @xmath252 ) , @xmath405 contains a neighborhood of @xmath406 , and @xmath247 can be taken aribitrary close to 1 if @xmath241 is large enough .
the conclusion in the general case follows from the case @xmath123 by applying the result in the case @xmath123 for @xmath256 where @xmath407 is the unique solution of the system @xmath408 the proof is complete .
@xmath257 [ cor1 ]
let @xmath409 , and define @xmath410 .
set @xmath411 and @xmath412 .
let @xmath101 be a real bounded function defined in @xmath413 , @xmath414 , and @xmath415 be such that @xmath416 for any @xmath209 , there exist two positive constants @xmath417 depending only on @xmath7 such that if @xmath418 , then , for some neighborhood @xmath373 of @xmath419 and for some positive constant @xmath192 , both independent of @xmath196 and @xmath33 , @xmath420 * proof . * set @xmath421 with @xmath361 for some @xmath422 . then @xmath423 satisfies the equation @xmath424 where , by lemma [ lem - to ] , @xmath425 for some positive constant @xmath426 such that @xmath427 ( such a constant @xmath426 exists since @xmath428 is a conformal map ) .
corollary [ cor1 ] is now a consequence of lemma [ lem-3s-1 ] by taking @xmath241 large enough .
@xmath257 [ lem-3s-3 ] let @xmath429 , @xmath359 , and define @xmath430 where @xmath361 . set @xmath365 and @xmath366 . let @xmath101 be a real bounded function defined in @xmath367 , @xmath368 , and
@xmath369 be such that @xmath370 for any @xmath209 , there exists @xmath371 depending only on @xmath7 such that if @xmath372 then , for some neighborhood @xmath373 of @xmath431 and for some positive constant @xmath192 both independent of @xmath196 and @xmath33 , @xmath432 [ cor2 ] let @xmath409 and define @xmath433 where @xmath361 .
set @xmath434 and @xmath435 .
let @xmath414 , @xmath101 be a real bounded function defined in @xmath413 , and @xmath415 be such that @xmath416 for any @xmath209 , there exist two positive constants @xmath436 depending only on @xmath7 such that if @xmath437 then , for some neighborhood @xmath373 of @xmath419 and for some positive constant @xmath192 both independent of @xmath196 and @xmath33 , @xmath438 * proof of theorem [ thm - cs-1-k ] . * the proof of theorem [ thm - cs-1-k ] is in the same spirit of the one of theorem [ thm - main - g ] .
applying lemma [ lem1 ] , we have @xmath439 here and in what follows in this proof , @xmath192 denotes a positive constant independent of @xmath34 and @xmath33 .
let @xmath92 and @xmath113 be the kelvin transform with respect to @xmath137 and @xmath138 .
define @xmath440 set @xmath441 by lemma [ lem - to ] , we have @xmath442 and @xmath443 combining , , and yields @xmath444 and @xmath445 applying lemma [ lem - to ] , we have @xmath446 and @xmath447 we derive that @xmath448 recall that @xmath449 denote @xmath450 the image of @xmath36 by @xmath92 and set @xmath451 .
we claim that there exist three positive constants @xmath452 , depending only on @xmath29 and @xmath453 , with @xmath454 , @xmath455 , such that , if @xmath456 , then for some neighborhood @xmath457 of @xmath458 , and some neighborhood @xmath459 of @xmath460 @xmath461 and @xmath462 we first derive from corollary [ cor1 ] . for
simplicity of notations , assume that @xmath463 .
applying corollary [ cor1 ] for @xmath464 , @xmath465 , we obtain assertion .
assertion can be obtained similarly by using corollary [ cor2 ] .
set @xmath466 it follows from and that @xmath467 and from and that @xmath468 define @xmath469 u_{2 , \delta } - ( u_{1 , \delta } - u_{\delta } ) & \mbox { in } o , \\[6pt ] u_{2 , \delta } & \mbox { in } o_2 . \end{array}\right.\ ] ] then @xmath470 for all @xmath2 and @xmath323 is an outgoing solution of the equation @xmath471 by the definition of @xmath472 and lemma [ lem - to ] .
this implies @xmath473 \le c \big ( \|[{\cal u}_\delta ] \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } + \| f\|_{l^2 } \big).\ ] ] from the definition of @xmath328 in , we derive from and that @xmath474 \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \le c \delta^{\alpha } \| u_\delta \|_{h^1(b_{r_3})}.\ ] ] combining and and using yield , for large @xmath322 such that @xmath475 , @xmath476 which implies , since @xmath477 , @xmath478 by , we obtain @xmath479 we derive from that @xmath480 \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \le c \delta^{\alpha - 1/2 } \| f\|_{l^2}.\ ] ] from and , without loss of generality , one may assume that @xmath342 weakly in @xmath481 as @xmath82 for any @xmath2 ; moreover , @xmath344 is the unique outgoing solution to the equation @xmath482 hence @xmath346 in @xmath483 .
since the limit is unique , the convergence holds for the whole family @xmath347 as @xmath82 . in this section ,
we revisit the construction of lenses using complementary proposed in @xcite and @xcite which has roots from @xcite .
we first consider the quasistatic regime . to magnify @xmath485-times
the region @xmath486 , for some @xmath487 and @xmath488 , following @xcite one puts a lens in @xmath489 whose medium is characterized by matrix @xmath0 with @xmath490 .
the lens construction in @xcite is related to but different from this .
our lens contains * two parts*. the first one is given by @xmath491 and the second one is @xmath492 here @xmath28 and @xmath29 are such that @xmath493 other choices for the first and the second layers are possible via the concept of complementary media ( see @xcite ) . given @xmath20 with compact support such that @xmath494 , and @xmath338 , we showed in @xcite , that @xmath495 here @xmath81 and @xmath496 are the unique solutions in @xmath497 of the equations @xmath498 where @xmath499 i & \mbox { otherwise}. \end{array}\right .
\quad \mbox { and } \quad s_\delta = \left\{\begin{array}{cl } -1 - i \delta & \mbox { in } b_{r_2 } \setminus b_{r_1 } , \\[6pt ] 1 & \mbox { otherwise}. \end{array}\right.\ ] ] for arbitrary uniformly elliptic matrix - valued function @xmath100 defined in @xmath500 : superlensing is achieved .
in fact , in @xcite , we considered bounded setting , nevertheless , the results stated here holds with similar proof . in comparison with the schemes suggested previously , the second part is added .
the second layer can be chosen thinner ( see @xcite ) .
nevertheless , as a consequence of theorem [ thm - cs-1 ] , the second layer in is necessary .
indeed , we have [ pro - lensing ] let @xmath501 , @xmath13 , and @xmath14 and let @xmath17 be a symmetric uniformly elliptic matrix - valued function defined in @xmath502 .
define @xmath503 i & \mbox { otherwise}. \end{array}\right .
\quad \mbox { and } \quad
s_\delta = \left\{\begin{array}{cl } -1 - i \delta & \mbox { in } b_{r_2 } \setminus b_{r_1 } , \\[6pt ] 1 & \mbox { otherwise}. \end{array}\right.\ ] ] given @xmath20 with compact support such that @xmath494 and @xmath504 with @xmath15 , let @xmath23 be the unique solutions of and respectively .
there exists @xmath505 depending only only on @xmath28 and @xmath29 , such that if @xmath506 then @xmath507 for an observer outside @xmath41 , the medium @xmath42 in @xmath41 looks like @xmath39 in @xmath41 : the object @xmath17 is disappeared .
thus the superlensing device becomes a cloak for small objects close to it .
similar conclusion also holds for the finite frequency regime as a consequence of theorem [ thm - cs-1-k ] . using theorem [ thm - main - g ] , in the same spirit with what presented in this section
, one could argue to show that the necessity of the second layer in the lens construction suggested in @xcite in three dimensions .
this section deals with cloaking using complementary media . we show that a modification proposed in @xcite ( see also @xcite ) from the suggestion in @xcite on cloaking using complementary media is necessary : without a modification , cloaking might be not achieved .
we only consider here the two dimensional finite frequency regime .
the two dimensional quasistatic case holds similarly .
we first describe how to cloak the region @xmath508 for some @xmath509 using complementary media as proposed in @xcite .
assume that the medium in @xmath508 is characterized by a symmetric uniformly elliptic matrix - valued function @xmath17 and a real function @xmath47 bounded above and below by positive constants .
we only consider here the two dimensional case ; even the construction in three dimensions hold similarly . the idea suggested by lai et al . in @xcite
is to construct a complementary media in @xmath38 for some @xmath510 . our cloak proposed in @xcite
is related to but different from @xcite .
it consists of * two parts*. the first one , in @xmath38 , makes use of reflecting complementary media to cancel the effect of the cloaked region and the second one in @xmath500 is to fill the space which disappears " from the cancellation by the homogeneous medium @xmath179 . for the first part , we also modified the strategy in @xcite . instead of @xmath508 , we consider @xmath511 for some @xmath512 as the cloaked region in which the medium is given by @xmath513 ( i , 1 ) & \mbox { in } b_{r_3 } \setminus b_{2 r_2}. \end{array } \right.\ ] ] the complementary medium in @xmath38 with @xmath514 is @xmath515 where @xmath516 is the kelvin transform with respect to @xmath137 .
concerning the second part , the medium in @xmath500 is given by @xmath517 set @xmath518 \big({f^{-1}}_*a_1 , { f^{-1}}_*\sigma_1 \big ) & \mbox { in } b_{r_2 } \setminus b_{r_1 } , \\[6pt ] \big ( i , r_3 ^ 2/r_1^{2 } \big ) & \mbox { in } b_{r_1 } , \\[6pt ] ( i , 1 ) & \mbox { otherwise } , \end{array}\right .
\quad \mbox { and } \quad s_\delta = \left\{\begin{array}{cl } -1 - i \delta & \mbox { in } b_{r_2 } \setminus b_{r_1 } , \\[6pt ] 1 & \mbox { otherwise}. \end{array}\right.\ ] ] given @xmath20 with compact support such that @xmath338 , we showed in ( * ? ? ?
* theorem 1 ) ( see ( * ? ? ?
* theorem 1 ) for the case @xmath519 ) that , if @xmath520 is large enough , @xmath521 here @xmath81 and @xmath496 are the unique outgoing solutions in @xmath522 of @xmath523 for an observer outside @xmath41 , the medium @xmath143 acts like @xmath61 : cloaking is achieved .
[ pro - cloaking ] let @xmath525 , and @xmath450 with @xmath15 .
let @xmath17 be a symmetric uniformly elliptic matrix - valued function and @xmath47 be a positive function bounded below by a positive constant both defined in @xmath526 .
define @xmath84 by where @xmath527 ( i , 1 ) & \mbox { in } ( b_{r_3 } \setminus b_{r_2 } ) \setminus b(x_3 , r_0 ) .
\end{array}\right.\ ] ] given @xmath20 with compact support and @xmath504 , let @xmath528 be the unique outgoing solution to the equation @xmath529 where @xmath84 and @xmath530 are given by .
there exists @xmath505 depending only on @xmath28 and @xmath29 such that if @xmath506 , then @xmath531 here @xmath532 is the unique outgoing solution to the equation @xmath533 ( i , 1 ) & \mbox { otherwise}. \end{array}\right.\ ] ] * proof . *
the proof is similar to the one of theorem [ thm - cs-1-k ] even simpler .
we only sketch the proof .
we have , by lemma [ lem1 ] , @xmath534 here and in what follows in this proof , @xmath192 denotes a positive constant independent of @xmath34 and @xmath33 .
define @xmath535 where @xmath92 and @xmath113 are
the kelvin transforms with respect to @xmath137 and @xmath138 respectively . by lemma
[ lem - to ] , as in and we have @xmath536 and @xmath537 applying lemma [ lem - to ] , we have @xmath538 and @xmath539 we derive that @xmath540 for simplicity of notations , we assume that @xmath541 .
applying corollary [ cor2 ] for @xmath542 , @xmath543 , and @xmath544 , there exist two positive constants @xmath545 such that if @xmath506 then for some neighborhood @xmath546 of @xmath547 @xmath548 set @xmath549 and define @xmath550 u_{2 , \delta } & \mbox { in } b_{r_2 } , \\[6pt ] u_\delta & \mbox { otherwise . } \end{array}\right.\ ] ] then @xmath551 for all @xmath322 and @xmath323 is an outgoing solution of the equation @xmath552 and , as in , @xmath329 \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \le c \delta^{\alpha } \| u_\delta \|_{h^1(b_{r_3})}.\ ] ] it follows that , for large @xmath322 such that @xmath475 , @xmath553 since @xmath554 , as in the proof of theorem [ thm - cs-1-k ] , one deduces that @xmath323 is bounded in @xmath555 , @xmath329 \|_{h^{1/2}(\partial o ) } + \|[\hat a \nabla { \cal u}_\delta \cdot \nu ] \|_{h^{-1/2}(\partial o ) } \to 0,\ ] ] and @xmath556 in @xmath557 as @xmath82 .
the details are left to the reader .
@xmath257 h. ammari , g. ciraolo , h. kang , h. lee , and g. w. milton , _ spectral theory of a neumann - poincar - type operator and analysis of cloaking due to anomalous localized resonance _ , arch .
rational mech .
* 218 * ( 2013 ) , 667692 .
g. w. milton , n. a. nicorovici , r. c. mcphedran , and v. a. podolskiy , _ a proof of superlensing in the quasistatic regime , and limitations of superlenses in this regime due to anomalous localized resonance _ , proc .
ser . a * 461 * ( 2005 ) , 39994034 . g. w. milton , n. a. nicorovici , r. c. mcphedran , k. cherednichenko , and z. jacob , _ solutions in folded geometries , and associated cloaking due to anomalous resonance _ , new j. phys . * 10 * ( 2008 ) , 115021 .
nguyen , _ cloaking via anomalous localized resonance . a connection between the localized resonance and the blow up of the power for doubly complementary media _
, c. r. math .
paris * 353 * ( 2015 ) , 4146 .
nguyen , _ negative index materials and their applications : recent mathematics progress_. ( 2016 ) , chin . ann .
math . , to appear , http://cama.epfl.ch/files/content/sites/cama/files/documents/survey.pdf .
nguyen and h. l. nguyen , _ cloaking using complementary media for the helmholtz equation and a three spheres inequality for second order elliptic equations _ ,
b * 2 * ( 2015 ) , 93 - 112 . | in this paper , we present various schemes of cloaking an arbitrary objects via anomalous localized resonance and provide their analysis in two and three dimensions .
this is a way to cloak an object using negative index materials in which the cloaking device is independent of the object . as a result
, we show that in two dimensional quasistatic regime an annular plasmonic structure @xmath0 cloaks small but finite size objects near by .
we also discuss its connections with superlensing and cloaking using complementary media .
in particular , we confirm the possibility that a lens can act like a cloak and conversely .
this possibility was raised about a decade ago in the literature .
* key words : * cloaking , superlensing , complementary media , localized resonance , three spheres inequality , conformal maps . |
there is considerable interest to understand the effects on the properties of a particle coming from interactions with an environment .
examples of such problems abound in condensed matter ; the problem discussed here is that of an electron coupled to lattice vibrations , i.e. of electron - phonon coupling . for example , such coupling is believed to be relevant for understanding certain aspects of the high - temperature superconductors behavior in the underdoped limit ( where the `` particle '' coupling to phonons is the doping hole already dressed by interactions with the electrons in the lower hubbard band)@xcite but there are many examples of other materials characterized by strong - electron phonon coupling , including polymers like polyacetylene , nanotubes , c@xmath0 molecules and other fullerenes.@xcite other problems of the same general type regard electrons coupled to spin - waves of a magnetically ordered system,@xcite or to orbitrons , for example in manganites,@xcite or to some combination thereof . in this work , we focus on the simplest hamiltonian describing an electron on a lattice interacting with an optical phonon mode , namely the holstein model:@xcite @xmath1 the first term is the kinetic energy of the electron , with @xmath2 and @xmath3 being the electron creation and annihilation operators . for the single dressed particle ( known as polaron , in this case ) problem of interest to us , the spin of
the particle is irrelevant and we suppress its index .
@xmath4 is the free - particle dispersion . in all results shown here
, we assume nearest - neighbor hopping on a @xmath5-dimensional simple cubic lattice of constant @xmath6 and a total of @xmath7 sites with periodic boundary conditions , so that @xmath8 but our results are valid for any other dispersion . the second term describes a branch of optical phonons of energy @xmath9 ( we set @xmath10 throughout this paper ) . @xmath11 and @xmath12 are the phonon creation and annihilation operators .
the last term is the on - site linear electron - phonon coupling @xmath13 , written in @xmath14-space .
all sums over momenta are over the first brillouin zone , @xmath15 .
the quantity of interest to us is the green s function of the single dressed particle , or polaron , defined as:@xcite @xmath16 | 0 \rangle,\ ] ] where @xmath17 is the ground state of the zero - particle system , which is the vacuum .
@xmath18 is the time ordering operator , and @xmath19 . since @xmath20 and @xmath21 , eq . ( [ eq : green_time_ordered ] )
simplifies to : @xmath22 where @xmath23 is the heaviside function . in other words ,
only the retarded part contributes .
this green s function gives the amplitude of probability that an electron introduced in the system at @xmath24 and removed at a later time @xmath25 , leaves the system in its ground state .
the usefulness of this quantity can be appreciated using its lehmann representation:@xcite @xmath26 where @xmath27 and @xmath28 are the complete set of one - particle eigenstates and eigenenergies , @xmath29 , @xmath30 .
thus , the poles of the green s function give the _
whole one - particle spectrum _ , while the associated residues , called quasi - particle ( @xmath31 ) weights , give partial information on the nature of the eigenstates .
moreover , the imaginary part of this green s function , called the spectral weight , is directly measured experimentally through angle - resolved photoemission spectroscopy.@xcite there has already been a large amount of work devoted to understanding the properties of the holstein polaron , and we briefly review some of it here . most of it is numerically - intensive work .
some examples are ( i ) exact diagonalization ( ed ) methods.@xcite these are usually hampered by the fact that even for a finite lattice , the hilbert space is infinite due to the infinite number of possible phonon configurations .
thus , some truncation scheme is needed , but for small phonon frequencies and/or large couplings , the cpu resources needed and run times become prohibitive .
this has led to a number of ( ii ) proposals based on variational approaches to decide which phonon configurations should be included,@xcite as well as ( iii ) various quantum monte carlo ( qmc ) methods.@xcite a brief review of these is provided in ref . .
of special interest are the so - called diagrammatic monte - carlo simulations@xcite which calculate directly the green s function in imaginary time , by _ numerically _ summing all diagrams in the perturbational expansion .
we make extensive use of comparisons with low - energy results of this method from ref . .
while this method allows in principle the exact calculation of the green s function , the requirement of convergence for the propagator series usually means that only low - energy properties are shown .
finally , there are methods suitable for some particular cases , such as density - matrix renormalization group for one - dimensional systems,@xcite and dynamic mean - field theory ( dmft ) for infinite - dimensional systems.@xcite of course , given the long history of this problem , many analytical techniques have been applied to it with varying degrees of success ( for a review , see ref . ) .
first of all , the green s function is known exactly in two asymptotic limits .
if there is no coupling , @xmath32 , then the green s function is that of the free electron .
the ground - state is at @xmath33 and the spectrum consists of a continuous band extending from @xmath34 $ ] ( for the tight - binding model ) .
the so - called impurity limit , with @xmath35 , also has an exact solution , given by the lang - firsov formula@xcite @xmath36 this can be viewed as the strong - coupling limit , since for @xmath35 , @xmath37 becomes the important energy scale in the system . in this limit
the electron is localized at one site in real space , therefore it is fully delocalized in @xmath38-space , and the green s function is independent of @xmath38 . the spectrum has the ground - state ( gs ) at @xmath39 and an infinite sequence of equidistant levels spaced by @xmath9 above it .
this is extremely different from the free - particle spectrum , and it is of considerable interest to understand not only how the ground - state evolves from @xmath40 to @xmath41 as the coupling @xmath37 is increased , but also the evolution of all the higher - energy spectral weight from a continuous , finite - width band to an infinite set of discrete levels .
note that it is customary to describe the effective coupling as the ratio of the two asymptotic ground - state energies , using as a new parameter @xmath42 most of the numerical methods reviewed above calculate only gs or low - energy properties , given the significant cpu time and numerical resources needed to calculate the whole spectrum .
very recently , several sets of whole - spectrum results have become available,@xcite however only for a few points in the parameter space , and generally for low dimensions .
it is of obvious interest to find an analytical approximation for the green s function that is simple to estimate , so that the whole parameter space can be studied easily , but also with high accuracy .
this is precisely what we propose here ( a short version of this work has been published in ref . ) .
we call our approximation the momentum average ( ma ) approximation ; its essence consists in analytically summing _ all the diagrams _ in the diagrammatic expansion , but with each diagram simplified in a certain way . before introducing this method ,
we briefly review here the other two simple ( in terms of computational effort ) analytical approximations for the green s function of the holstein polaron , available in the literature .
the first is the self - consistent born approximation ( scba ) , which consists of summing exactly only the non - crossed diagrams .
the percentage of diagrams kept decreases fast with increasing order ( see table [ tab1 ] ) .
if the coupling is small , the sum is dominated by the low order diagrams and scba works reasonably well . at strong coupling ,
the contribution of higher order diagrams becomes essential , and scba is expected to fail ( see below ) . in this approximation , the green s function is written in terms of a self - energy : @xmath43 with the self - consistency condition @xmath44 note that @xmath45 is independent of @xmath38 .
this is a consequence of the simplicity of the holstein model : if either the coupling @xmath37 or the dispersion @xmath9 were functions of the phonon momentum @xmath46 , the scba self - energy would depend explicitly on @xmath38.@xcite @xmath45 can be expressed@xcite as a function of the average of the free propagator over the brillouin zone ( bz ) and can be evaluated very efficiently .
.comparison between the total number of diagrams of a given order in the proper self - energy @xmath47 vs. the number of diagrams of a given order kept within scba .
[ cols="^,^,^,^,^,^,^,^,^",options="header " , ] the other simple analytical approximation for the green s function of the holstein model is the generalized lang - firsov ( lf ) expression.@xcite it is reminiscent of the lang - firsov expression of eq .
( [ eq : green_lf_zerohopping ] ) : @xmath48 this expression is exact in both asymptotic limits @xmath49 ( @xmath32 ) and @xmath50 ( @xmath51 , but less accurate for finite @xmath52 and @xmath37 , as we show below .
the article is organized as follows . in section
ii , we derive the momentum average approximation .
its diagrammatic meaning is discussed in section iii , where we also estimate its corresponding spectral weight sum rules .
we show that ma satisfies exactly the first 6 sum rules , but more importantly , it remains highly accurate for higher order sum rules .
this is a strong argument in favor of its accuracy .
the accuracy is gauged in more detail in section iv , where we compare the ma predictions against the scba and generalized lf predictions , but also against a host of numerical results .
this indeed demonstrates that the ma approximation is remarkably accurate for all parameter values , especially given its simplicity . in this section
we also present some new results regarding various properties of the holstein polaron .
finally , section v contains our summary and conclusions .
as is always the case for green s functions , one can use the equation of motion technique to generate an infinite hierarchy of coupled equations for an infinite number of related green s functions .
we derive it here for the holstein polaron . in the frequency domain ,
this approach is equivalent to using repeatedly dyson s identity @xmath53 , which holds for any green s operators @xmath54^{-1}$ ] , @xmath55^{-1}$ ] and for any hamiltonians @xmath56 . as customary
, we take @xmath57 to be the electron - phonon interaction . applying dyson s identity once
, we obtain : @xmath58,\ ] ] where @xmath59 is the free particle green s function .
we made use of the equality @xmath60 and defined a new green s function : @xmath61 @xmath62 is related to the amplitude of probability to start with the electron and a phonon at the initial time , and find only the electron in the system at the final time .
its own equation of motion relates back to @xmath63 but also to a new green s function with two phonons initially . in general ,
if we define : @xmath64 where @xmath65 is the total momentum of the @xmath66 initial phonons , using dyson s identity we find its equation of motion to be ( @xmath67 ) : @xmath68,\end{gathered}\ ] ] i.e. related to the green s functions with @xmath69 and @xmath70 initial phonons .
( [ eq : g ] ) and ( [ eq : f ] ) form the exact infinite hierarchy of coupled equations whose solution is the holstein polaron green s function @xmath71 .
obviously , this system of coupled equations can be solved trivially in the limit @xmath72 , in which case @xmath73 directly from eq .
( [ eq : g ] ) .
an exact solution equal to the lang - firsov result must also exist if @xmath74 .
indeed , in this limit all green s functions become independent of all momenta , and eqs .
( [ eq : g ] ) and ( [ eq : f ] ) simplify to : @xmath75\\ f_n(\omega ) = gg_0(\omega - n\omega)\left [ \frac{n}{\sqrt{n}}f_{n-1}(\omega ) + \sqrt{n}f_{n+1}(\omega ) \right ] , \nonumber\end{gathered}\ ] ] where @xmath76 .
these recurrence equations can be solved in terms of continued fractions .
we briefly review the solution here .
we suppress the functional notation and rewrite @xmath77 , where @xmath78 , @xmath79 . on physical grounds we expect that @xmath80 becomes vanishingly small for a sufficiently large @xmath66 , since it describes physical processes which are less and less likely .
this allows one to solve these equations iterationally starting from this sufficiently large @xmath66 : @xmath81 , to find after solving for @xmath82 , @xmath83 , etc .
, that : @xmath84 allowing the continued fraction to be infinite instead of truncated after @xmath66 steps gives the exact solution .
recalling that @xmath85 , one can now solve for @xmath86 . with the original notation
, we find : @xmath87 after some further work , this can indeed be shown to equal the lang - firsov expression of eq .
( [ eq : green_lf_zerohopping ] ) .
this exact hierarchy of coupled equations [ eqs .
( [ eq : g ] ) and ( [ eq : f ] ) ] can also be solved in the general case of finite @xmath37 and finite @xmath52 by iteratively solving for @xmath62 , then @xmath88 , etc . , and removing them from this coupled system .
it is straightforward to verify that the solution obtained in this case is the diagrammatic expansion , which can be partially resummed to the expected form : @xmath89^{-1 } -\sigma({\mathbf{k}},\omega)},\ ] ] where the self - energy @xmath47 is the sum of all proper self - energy diagrams , the first few of which are shown in fig .
[ fig : self_energy ] . while this solution as a sum of an infinite number of diagrams is exact , it is clearly not very useful if the sum can not be performed .
one typical strategy in such cases is to sum only a subset of these diagrams ( _ e.g. _ the non - crossed ones , in the self - consistent born approximation ) .
this is reasonable when one can argue that the diagrams kept contribute much more than the neglected diagrams , which is not the case for scba in this problem ( see below ) .
we propose a new strategy , explained in the next subsection , to find an approximative solution of these equations in the case of finite @xmath52 and finite @xmath37 . to obtain an approximate solution in the case of finite @xmath52 and @xmath37
, we proceed as follows .
we first note that @xmath63 depends on @xmath62 only through its average over the brillouin zone @xmath90 , @xmath91.\ ] ] we define the _ momentum averaged _ green s functions : @xmath92 where all momenta sums run over the bz , and attempt to express eqs .
( [ eq : f ] ) in terms of these simpler quantities , by performing the corresponding momenta averages on both sides .
while the first term on the right - hand side can be written in terms of @xmath93 exactly , the second term requires an approximation , which we choose to be : @xmath94 where @xmath95 is the free propagator momentum - averaged over the bz .
one justification for replacing @xmath96 by its momentum average @xmath97 in eq .
( [ app ] ) is that @xmath98 takes , with equal probability , any value in the bz .
moreover , in the impurity limit @xmath35 where all green s functions are momentum independent , eq .
( [ app ] ) is exact .
this suggests that our approach should be reasonable at least in the strong - coupling limit @xmath99 . in a more practical sense
, this approximation allows us to write @xmath100 in terms of @xmath101 and @xmath102 only , which is what we require to be able to obtain an analytical expression for @xmath103 .
we discuss the meaning and consequences of this approximation in more detail below . after the approximation of eq .
( [ app ] ) , eqs .
( [ eq : f ] ) become : @xmath104.\ ] ] together with eq .
( [ a1 ] ) these simplified recurrence relations can be solved similarly to the @xmath35 case .
we find : @xmath105 where the self - energy is , within the ma approximation , @xmath106 this is the main new result of this work . as discussed , the ma approximation becomes exact in the limit of zero hopping ( @xmath35 ) , where @xmath107 , but also for zero coupling , @xmath32 . in the following sections we show that the range of validity of this approximation extends well beyond these asymptotic limits , and that in fact the ma expression is reasonably accurate over the entire parameter space . note that although the ma self - energy looks similar to the dmft self - energy,@xcite this is in fact a very different approximation .
this issue is discussed in appendix [ dmft ] .
to understand the diagrammatical meaning of the ma approximation we expand eq .
( [ eq : sigma_ma ] ) in powers of @xmath108 : @xmath109 \\ + g^6 \left [ 4 \bar{g}_0 ^ 3(\omega-\omega ) \bar{g}_0 ^ 2(\omega-2\omega ) + 6 \bar{g}_0 ^ 2(\omega-\omega ) \right . \\ \left .
\times \bar{g}_0 ^ 2(\omega-2\omega ) \bar{g}_0(\omega-3\omega)\right ] + \mathcal{o}(g^8).\end{gathered}\ ] ] and analyze the various terms .
first , one can verify that the ma approximation generates the correct number of proper self - energy diagrams to all orders .
indeed , there is one term of order @xmath108 , two terms of order @xmath110 , 4 + 6=10 terms of order @xmath111 , and so on ( see table [ tab1 ] ) .
moreover , to each of these terms we can associate an ma diagram .
these have the same topology as the exact proper self - energy diagrams .
the difference is that each free propagator @xmath112 in the exact self - energy diagrams is replaced with a momentum averaged free propagator @xmath113 ( with the correct frequency ) in each ma diagram . using eq .
( [ eq : g0 ] ) , the first - order self - energy diagram is ( see fig . [ fig : self_energy ] ) : @xmath114 i.e. @xmath115 , and thus ma is exact to first order .
differences appear from the second order diagrams , where the two exact contributions ( see fig . [
fig : self_energy ] ) : @xmath116\end{gathered}\ ] ] are replaced , within ma , by two equal contributions : @xmath117 comparing eq .
( [ eq : second_exact ] ) to eq .
( [ eq : second_ma ] ) , we see that the ma self - energy diagrams have the correct number of free propagators with the correct frequencies , however the momenta of the free propagators are un - correlated and individually averaged over .
it is as if there is no connection between the momentum carried by a cloud phonon when it is emitted and when it is re - absorbed by the electron .
precisely the same holds for all higher order self - energy diagrams . to gain a better understanding of the difference between the exact and the ma diagrams ,
let us further analyze the dependence on @xmath52 of eqs .
( [ eq : second_exact ] ) and ( [ eq : second_ma ] ) . for @xmath35
the expressions are identical , because the free propagators become independent of momenta and , as already discussed , ma becomes exact . for finite @xmath52
, we expand each free propagator as : @xmath118,\ ] ] where @xmath119 . inserting this expansion into eqs .
( [ eq : second_exact ] ) and ( [ eq : second_ma ] ) and collecting powers of @xmath52 yields the following .
all @xmath120 terms vanish in both the exact and the ma diagrams because they are proportional to a @xmath121 .
in fact , all odd - order powers in @xmath52 vanish because @xmath122 .
next , consider terms of order @xmath123 .
such terms arise either from expanding one free propagator to @xmath124 , or from expanding two different free propagators to @xmath120 .
the former case leads to the same result for both the exact and the ma diagrams , and we obtain 6 contributions proportional to @xmath125 .
the latter case , however , reveals a difference .
five of the six @xmath124 such contributions from the exact diagrams vanish because they involve propagators carrying different momenta , and , for example , @xmath126 .
the exception comes from the two outside free propagators of the non - crossed diagram , which carry the same momentum and result into another @xmath127 contribution .
this is absent in the ma approximation , where different propagators always carry different momenta .
it follows that the ma second order self - energy diagrams capture 6 out of the 7 finite @xmath124 contributions correctly .
similar considerations apply for higher order @xmath52 powers and for higher order diagrams , differences between the ma and the exact self - energy diagrams coming only from terms involving free propagators carrying equal momenta in non - crossed diagrams .
however , the error from such missed terms becomes smaller and smaller as one goes to higher order diagrams because the percentage of self - energy diagrams with one or more pairs of free propagators of equal momenta decreases exponentially .
we conclude that the ma approximation captures most of the @xmath52 dependence of each self - energy diagram , while summing over all diagrams .
this analysis suggests that ma should be quite accurate for any finite @xmath37 and @xmath52 values . in the next section
, we reinforce this conclusion by analyzing the sum rules of the spectral weight .
for an even better idea of the accuracy of the ma approximation , we consider the sum rules for the spectral weight @xmath128 : @xmath129 for a problem of this type ( single dressed particle ) , the sum rules can be calculated exactly to arbitrary order .
the usual approach is based on the equation of motion technique.@xcite we review it briefly here in order to make a few useful observations .
the key step is to rewrite @xmath130 , so that we have : @xmath131 the integral is now simply @xmath132 . using the definition of eq .
( [ eq : green_simplified ] ) , we find , for any @xmath133 : @xmath134 so that the sum rules simplify to : @xmath135 these vacuum expectation values can be evaluated directly with some effort .
we find @xmath136 , @xmath137 , @xmath138 , @xmath139 , etc .
one very important conclusion that can be drawn from this derivation is that these sum rules have the same functional dependence on the energy scales @xmath140 anywhere in the parameter space . of course , in various asymptotic regimes ,
different terms dominate the overall value ( _ e.g. _ @xmath141 if @xmath142 while @xmath143 if @xmath144 ) .
however , this shows that if one can evaluate the sum rules _ exactly _ in any asymptotic regime , for instance by using perturbation theory , the results hold true _ everywhere _ in the parameter space , even where perturbation fails .
the second important conclusion one can draw from eq . ( [ sr ] ) is that each term in @xmath145 is proportional to @xmath146 , where @xmath147 are integers , i.e. the sum rule @xmath148 is a polynomial of total order @xmath66 in the energy scales of the problem .
more complicated dependence on @xmath149 , for example through @xmath150 , simply can not appear ( see below ) .
the first conclusion suggests an alternative derivation of the sum rules , which can also be used for the ma and scba sum rules .
namely , we use the diagrammatic perturbational expansion of the green s function valid for @xmath144 to evaluate directly the integrals @xmath151 and retain the imaginary part . in this case , @xmath152 , where @xmath153 are all green s function diagrams of order @xmath66 , i.e. containing @xmath66 phonon lines . the multiplicity @xmath154 .
each diagram @xmath155 is a product of @xmath156 free propagators , summed over internal phonon momenta . since for large frequency each @xmath157
, it follows that for @xmath158 , each @xmath159 . since any integrand that decreases faster than @xmath160 has a vanishing contribution to eq .
( [ eq : sum_rule ] ) , it follows that the diagrams of order @xmath66 only contribute to the sum rules @xmath161 with @xmath162 .
thus , even though @xmath103 is the sum of an infinite number of diagrams , only a finite number of them , of low order , contribute to any given sum rule and the calculation can be done .
the same holds true for the ma sum rules , the only difference being that the self - energy parts in the green s functions diagrams are replaced with the corresponding ma self - energy parts .
let us analyze the differences between contributions of the exact and of the ma diagrams to the sum rules .
it is straightforward to verify that : @xmath163 @xmath164 and @xmath165.\end{gathered}\ ] ] both the exact and the ma diagrams of order @xmath66 contain products of the general form @xmath166 .
some of the free propagators are actually @xmath112 ( always the first and the last one , but there can also be intermediary ones connecting proper self - energy parts ) .
all other free propagators have momenta dependent on the phonon momenta , which are summed over ( in the exact diagrams ) , or are individually averaged over ( in the ma diagrams ) .
since there is one - to - one correspondence between the number of exact vs. ma diagrams and their topologies , and since @xmath167 , it follows that the exact and the ma diagrams of order @xmath66 give precisely the same contributions to @xmath168 and @xmath169 .
differences appear in the contribution to @xmath170 if there is at least one pair of propagators in any of the self - energy parts of the exact diagram that carry the same momenta . in this case , the corresponding @xmath171 averages to @xmath172 when the sums over phonon momenta are carried for the exact diagrams , whereas these terms always average to zero for the ma diagrams .
since most free propagators in the self - energy parts have different momenta , such differences are quantitatively small .
this is especially true for large phonon frequencies @xmath9 , where the contributions proportional to @xmath9 captured correctly by ma scale like some power of @xmath156 .
this analysis can be continued for higher sum rules , with similar conclusions .
we can now summarize our findings .
only the @xmath173 order green s function diagram ( the free propagator @xmath174 ) contributes to @xmath175 and @xmath176 .
since this is included correctly in the ma and the scba cases , both give the correct @xmath175 and @xmath176 .
in fact , this diagram contributes an @xmath177 to @xmath145 , which is always the leading power in @xmath52 contribution .
the 1@xmath178 order diagram is also exact in both the ma and scba cases , therefore both give the correct @xmath179 and @xmath180 sum rules as well .
differences appear from @xmath181 onwards . because scba only keeps 2 out the 3 second order green s function diagrams , and 5 out of 15 third order diagrams , etc . , the leading terms in @xmath37 are @xmath182 instead of @xmath183 in @xmath184 , @xmath185 instead of @xmath186 in @xmath187 , etc .
this shows that scba fails all sum rules with @xmath188 significantly in the strong coupling regime where the term proportional to @xmath189 gives the most significant contribution to @xmath168 ( similar conclusions hold for odd sum rules ) .
so even though scba always satisfies exactly the first 4 sum rules , it is a bad approximation for large @xmath37 , where big discrepancies appear for @xmath188 . .
results are for 1d and @xmath190 , @xmath191 and @xmath192 , 2 . ]
ma satisfies @xmath184 and @xmath193 exactly as well , because these only depend on having the correct number and topology for the 2nd order diagrams .
ma fails from @xmath187 onwards , however in a very different manner than scba .
the leading term in @xmath111 has the correct prefactor , because ma has the correct number of 3rd order diagrams .
the error comes from the 2nd order diagram containing the non - crossed self - energy diagram , as discussed .
indeed , instead of the exact sum rule : @xmath194 \\ + g^4 ( 18dt^2 + 12 \varepsilon_{{\mathbf{k}}}^2 + 22 \varepsilon_{{\mathbf{k}}}\omega + 25
\omega^2 ) + 15g^6,\end{gathered}\ ] ] ma finds a sum rule @xmath195 .
the leading terms in the @xmath144 and @xmath196 limits are always exact ( as expected , since ma becomes exact in these limits ) , and this is true for all orders @xmath66 . for @xmath197
some of the cross terms are missing , but these are a minority related to non - crossed diagrams , as explained .
we therefore expect the ma sum rules to remain highly accurate for larger @xmath66 values . that this is indeed true for higher sum rules
is shown numerically in figs .
[ sumrules0 ] and [ sumrulespi ] , where we plot the ratio of the ma respectively scba sum rules , and the exact sum rules of same order @xmath66 .
the results shown are for 1d and @xmath198 , but similar trends are found in the other cases . for @xmath190
all the spectral weight is at negative frequencies , therefore @xmath199 alternate signs for even / odd @xmath66 , and this is reflected in the non - monotonic behavior with @xmath66 . for @xmath200 ,
most of the weight is at positive frequencies and sum rules are always positive .
the magnitude of the exact sum rules increases roughly exponentially with @xmath66 , for instance for @xmath201 and @xmath191 , @xmath202 . for @xmath190 and @xmath203 ,
both ma and scba are reasonably accurate , with a slight edge for ma at higher @xmath66 .
however , ma is clearly much more accurate for all the other cases shown , and its accuracy is expected to improve even more as one moves further into the asymptotic regions of weak or strong coupling . .
results are for 1d and @xmath200 , @xmath191 and @xmath192 , 2 . ] before ending this section , one more issue needs to be addressed . it is obvious that the ma sum rules must capture the contributions to @xmath145 proportional respectively to @xmath204 and @xmath205 exactly , since ma is exact for both @xmath32 and @xmath35 .
one may assume that this alone is sufficient for a good interpolation at finite @xmath52 and @xmath37 . that this is not so is shown by the generalized lf approximation , which is also exact for @xmath35 or @xmath32 .
however , in this approximation one finds @xmath206 , etc . @xmath207 and all higher sum rules have unacceptable dependence on the energy scales @xmath37 and @xmath208 ( see second observation above ) , even though they become exact asymptotically .
as shown in the next section , this approximation indeed performs rather poorly for finite @xmath52 and @xmath37 .
we conclude that while ma satisfies exactly the first 6 sum rules , it remains accurate for all higher order sum rules , and is asymptotically exact .
this is another argument in favor of the accuracy of this approximation over the whole parameter space . in the next section ,
we compare the ma predictions to those of existing numerical simulations to further support this claim .
we first list the explicit expressions of the momentum averaged green s function @xmath113 of eq .
( [ eq : g0 ] ) . for nearest - neighbor hoping , it is straightforward to derive : @xmath209 @xmath210 and @xmath211 respectively , where @xmath212 and @xmath213 is the complete elliptical function of the first kind.@xcite these integrals can be performed numerically very efficiently . more generally , for any free electron dispersion @xmath214 to which corresponds the free electron density of states ( dos ) @xmath215 , we have @xmath216 ( also see appendix [ dmft ] ) .
the self - energy @xmath217 is then calculated easily from eq .
( [ eq : sigma_ma ] ) by truncating the continued fraction to a high - enough level . for an error of order
@xmath218 it is necessary to go to a level with @xmath66 such that @xmath219 .
using the fact that for large enough @xmath66 we can approximate @xmath220 , it follows that we must have @xmath221 this result is expected , since @xmath222 is roughly the average number of phonons in the polaron cloud ( see below ) .
this condition shows that all diagrams with at least that many phonons have to be included . in practice
, we always use @xmath66 large enough so that the change in @xmath217 after doubling @xmath66 is below a threshold much smaller than @xmath223 .
all ma error bars in the figures we show are less than the thickness of lines / symbols used for the plots . with an explicit form for @xmath217 we are now in a position to calculate the green s function @xmath224 and
extract various polaron properties .
we begin by discussing polaron ground - state properties .
most of these are already known from numerical studies , but they give us an opportunity to further test the accuracy of the ma approximation . given the simplicity and efficiency of the ma approximation , we can also investigate higher dimensionality and larger parameter ranges than typical numerically intensive approaches . in this section
, we use for comparison 1d and 2d numerical results obtained from diagrammatic quantum monte carlo ( qmc ) simulations.@xcite for @xmath190 , we track the energy and weight of the lowest pole in the spectral weight , which give the ground - state energy @xmath225 and the ground - state quasi - particle weight @xmath226 using the hellmann - feynman theorem , @xcite we then find the average number of phonons in the ground - state to be @xmath227 note that one can also calculate the correlation function : @xmath228 just as easily .
we do not show it here because we do not have the corresponding numerical data for the comparison , however some typical results for this quantity are shown at the end of this section .
also note that all these quantities can be calculated similarly for other eigenstates .
we will show such results in other sections .
we also show the effective mass , @xmath229 .
because the ma self - energy is momentum independent for this simple holstein model , one has:@xcite @xmath230 this result also gives us a consistency check on our calculations . for ma
we generally show effective mass results obtained from the first equality .
are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] a comparison of the results for these four quantities as obtained with qmc ( black circles ) and the different approximations is shown in fig .
[ fig:1d_e ] .
as expected , scba ( blue line ) fares well at small couplings @xmath235 but very poorly at strong couplings .
the generalized lf ( green line ) is asymptotically exact , but quite wrong at finite @xmath235 .
because these are gs properties , we can also use perturbation theory in the two asymptotic limits to estimate them . at weak coupling
we use the rayleigh - schrdinger perturbation theory ( rs , violet line ) which gives the lowest energy for a state with total momentum @xmath14 as:@xcite @xmath236 at strong couplings we use the second order perturbation theory result ( pt , cyan line):@xcite @xmath237 the gs energy is simply @xmath238 .
@xmath232 is obtained from @xmath225 as before ; @xmath229 can be evaluated from the second derivative of @xmath239 with respect to @xmath14 , and @xmath231 is extracted from the effective mass [ eq .
( [ eq : effective ] ) ] .
[ fig:1d_e ] shows that one or the other of these perturbational values describe the gs energy @xmath225 quite well , especially for the larger @xmath9 value .
however , the agreement for the other quantities is somewhat poorer , especially at strong couplings ( pt and lf give identical results for @xmath231 ) .
part of the reason is that the @xmath123 term in eq .
( [ eq : e_pt ] ) has in fact a more complicated dependence of @xmath37 and @xmath240 , which is only asymptotically equal to the one used here.@xcite more importantly , neither perturbational theory describes well the transition area , or can be easily applied to high energy states . ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath242.,title="fig : " ] ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath242.,title="fig : " ] ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath242.,title="fig : " ] ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath242.,title="fig : " ] ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath243.,title="fig : " ] ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath243.,title="fig : " ] ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath243.,title="fig : " ] ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 as a function of @xmath235 and of @xmath241 , for @xmath243.,title="fig : " ] clearly , ma ( red line with red symbols , for easier comparison with qmc data ) has the best agreement with the qmc data . as expected from the sum rule analysis and the discussion on the convergence of @xmath244 , ma improves for larger @xmath9 .
the worst disagreements we ever found are the ones shown in the intermediary @xmath235 regime for @xmath245 . even there , the error in the gs energy is always below @xmath246 .
the @xmath31 weight has a more significant disagreement , however note that it indeed becomes asymptotically correct for @xmath247 and @xmath248 .
the second claim is supported by the @xmath229 data , which indeed shows convergence towards the expected pt values .
most importantly , even though it is quantitatively somewhat wrong in this intermediary regime for small @xmath9 , the ma approximation is the only one that clearly captures the crossover from the large to the small polaron , which is accompanied by the collapse of the @xmath31 weight and the increase in the number of trapped phonons and thus of the effective mass . the agreement with the qmc data
is significantly improved for 2d polarons , as shown in fig .
[ fig:2d_e ] . here
ma gives excellent agreement at all couplings @xmath235 .
this is all the more remarkable when considering what a numerically trivial task it is to evaluate the ma results compared to the qmc simulations .
the physics is similar to that seen in 1d , however the cross - over from the large ( light ) polaron at weak couplings , to the small ( heavy ) polaron at strong couplings , becomes somewhat sharper , especially for smaller @xmath241 values .
given the simplicity of ma , we can also generate contour plots of these quantities as a function of both @xmath235 and @xmath241 , and investigate the entire parameter space .
such results are shown in figs .
[ fig:1d_c ] and [ fig:2d_c ] for @xmath242 and 2 , respectively .
the cross - over from the large to the small polaron can now be tracked ( for instance , from the collapse of @xmath231 ) and one can quantitatively trust the results to a high degree .
the transition occurs for @xmath249 , although this shifts to lower values as @xmath9 decreases .
are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] are the ground state energy @xmath225 ; the @xmath31 weight @xmath231 ; the average number of phonons @xmath232 and the effective mass @xmath229 on a logarithmic scale .
the left panels correspond to @xmath233 and the right ones to @xmath234.,title="fig : " ] , for @xmath191 and @xmath250 , 3.,title="fig : " ] , for @xmath191 and @xmath250 , 3.,title="fig : " ] , for @xmath191 and @xmath250 , 3.,title="fig : " ] we also show ma results in 3d , see fig .
[ fig:3d_e ] .
here we do not have qmc results for comparison , however the good agreement with one or the other perturbational theories for most coupling strengths suggests that the ma results are probably even more accurate than in 2d .
this is consistent with the expectation of improved agreement in higher dimensions .
the crossover from large to small polaron becomes even sharper , especially for lower @xmath9 .
it is still located in the neighborhood of @xmath249 . to the best of our knowledge , the highly accurate three - dimensional results shown in fig .
[ fig:3d_e ] are the first of their kind .
this is likely due to the fact that the numerically intensive techniques require far too much computational effort to investigate such cases.@xcite using the hellmann - feynman theorem like in eqs .
( [ eq : phonons ] ) and ( [ eq : corr ] ) also allows us to separate the individual contributions of @xmath251 @xmath252 and @xmath253 to the total gs energy .
plots of these individual contributions as a function of coupling strength @xmath235 are shown in fig .
[ fig:1d_components_0 - 5 ] for @xmath254 . as expected , the kinetic energy is close to @xmath40 at weak couplings , but it becomes vanishingly small in the strong coupling limit , where the polaron becomes very heavy .
the phonon energy @xmath255 increases roughly like @xmath256 in the strong coupling limit , where @xmath257 .
it follows that the decrease in the total gs energy is due to the interaction term , as expected .
note that this energy is proportional to the correlation of eq .
( [ eq : corr ] ) . since @xmath258 in the strong coupling limit
( see agreement with pt results ) , it follows that this correlation becomes asymptotically equal to @xmath259 in the strong coupling limit . , @xmath260 , and the corresponding @xmath31 weight @xmath261 and average phonon number @xmath262 .
results are for @xmath263 , @xmath234 , @xmath264 and 1 ( for @xmath260 and @xmath261 ) as well as 1.96 , for @xmath262 . only half of the bz is shown .
, title="fig : " ] , @xmath260 , and the corresponding @xmath31 weight @xmath261 and average phonon number @xmath262 .
results are for @xmath263 , @xmath234 , @xmath264 and 1 ( for @xmath260 and @xmath261 ) as well as 1.96 , for @xmath262 .
only half of the bz is shown .
, title="fig : " ] , @xmath260 , and the corresponding @xmath31 weight @xmath261 and average phonon number @xmath262 .
results are for @xmath263 , @xmath234 , @xmath264 and 1 ( for @xmath260 and @xmath261 ) as well as 1.96 , for @xmath262 .
only half of the bz is shown .
, title="fig : " ] we can also calculate the same properties for the lowest energy state corresponding to each given momentum @xmath265 , to find the low - energy behavior of the polarons . in this section we present comparisons with available qmc results@xcite for 1d and 2d systems . in fig .
[ fig:1d_ek ] we show 1d results for the polaron dispersion @xmath260 , the associated @xmath31 weight @xmath261 and average phonon number @xmath262 for two couplings . for the very weak @xmath264 , we see that ma and scba are equally good at small @xmath266 , however for large @xmath266 ma overestimates the energy , such that the continuum which is expected to appear at a distance @xmath9 above the gs energy @xmath225 , is in this case pushed somewhat higher .
we will return to a discussion of this discrepancy later on .
as expected , the @xmath31 weight is large for small @xmath266 , where the main contribution to the eigenstate comes from the free electron state @xmath267 .
however , @xmath261 goes to zero for larger @xmath266 , since these are primarily linear combinations of states of type @xmath268 , as confirmed also by the the average phonon number of about unity .
note that the rs perturbation also works well for small @xmath266 , as expected .
it however breaks done at a finite @xmath266 where @xmath269 , i.e. the free electron dispersion crosses into the continuum of electron plus one phonon states . here
, rs predicts an unphysical peak in the polaron dispersion ( see the denominator of ( [ eq : e_rs ] ) ) and it fails for larger @xmath266 . , @xmath239 , and the corresponding @xmath31 weight @xmath270 and average phonon number @xmath271 .
results are for @xmath272 , @xmath234 , @xmath273 .
three high - symmetry cuts in the brillouin zone are shown .
, title="fig : " ] , @xmath239 , and the corresponding @xmath31 weight @xmath270 and average phonon number @xmath271 .
results are for @xmath272 , @xmath234 , @xmath273 .
three high - symmetry cuts in the brillouin zone are shown .
, title="fig : " ] , @xmath239 , and the corresponding @xmath31 weight @xmath270 and average phonon number @xmath271 .
results are for @xmath272 , @xmath234 , @xmath273 .
three high - symmetry cuts in the brillouin zone are shown .
, title="fig : " ] the second coupling @xmath274 is roughly in the cross - over regime , see also fig .
[ fig:1d_e ] . here
ma gives a much better agreement with qmc than scba or the perturbational theories .
the polaron bandwidth is already renormalized and well below the weak coupling value of @xmath240 .
in fact , as we show later , there is another bound state between these states and the continuum .
this is not captured in scba , which always predicts a polaron bandwidth of @xmath9 with a continuum above , and roughly between zero and one average number of phonons as @xmath266 increases from 0 to @xmath275 .
for the strong coupling @xmath276 , low - energy properties become almost @xmath266-independent , as expected since the lang - firsov impurity limit is being approached .
a second such comparison is possible for the 2d case with @xmath191 and @xmath273 , where qmc data is available , see fig .
[ fig:2d_ek ] .
this coupling is on the weak side of the crossover , with the gs @xmath31 weight still large , @xmath277 . here
ma is already doing better than scba .
as in the 1d weak - coupling case , one can again see that the ma polaron bandwidth is slightly larger than @xmath9 .
ma also somewhat overestimates the average number of phonons for large @xmath14 values .
overall , given that all this data is in the crossover regime where the ma is at its worst , one can conclude that ma is also reasonably accurate in capturing low - energy polaron behavior .
the main motivation in trying to find an approximation for the green s function is that this quantity gives not only low - energy information , but the whole spectrum .
here we compare ma predictions with various high - energy results available in the literature .
unfortunately there are much fewer of these , because the computational effort to obtain the whole spectrum through the usual numerical approaches is generally forbidding . between gs and first excited @xmath190 state .
a second bound state is stable if @xmath278 ( here @xmath279 ) ; ( b ) line below which a second bound state appears : ed@xcite ( circles ) and ma ( squares ) ; ( c ) ma @xmath31 weight of the second bound state when stable ( red squares ) , and that of the gs ( black squares ) , for @xmath234 .
the circle is the one qmc result available for the @xmath31 weight of the second bound state.@xcite these results are for 1d . ] vs. @xmath280 , for @xmath281 and @xmath275 .
ma results ( red line ) vs. data from ref .
( black line ) .
parameters are @xmath282 , @xmath203 , @xmath283 . ]
we begin with a comparison against exact diagonalization ( ed ) 1d data , from ref . [ ] .
the results are shown in fig .
[ fig:2dpeak ] .
as already discussed , for weak coupling there is a continuum starting at @xmath284 , but for stronger coupling a second so - called bound state appears below the continuum . in fig .
[ fig:2dpeak](a ) we track the energy @xmath285 of the second @xmath190 state . for small couplings ,
the data actually shows the maximum dos in the continuum , not its edge ( the maximum is generally located close to the lower edge .
this data shows again that ma somewhat overestimates this energy , which should be @xmath286 ) . when @xmath287 , there is a true discrete state . note
that panel @xmath288 is in very good quantitative agreement with similar data shown in fig .
8 of ref . .
the only difference is for strong coupling , where the ed data shows @xmath289 again , however , with significant finite - size dependence on the chosen hilbert space cutoff .
we can thus find the coupling @xmath290 where the second bound state appears , for different values of @xmath241 .
this line is shown in panel ( b ) , together with the ed results .
the agreement between the two data sets is excellent , even at small @xmath241 values where we expect ma to be less accurate .
we also show in panel ( c ) the @xmath31 weight of this second bound state , where stable .
this data is not given in ref .
, however one qmc point is available in ref .
, in good agreement with the ma prediction . vs. @xmath280 , for @xmath281 and @xmath275 .
ma results ( red line ) vs. data from ref .
( black line ) .
parameters are @xmath282 , @xmath274 , @xmath283 . ] vs. @xmath280 , for @xmath281 and @xmath275 .
ma results ( red line ) vs. data from ref .
( black line ) .
parameters are @xmath282 , @xmath201 , @xmath283 . ]
we now move to comparisons for the entire spectral weight @xmath291 . in figs .
[ cat05 ] , [ cat1 ] and [ cat2 ] we show comparisons for a 1d system with @xmath282 and three different coupling strengths @xmath192 and 2 , respectively . in each case
, data for 5 values of @xmath266 , namely @xmath292 and @xmath275 are shown .
the numerical data ( black line ) is obtained using a variational method by de filippis et al .
@xcite numerical data obtained through exact diagonalization of a finite system and by qmc , for the same parameters but somewhat different @xmath266 values , is also presented by hohenadler et al . in refs . .
these sets of numerical data are in good agreement with one another . in all three cases
the agreement between ma results and the numerical data is very good .
as expected , it is best for the largest @xmath235 , but even for the smaller @xmath235 values , which are just below and within the cross - over region , the agreement is very satisfactory . for @xmath293 and @xmath190 ( upper panel of fig . [ cat05 ] ) we see the polaron state as a lorentzian peak ( a broadening @xmath283 was used ) which accounts for most of the weight , and a small continuum at a higher energy .
ma overestimates the gap between the two , which should be @xmath9 . as @xmath266 increases
, the polaron peak disperses but also looses significant weight , as discussed in the previous section .
most of the weight is now in the high energy continuum , located roughly near the corresponding @xmath294 value .
this simply shows that these higher energy states are not significantly affected by this rather weak coupling .
the vm data shows somewhat more structure in these continua than the ma data , but most of the weight occupies similar frequency ranges . for @xmath274 and @xmath190 ( upper panel of fig .
[ cat1 ] ) , the ma data shows 3 lorentzian peaks plus a continuum starting at @xmath295 . for the rather large @xmath223
used it is hard to distinguish which peaks come from individual poles , and which are true continua .
this can be easily done by studying their behavior as a function of the broadening @xmath223 , as shown in fig .
[ etacomp ] .
the height of peaks corresponding to discrete states scales precisely like @xmath296 , as expected for lorenzians .
the continuum is affected very little by changes in @xmath223 , except the peak near its lower edge where the finite @xmath223 smoothes out a singularity in the dos . since
this singularity is not of the @xmath297 type , its scaling with @xmath223 is different from that of the lorentzians .
the two lower states are closer to one another than @xmath9 , however the ma data shows no sign of the continuum that is expected to start at @xmath298 .
note that the numerical data in fig .
[ cat1](a ) shows more structure , that could be consistent with this continuum .
we will address the issue of this continuum below . as @xmath266 is increased ( see fig .
[ cat1 ] ) the low - energy peaks show some dispersion , but with a strongly renormalized bandwidth . at higher @xmath266 most weight shifts again at high energies , in a rather broad continuum . finally , for @xmath201 and @xmath190 , fig .
[ cat2 ] shows even more discrete peaks spaced by @xmath9 .
the gs is at @xmath299 , but its weight is so small that it can not be seen on this scale , unless @xmath223 is decreased significantly .
a continuum is seen above @xmath300 .
as @xmath266 is now increased , there is almost no dispersion of the discrete peaks , however the weight shifts again to an even broader high - energy continuum . vs. @xmath280 , for @xmath282 , @xmath274 , and @xmath301 .
the first three peaks are discrete states ( lorentzians ) whereas the fourth marks the band - edge singularity of the continuum . ] for @xmath191 , @xmath302 and various @xmath235 values , from exact diagonalization@xcite ( black ) and ma ( red ) . for @xmath303 ,
the arrows point the gs location .
the insets show the same data on a smaller scale , so that low - weight states are more visible .
the arrows show the gs and the state appearing at precisely @xmath9 above gs , in the ed results . ]
the issue of the continuum at @xmath284 in the exact case , and of its absence in the ma approximation for moderate and large couplings can be understood from fig .
[ bayo ] . here
we show a comparison of @xmath304 in 2d obtained from exact diagonalization,@xcite vs. ma results , for various couplings @xmath235 .
the first remark is that ma captures quite well all the large - weight features , both as far as their energy and their weight are concerned .
this is expected , given the good sum rules agreement demonstrated previously . however , the ed results clearly show more states than ma predicts .
there is always a low - weight peak at precisely @xmath305 ( the gs and this state are marked by arrows in the insets ) . for couplings of up to @xmath306 , this peak
is followed by several nearby peaks with comparably low - weight , which can be argued to be part of the expected continuum . for larger @xmath235 , however , only the state at @xmath305 can still be seen , although more states suggesting more continua are seen between the large higher - energy peaks .
the gradual disappearance of the first continuum is not surprising , since one expects its width to narrow exponentially as the coupling increases .
moreover , one expects that the largest contribution to this continuum is from states with one or more phonons , explaining their low @xmath31 weight .
as far as ma is concerned , fig . [ bayo ] suggests that for couplings where there is more than one discrete state , the very little weight in the @xmath305 and similar higher - energy continua is absorbed in the discrete states predicted by ma .
this is consistent with the systematic up - shift of the ma peaks compared to the ed data .
in fact , it is straightfoward to see that the ma approximation can only predict a continuum starting at @xmath307 .
a continuum is signaled by a finite imaginary part of @xmath217 , and the lowest frequency where this can occur is that for which @xmath308 acquires a finite imaginary part [ see eq .
( [ eq : sigma_ma ] ) ] . however , the imaginary part of @xmath113 is proportional to the total density of states of the free model , i.e. it is finite for @xmath309 $ ] for nearest - neighbor hopping .
it follows that the ma continuum always starts precisely at @xmath310 .
this explains why for small @xmath235 , where there is only one peak below this continuum , the gap between the two is somewhat larger that the expected @xmath9 value : the gs energy decreases below @xmath40 with increasing @xmath235 , whereas the continuum edge is pinned at @xmath311 , in the ma approximation .
as the coupling increases , bound states start to split from this continuum , and spectral weight is shifted to lower energies , in good agreement with the sum - rule predictions of the exact solution .
these new bound states have to account for the ( small ) weight that is present in lower energy continua , in the exact solution , and this is precisely what fig .
[ bayo ] shows . clearly , a self - energy that would account for these continua as well would have to be a lot more complicated than that of eq .
( [ eq : sigma_ma ] ) .
weight and average number of phonons in the gs ( black line ) and the next three higher @xmath190 bound states , when they become stable according to ma .
results are for @xmath263 , @xmath191 . ]
this shows again that ma is remarkably successful in predicting the main features of the green s function , given its simplicity and trivial numerical cost .
this should make it ( and generalizations of it to other models ) of large interest for comparison against experiments . in the following , we use ma to investigate more properties of the green s function . to our knowledge , there are few or no similar results with similar accuracy published in the literature .
we begin with fig .
[ peaks ] , where we plot the @xmath31 weight and average phonon numbers for a few of the higher - energy peaks , once they appear below the continuum . for comparison
, we also show the already discussed gs results ( black line ) .
results in higher dimension are qualitatively similar and we do not show them here . unlike for the gs ,
both these quantities are non - monotonic functions of @xmath235 for all higher - energy bound states .
each of these states disperses with @xmath266 , like in figs .
[ cat1 ] and [ cat2 ] ( more data for this is shown below ) , so an effective mass can be associated with each such band .
this effective mass satisfies @xmath312 , and therefore also shows non - monotonic behavior , first decreasing and then increasing as @xmath235 is increased . the average phonon number in the @xmath66th state must approach @xmath313 asymptotically , as can be verified in the lang - firsov limit
this is indeed observed in fig .
[ peaks ] , however the plateaus seen at moderate @xmath235 suggest some cross - over from one to another type of wavefunction associated with these higher levels . as @xmath50 , an infinite sequence of such bound states appear , as expected in the lang - firsov limit . as a function of @xmath266 and @xmath280 .
the intensity scales are shown to the right of each plot .
parameters are @xmath314 , @xmath315 and @xmath316 , respectively , and @xmath317 .
these results are in excellent agreement with the results of hohenadler _ et al .
_ , ref . , and the gs results of bonca _ et al .
_ , ref . . , title="fig : " ] as a function of @xmath266 and @xmath280
the intensity scales are shown to the right of each plot .
parameters are @xmath314 , @xmath315 and @xmath316 , respectively , and @xmath317 .
these results are in excellent agreement with the results of hohenadler _ et al .
_ , ref . , and the gs results of bonca _ et al .
_ , ref . .
, title="fig : " ] as a function of @xmath266 and @xmath280 .
the intensity scales are shown to the right of each plot .
parameters are @xmath314 , @xmath315 and @xmath316 , respectively , and @xmath317 .
these results are in excellent agreement with the results of hohenadler _ et al .
_ , ref . , and the gs results of bonca _ et al .
_ , ref . .
, title="fig : " ] for a better illustration of the appearance of these bound states and of their evolution , we show contour plots of the spectral weight @xmath318 .
we begin by plotting @xmath318 as a function of @xmath14 and @xmath280 , for fixed parameters @xmath319 and @xmath9 . in fig .
[ 1dkcont ] we show 1d results corresponding to @xmath314 and @xmath315 and 2 , respectively .
only half of the bz is shown , since time - invariance guarantees that @xmath320 .
each of these ma contour plots takes below ten seconds to generate .
note that similar plots for the same parameters were provided by hohenadler _
_ in ref . , based on a cluster perturbation theory approach
the agreement between the main features of our and their plots is excellent .
as expected , their data does show a few more low - weight features at lower energies , below @xmath321 , in this case , where our continuum starts .
such contour plots are richer versions of plots like those shown in figs .
[ cat05 ] , [ cat1 ] and [ cat2 ] .
they illustrate basically the same points , although one can now also see clearly the dispersion of various features . for the lowest @xmath235 ,
the free electron dispersion @xmath322 is still almost visible , except that electron - phonon interactions split it into the lower polaron band , and the higher continuum .
this continuum is not featureless , instead one can already see weight accumulating near its lower edge . as @xmath235 increases , a new bound state will split off from this .
indeed , this is seen for @xmath274 , where there are 2 bound state below the continuum starting at @xmath323 ( for these parameters ) .
the bandwidth of each of these states is now narrowed below @xmath9 .
the weight in the continuum at higher energies is redistributed suggesting the impending formation of yet more bound states .
indeed , the @xmath201 data shows 4 , even narrower bound states below the continuum , which is showing signs of further fragmentation at multiples of the phonon frequency . as a function of @xmath14 and @xmath280 , along several cuts in the bz .
the intensity scales are shown to the right of each plot .
parameters are @xmath324 , @xmath325 and @xmath316 , respectively , and @xmath326 .
the middle panel is in excellent agreement with the results of hohenadler _ et al . _ given in ref . .
, title="fig : " ] as a function of @xmath14 and @xmath280 , along several cuts in the bz .
the intensity scales are shown to the right of each plot .
parameters are @xmath324 , @xmath325 and @xmath316 , respectively , and @xmath326 .
the middle panel is in excellent agreement with the results of hohenadler _ et al . _ given in ref . .
, title="fig : " ] as a function of @xmath14 and @xmath280 , along several cuts in the bz .
the intensity scales are shown to the right of each plot .
parameters are @xmath324 , @xmath325 and @xmath316 , respectively , and @xmath326 .
the middle panel is in excellent agreement with the results of hohenadler _ et al . _ given in ref . .
, title="fig : " ] similar behavior is expected , and indeed seen , in higher dimensions . here
we only show similar 2d contour plots , for @xmath324 and @xmath325 and @xmath316 , in fig .
[ 2dkcont ] .
the middle panel again agrees very well with data shown in ref . .
in this case , the ma continuum starts at @xmath327 .
as @xmath235 increases , we see again first 1 , then 2 and then 4 bound states below the continuum .
their bandwidths decrease with increasing @xmath235 , so that the lowest band for @xmath201 is already almost dispersionless , even though its weight still varies with @xmath14 .
as in the 1d case , as the interaction becomes stronger , the weight in the continuum also redistributes itself , with strong resonances seen around multiples of the phonon frequency . and @xmath328 .
( a ) @xmath329 vs. @xmath280 and @xmath235 ; ( b ) @xmath330 vs. @xmath280 and @xmath235 , on a linear scale ; ( c ) @xmath330 vs. @xmath280 and @xmath235 , on a logarithmic scale ; ( d ) total spectral weight @xmath331 vs. @xmath280 and @xmath235 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] and @xmath328 .
( a ) @xmath329 vs. @xmath280 and @xmath235 ; ( b ) @xmath330 vs. @xmath280 and @xmath235 , on a linear scale ; ( c ) @xmath330 vs. @xmath280 and @xmath235 , on a logarithmic scale ; ( d ) total spectral weight @xmath331 vs. @xmath280 and @xmath235 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] and @xmath328 .
( a ) @xmath329 vs. @xmath280 and @xmath235 ; ( b ) @xmath330 vs. @xmath280 and @xmath235 , on a linear scale ; ( c ) @xmath330 vs. @xmath280 and @xmath235 , on a logarithmic scale ; ( d ) total spectral weight @xmath331 vs. @xmath280 and @xmath235 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] and @xmath328 .
( a ) @xmath329 vs. @xmath280 and @xmath235 ; ( b ) @xmath330 vs. @xmath280 and @xmath235 , on a linear scale ; ( c ) @xmath330 vs. @xmath280 and @xmath235 , on a logarithmic scale ; ( d ) total spectral weight @xmath331 vs. @xmath280 and @xmath235 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] another way to understand the dependence on the coupling @xmath235 ( or any other parameter ) is to plot a contour of the spectral weight @xmath318 vs. @xmath235 and @xmath280 for a fixed value of @xmath14 . such a task is equally trivial at the ma level .
in fact , one can also just as easily calculate and plot the total density of states , or spectral weight : @xmath332 since within the ma approximation this is given by : @xmath333 \\ \nonumber & = & -\frac{1}{\pi } \textrm{im } \
, \left\ { \bar{g}_0 \left [ \omega - \sigma_{\textrm{ma}}(\omega ) \right ] \right\}.\end{aligned}\ ] ] in fig .
[ 1dc ] we show 4 such contour plots for the 1d case .
the uppermost one shows @xmath329 vs. @xmath280 . for @xmath49 ( noninteracting case ) ,
only one state exists at @xmath334 , as expected .
as the coupling turns on , the energy of this state ( the ground - state ) decreases , but @xmath190 weight is also transferred to higher energies , due to hybridization with the states in the electron - plus - one - phonon continuum .
the ma continuum here starts at @xmath335 . for moderate and larger
@xmath235 one can clearly see how weight is re - arranged inside the continuum as @xmath235 increases , and new bound states split from it and move towards lower energies .
the apparent `` break '' in the slope of the gs energy , as @xmath235 increases , is now seen to occur when the first bound state approaches the gs , and is suggestive of an avoided crossing . from this point on the gs
lowers its energy much faster , but its weight also decreases dramatically and it becomes difficult to see .
the second panel of fig .
[ 1dc ] shows @xmath330 vs. @xmath280 , on a linear scale . at @xmath49 , there is only one peak at @xmath336 , as expected .
as the coupling is turned on , this weight seems to spread around in a rather featureless , broad continuum .
in fact , on a logarithmic scale ( panel ( c ) of fig . [ 1dc ] ) one can see that @xmath200 weight is pulled down into all the bound states , in agreement with the previous data we showed .
this weight , however , is so small that it is not visible on the linear scale . finally , the lowest panel of fig .
[ 1dc ] shows the total spectral weight , or dos . at @xmath49
we see the usual 1d dos , with the singularities near the band - edges rounded off because of the finite @xmath223 used . as @xmath235
is turned on , one can recognize both the contribution from the @xmath190 and @xmath200 states to the total dos : each bound state has a finite bandwidth due to its dispersion ( this is to be contrasted to the upper panels , where the bound states are true delta - functions , with a width defined by the broadening @xmath223 ) .
as the coupling strength increases , the number of bound states increases ; they are spaced by roughly @xmath9 , their bandwidths narrow down , and their weights also decrease . in other words , they approach the expected lang - firsov behavior . thus , this figure actually answers the question posed in the introduction , regarding the evolution of the spectral weight from that of a free electron towards that of the impurity limit .
while the ma results are certainly not exact , we can claim with a high degree of certainty that the main features are accurately captured , especially in the weak and in the strong coupling limits .
this suffices to understand the physics of this problem , and given the simplicity of the approximation , to investigate in detail any number of other quantities we have not shown here , such as the self - energy .
of course , if one is interested in exact results for some particular set of parameters , numerically - intensive methods have to be used . and @xmath328 .
top : @xmath337 vs. @xmath280 and @xmath235 .
bottom : total 2d spectral weight @xmath331 vs. @xmath280 and @xmath235 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] and @xmath328 .
top : @xmath337 vs. @xmath280 and @xmath235 .
bottom : total 2d spectral weight @xmath331 vs. @xmath280 and @xmath235 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] .
top : @xmath338 vs. @xmath280 and @xmath235 , for @xmath339 .
bottom : total 3d spectral weight @xmath331 vs. @xmath280 and @xmath235 , for @xmath328 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] .
top : @xmath338 vs. @xmath280 and @xmath235 , for @xmath339 .
bottom : total 3d spectral weight @xmath331 vs. @xmath280 and @xmath235 , for @xmath328 .
the intensity scales are shown to the right of each plot .
, title="fig : " ] qualitatively similar plots are obtained in 2d and 3d , as shown in figs .
[ 2dc ] and [ 3dc ] .
of course , the @xmath49 dos is very different , with a van - hove singularity at @xmath340 for the 2d case , and the characteristic nearest - neighbor hopping dos in the 3d case .
however , the appearance of multiple bands below the continuum as @xmath235 increases and all the remaining phenomenology is very similar .
thus , we see no evidence of any qualitative differences in the polaron physics due to different dimensionality .
in this paper we analyzed the green s function of the holstein polaron , using the _ momentum average _ approximation , which consists in summing all the self - energy diagrams , but with each individually averaged over all its free propagator momenta .
the resulting self - energy can be written as an infinite continuous fraction that is numerically trivial to evaluate .
this procedure becomes exact in the limits @xmath32 and @xmath35 .
we gauged the accuracy of this approximation by computing its corresponding spectral weight sum rules and comparing them against the exact sum rules , which are known for this type of hamiltonian .
we showed that the ma spectral weight satisfies exactly the first 6 sum rules . even though this is quite impressive at first sight , it is actually no guarantee of overall accuracy , as the case of scba demonstrates .
the scba spectral weight always satisfies exactly the first 4 sum rules , even at large couplings @xmath235 where it predicts very wrong results !
the meaningful criterion of accuracy for the sum rules is to show that the vast majority of terms in _ all sum rules _ , and in particular the dominant terms in the various asymptotic limits , are captured by the approximation .
ma indeed satisfies this very restrictive criterion .
the accuracy of the approximation was also tested by direct comparison with data obtained through numerically intensive methods . in all cases
we obtain remarkable agreement , especially considering the ease of evaluation of the ma results .
the ma approximation is not exact and some features are not correctly captured ( for example , the continuum starting at @xmath284 ) however , in all cases , all the higher - weight features in the spectral weight are quantitatively and qualitatively well described by the ma approximation . trading some of the accuracy of numerically exact but time consuming methods in exchange for very fast results which capture the main features accurately is a useful approach when trying to understand the main aspects of the physics of a problem , as well as when one is concerned about comparison with experiments .
it is very unlikely that arpes could capture very low - weight features , these would be lost in noise statistics .
thus , an approximation like ma , that quickly but accurately estimates results is very useful , to be followed , of course , by detailed numerics for the parameter sets of interest .
it is important to note that the existence and accuracy of approximations like ma is not guaranteed , in fact it can be regarded as a surprise for the case of the holstein polaron , that has been under investigation for almost five decades .
however , this demonstration of its existence and efficiency in the holstein polaron problem gives some hope for making progress for the general class of strongly - correlated systems problems .
one can always use some flavor of perturbation theory to understand behavior in asymptotic limits , but the really challenging problems are set in regimes where perturbation does not apply .
the ma approximation suggests that one way to make non - trivial progress is to sum all diagrams , with each simplified enough so as to make the calculation feasible , but not so much as to really alter the physics .
this is a very different approach from the usually employed summation of a subclass of diagrams .
note that there are many classes of problems with diagrams similar to the ones arising in the single polaron problem , although of course with different propagators and/or vertices .
a first possible generalization of this work is to models with several phonon branches , and/or momentum - dependent coupling @xmath341 , and/or dispersive phonons , @xmath342 .
the way to achieve this for a single polaron is briefly discussed in ref . .
given the length of this article , we postpone this discussion for future publications where results can also be shown .
other directions of generalization are to finite particle densities and/or finite temperatures , and indeed to hamiltonians which also include electron - electron interactions . such work is currently in progress .
it is still far from clear which cases will admit useful generalizations , however the proof of existence of this method for the holstein polaron problem is in itself encouraging .
* acknowledgments : * we thank t. devereaux , f. marsiglio , a. mishchenko and n. nagaosa for useful discussions , and a. macridin , v. cataudella , g. de filippis and b. lau for allowing us to use their numerical results .
this work was supported by nserc and cfi , and by ciar nanoelectronics and the alfred p. sloan foundation ( m.b . ) and ciar quantum materials ( g.s . ) .
the ma self - energy of eq .
( [ eq : sigma_ma ] ) looks similar to the dmft self - energy , discussed in ref . .
this is not so surprising , since both become equal to the exact lang - firsov limit if the bandwidth goes to zero , and this can be re - written as an infinite continued fraction , as shown in eq .
( [ t0 ] ) . however , this similarity may raise questions about the relationship between the two approximations . in this appendix , we briefly show that the two approximations are very different at all finite @xmath52 and @xmath37 .
the meaning of the ma @xmath113 and of corresponding dmft @xmath343 ( notation of ref . ) is very different .
the dmft @xmath343 is obtained by solving exactly the problem of an impurity coupled to an environment in the @xmath344 limit , and then imposing the self - consistency condition that the impurity site behaves similarly to all other sites in the environment .
the dmft @xmath343 is calculated self - consistently using the following steps:@xcite ( i ) with some initial guess for @xmath343 , one calculates the dmft self - energy , given by a formula similar to eq .
( [ eq : sigma_ma ] ) , with @xmath113 replaced by @xmath343 ; this self - energy is then used to calculate the total green s function @xmath345 where @xmath346 is the density of states of the non - interacting electrons .
the usual procedure is to take @xmath344 and use as dos the semielliptical formula corresponding to an infinitely branched bethe tree , @xmath347 where @xmath348 is the bandwidth for the non - interacting system .
then , ( iii ) the new @xmath343 is extracted from the condition that @xmath349^{-1}$ ] and the procedure is repeated until self - consistency is reached .
reaching self - consistency is a non - trivial numerical task , especially compared to obtaining the ma @xmath113 ( see below ) .
more importantly , the dmft @xmath343 is an explicit function of @xmath37 and @xmath9 .
in contrast , the ma @xmath113 is the momentum average of the free propagator : thus , it is a known function , independent of the parameters @xmath37 and @xmath9 . in particular , for the semi - elliptical dos , we have simply : @xmath352.\end{aligned}\ ] ] a comparison of these functions is provided in fig . [ fig : dmft ] .
they are clearly different . moreover ,
note that in the ma approximation , @xmath103 is an explicit function of the momentum .
the ma self - energy for the holstein polaron happens to be independent of the momentum , but this a consequence of the simplicity of the model , not an `` in - built '' feature like in dmft .
generalizations to models which have a momentum - dependent coupling and/or dispersive phonons , lead to momentum - dependent self - energies.@xcite a. lanzara , p. v. bogdanov , x. j. zhou , s. a. kellar , d. l. feng , e. d. lu , t. yoshida , h. eisaki , a. fujimori , k. kishio , j .-
shimoyama , t. noda , s. uchida , z. hussain and z .- x .
shen , nature * 412 * , 510 ( 2001 ) .
k. m. shen , f. ronning , d. h. lu , w. s. lee , n. j. c. ingle , w. meevasana , f. baumberger , a. damascelli , n. p. armitage , l. l. miller , y. kohsaka , m. azuma , m. takano , h. takagi , and z .- x .
shen , phys .
lett . * 93 * , 267002 ( 2004 ) .
alexandrov , v.v .
kabanov , and d.k .
ray , phys .
b * 49 * , 9915 ( 1994 ) ; g. wellein , h. rder , and h. fehske , phys . rev .
b * 53 * , 9666 ( 1996 ) ; g. wellein and h. fehske , phys . rev .
b * 56 * , 4513 ( 1997 ) ; e.v.l . de mello and j. ranninger , phys .
b * 55 * , 14872 ( 1997 ) ; m. capone , w. stephan , and m. grilli , phys . rev .
b * 56 * , 4484 ( 1997 ) ; g. wellein and h. fehske , phys . rev .
b * 58 * , 6208 ( 1998 ) ; o. s. barisic , phys . rev .
b * 65 * , 144301 ( 2002 ) .
f. marsiglio , phys .
b * 42 * , 2416 ( 1990 ) ; h.
de raedt and a. lagendijk , phys .
lett . * 49 * , 1522 ( 1982 ) ; _ ibid _ phys .
b * 27 * , 6097 ( 1983 ) ; _ ibid _ * 30 * , 1671 ( 1984 ) ; p. kornilovitch , j. phys . : condens .
matter * 9 * , 10675 ( 1997 ) ; p. e. kornilovitch and e. r. pike , phys .
b * 55 * , r8634 ( 1997 ) ; p. e. kornilovitch , phys . rev . lett . *
81 * , 5382 ( 1998 ) ; p. e. kornilovitch , phys .
b * 60 * , 3237 ( 1999 ) . | we present a novel , highly efficient yet accurate analytical approximation for the green s function of a holstein polaron .
it is obtained by summing all the self - energy diagrams , but with each self - energy diagram averaged over the momenta of its free propagators .
the result becomes exact for both zero bandwidth and for zero electron - phonon coupling , and is accurate everywhere in the parameter space .
the resulting green s function satisfies exactly the first six spectral weight sum rules .
all higher sum rules are satisfied with great accuracy , becoming asymptotically exact for coupling both much larger and much smaller than the free particle bandwidth .
comparison with existing numerical data also confirms this accuracy .
we use this approximation to analyze in detail the redistribution of the spectral weight as the coupling strength varies . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.